English in Elementary Schools Multilingualism and Language Teaching Herausgegeben von Thorsten Piske (Erlangen), Silke Jansen (Erlangen) und Martha Young-Scholten (Newcastle) Band 7 Anja Steinlen English in Elementary Schools Research and Implications on Minority and Majority Language Children’s Reading and Writing Skills in Regular and Bilingual Programs © 2021 · Narr Francke Attempto Verlag GmbH + Co. KG Dischingerweg 5 · D-72070 Tübingen Das Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt. Jede Verwertung außerhalb der engen Grenzen des Urheberrechtsgesetzes ist ohne Zustimmung des Verlages unzulässig und strafbar. Das gilt insbesondere für Vervielfältigungen, Übersetzungen, Mikroverfilmungen und die Einspeicherung und Verarbeitung in elektronischen Systemen. Internet: www.narr.de eMail: info@narr.de CPI books GmbH, Leck ISSN 2199-1340 ISBN 978-3-8233-8451-9 (Print) ISBN 978-3-8233-9451-8 (ePDF) ISBN 978-3-8233-0252-0 (ePub) Bibliografische Information der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über http: / / dnb.dnb.de abrufbar. www.fsc.org MIX Papier aus verantwortungsvollen Quellen FSC ® C083411 ® www.fsc.org MIX Papier aus verantwortungsvollen Quellen FSC ® C083411 ® 9 1. 11 2. 15 2.1 15 2.1.1 15 2.1.2 17 2.1.3 19 2.1.4 21 2.1.5 22 2.2 23 2.2.1 24 2.2.2 25 2.2.3 27 2.2.4 27 2.2.5 29 2.3 30 2.3.1 31 2.3.2 33 2.3.3 34 2.3.4 37 2.3.5 38 2.3.6 39 2.3.7 40 2.3.8 40 2.3.9 41 Contents Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Literature review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . L1 reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . L2 reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . L1 writing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . L2 writing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Some problems relating to phoneme-grapheme correspondences in German and English . . . . . . . . . . . Regular elementary school English programs in Germany . . . Controversial views on EFL reading and writing in elementary school . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Curricula for English-as-a-subject relating to reading and writing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . English textbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Studies on FL reading in regular EFL programs . . . . . . Studies on FL writing in regular EFL programs . . . . . . Bilingual programs in Germany . . . . . . . . . . . . . . . . . . . . . . . . . Differences between bilingual programs and EFL programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Different bilingual programs: CLIL vs. IM . . . . . . . . . . . Intensity of different bilingual programs . . . . . . . . . . . . Bilingual programs as “elitist” programs . . . . . . . . . . . . Curricula for bilingual teaching in Germany . . . . . . . . . Teacher supply for bilingual programs in Germany . . . Supply of materials for bilingual programs . . . . . . . . . . Assessment in bilingual programs . . . . . . . . . . . . . . . . . Studies on L1 and FL reading and writing in bilingual programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.10 42 2.3.11 44 2.3.12 45 2.4 46 2.4.1 47 2.4.2 49 2.4.3 51 2.5 62 2.5.1 62 2.5.2 63 2.5.3 64 2.5.4 68 2.6 69 2.6.1 70 2.6.2 71 2.7 72 2.7.1 74 2.7.2 74 2.7.3 75 3. 77 4. 79 4.1 79 4.1.1 79 4.1.2 80 4.1.3 81 4.1.4 82 4.2 83 4.2.1 83 4.2.2 84 4.2.3 85 4.3 86 L1 reading and writing in bilingual programs . . . . . . . . FL skills in regular vs. bilingual programs . . . . . . . . . . . FL reading and writing in bilingual programs with different FL intensity . . . . . . . . . . . . . . . . . . . . . . . . . . . . Multilingual background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Facts and figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . L2 literacy (German) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . L3 literacy (English) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gender . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . L1 literacy skills in regular programs by gender . . . . . . L1 literacy skills in bilingual programs by gender . . . . FL literacy skills by gender . . . . . . . . . . . . . . . . . . . . . . . The interaction of gender and minority language background in language learning . . . . . . . . . . . . . . . . . . Cognitive variables (with a focus on nonverbal intelligence) . Nonverbal intelligence and bilingual children . . . . . . . Literacy skills by nonverbal intelligence . . . . . . . . . . . . Social background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . L1 literacy skills by SES . . . . . . . . . . . . . . . . . . . . . . . . . . FL literacy skills by SES . . . . . . . . . . . . . . . . . . . . . . . . . . The interaction of SES and language background on literacy skills . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Research questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The schools and their programs . . . . . . . . . . . . . . . . . . . . . . . . . The Bili-50 program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The EFL-program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Bili-20 program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Bili-70 program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Test material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Control measures: Family background and cognitive background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . English tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . German tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Contents 5. 89 5.1 89 5.2 91 6. 97 6.1 97 6.1.1 97 6.1.2 99 6.2 106 6.2.1 106 6.2.2 109 6.3 115 6.3.1 115 6.3.2 118 6.3.3 119 6.4 120 6.4.1 121 6.4.2 123 6.5 124 6.5.1 124 6.5.2 126 6.5.3 129 7. 133 7.1 134 7.2 136 7.3 139 7.3.1 140 7.3.2 141 7.3.3 144 7.4 145 7.5 149 7.6 150 7.7 152 7.7.1 152 7.7.2 156 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Within-group comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Between-group comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Effectivity of the FL programs . . . . . . . . . . . . . . . . . . . . . . . . . . German . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . English . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Language background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . German . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . English . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gender . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . German . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . English . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interaction between gender and language background Control variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cognitive background . . . . . . . . . . . . . . . . . . . . . . . . . . . . Social background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary, limitations and future studies . . . . . . . . . . . . . . . . . . Summary of results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Future studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Recommendations for teaching reading and writing in the FL classroom Starting with FL literacy activities as early as possible . . . . . . Creating a literacy-rich environment . . . . . . . . . . . . . . . . . . . . . Scaffolding for FL reading and writing activities . . . . . . . . . . . Verbal scaffolding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Content scaffolding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Learning process scaffolding and procedural scaffolding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Regular FL literacy activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . Awareness-raising literacy activities . . . . . . . . . . . . . . . . . . . . . Relevant and authentic literacy activities . . . . . . . . . . . . . . . . . Some notes on FL spelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Invented spelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Spelling activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Contents 7.7.3 158 7.7.4 159 7.8 161 7.9 167 8. 169 9. 173 The role of spelling errors in the FL classroom . . . . . . . Teaching spelling rules . . . . . . . . . . . . . . . . . . . . . . . . . . . Teacher feedback on FL students’ writing . . . . . . . . . . . . . . . . Conclusion on recommendations . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Contents Acknowledgements I would like to acknowledge a debt of gratitude to a great number of friends, family and colleagues who made it possible for this work to be carried out and who did not cease to offer their encouragement and support. First, I would like to thank the committee members Thorsten Piske, Thomas Herbst, Eva Breindl (Friedrich-Alexander-University Erlangen-Nürnberg) as well as Jens Möller (Christian-Albrechts-University Kiel) and Angela Hahn (Ludwig-Maximilians-University Munich) for evaluating my work. Although I claim as my own any deficiencies of this study, much of what it contains of value must be credited to Thorsten Piske. It has been a great pleasure to work with you over these many years. I would like to thank you for your friendship, your support and encouragement, your critical questions, your constructive feedback, and for countless stimulating and inspiring discussions. I would also like to thank Jens Möller (CAU Kiel) who has been with this project from its very beginning to the very end. His team has always been very helpful, and I am particularly grateful to Anna Zaunbauer and Johanna Fleckenstein for many valuable discussions (about this topic and much more). Andrea Abele-Brehm was my mentor in the ARIADNE mentoring program at FAU Erlangen-Nürnberg (2013-2015). Thank you for the right words at the right time and your advice regarding my first regression analyses. This project would not have taken place without the three elementary schools, namely the Muhlius Schule in Kiel (Ulrike Gerdes and Henning Roose) and the Platanus Schule Berlin (Dr. Karin Krüger and Zoia Grüning). In the Hügelschule in Tübingen (Eva Herzog and Jutta Friesch), we have been able to conduct tests annually since 2012: thank you so much for your ongoing support and patience! In particular, I would like to express my gratitude to the teachers of the three schools who tirelessly strive for supporting the children in their classes! You make this world a better place! After the tests had been carried out, the research assistants (Lea Pöschik, Jessica Schmidt, Anna Schrötter, and Johannes Schürmann, to name just a few) at Thorsten Piske’s Department for Foreign Language Education (Frie‐ drich-Alexander-University Erlangen-Nürnberg, Germany) always made sure that the data was ready for further analyses. I would have not been able to process these large amounts of data without you. Thank you! A good number of people also contributed substantially to the end product of this research: Nina Rogotzki has been my never-tiring proofreader for many years, and there is hardly any manuscript that did not go through her hands (or eyes) - thank you for many insightful comments on style and logic, and thank you for being my partner in crime on various trips to South America (and elsewhere). Henning Wode (CAU Kiel) was the first source of inspiration, as he was the one who directed my attention to bilingual school programs. In 1995 I set foot in a bilingual elementary school for the first time and helped with collecting data on the infamous “Frog” story. Thank you for supporting me ever since! I am also grateful for the fact that the “Wode Clan” (Petra Burmeister, Kristin Kersten, Annette Lommel, Thorsten Piske, Andreas Rohde, to name just a few) has been successfully collaborating over the years. I would like to thank the Narr Verlag, in particular Kathrin Heyng and her team, respectively, for their support of this book. In addition, I would like to extend my thanks to Keeley Madison for proof-reading the manuscript. Nothing would have been accomplished without my family. I would particu‐ larly like to thank my husband Jörg Wettlaufer for being the constant source of support, stability and love that helped me carry this project through to completion. I would also like to extend my gratitude to my mother for her never-ending belief and trust in me; and to Clemens, who lets me see the world with different eyes. Large parts of this and other manuscripts were written on countless train rides (“Thank you for travelling with Deutsche Bahn”) and also at the Bakery Küster, Windausweg in Göttingen: I would like to thank the staff for the friendly atmosphere and the endless supply of tea. This book is dedicated to the children for their enthusiastic participation, their endless curiosity and willingness to learn. You are our future! No wo/ man is an island. Thank you all for all kinds of support and encour‐ agement! 10 Acknowledgements 1 In Germany, elementary school generally starts at the age of 6. After four years (i.e., after grade 4), children enter secondary school (which generally consists of Hauptor Realschule, offering a six-year middle-school degree, and Gymnasium or Gesamtschule, offering high-school diplomas that provide access to tertiary education, e.g., KMK, 2013a). 1. Introduction This book deals with the development of literacy skills in elementary schools in Germany 1 , which offer English programs with different degrees of intensity, ranging from regular programs (with English-as-a-subject for two hours per week) to bilingual programs, in which one subject or several subjects (such as science, math, or music) are taught in the new language (in this case English). Literacy skills (i.e., reading and writing skills) constitute the focus of this book because they play a key role in the acquisition of academic knowledge and participation in education and society. Of special interest are minority language children (sometimes also called “children with a migration background”) who have often been reported to constitute an “at-risk group” in terms of academic achievement in school. In second and foreign language research, various factors have been claimed to affect language learning (e.g., Kersten, 2019, for a review). These effects are often subdivided into child-internal and child-external factors. Child-inter‐ nal factors, for example, may include language skills (i.e., knowledge and proficiency in the first, second and any additional language), gender, cognitive skills (e.g., nonverbal intelligence, working memory, phonological short-term memory, phonological awareness, executive control) as well as personality factors (e.g., traits, attitudes, motivation). Child-external factors involve the learner’s family/ social environment (e.g., parental education, socio-economic background, early cognitive stimulation, cultural capital), the environment in which the language is learned (which, for the educational context, may refer to educational policies, administration, program intensity and duration, teachers and classmates) as well as language input (provided by the teacher at school, which may vary in terms of quality). All these internal and external factors (and many more) affect children’s development in any language. However, a great number of studies have shown inconsistent results re‐ garding the impact of individual factors. For example, some studies have reported gender to play an important role in (foreign) language acquisition (e.g., BIG-Kreis, 2015), while others have not (e.g., Schmenk, 2002). Inconsistencies have also been reported with respect to students with a minority language background (who often have a migration background). In some studies, such students performed lower on foreign language tests (e.g., Elsner, 2010), while in other studies they performed just as well as comparable majority language peers (e.g., Kessler & Paulick, 2010). It is the aim of this book to disentangle some of these effects and to describe their impact in more detail, especially with regard to German and English reading and writing skills in the elementary school context. Therefore, the purposes of this book are a) to provide a summary of the research covering minority language children and language learning in various elementary school programs with different degrees of foreign language inten‐ sity; b) to introduce new, unpublished data to extend said research findings; and finally, c) to present recommendations regarding foreign language reading and writing activites in the elementary school classroom. The outline of this book is as follows: Chapter 2 presents a comprehensive literature review on the reading and writing skills of elementary school children. Characteristics of reading and writing in different acquisition settings are illustrated in chapter 2.1. Chap‐ ters 2.2 and 2.3 compare the curricular guidelines of regular and bilingual programs, as well as the supply of staff and materials in such programs; it also presents findings of empirical (often large-scale) studies on L1 (first language) and L2 (second / foreign language) reading and writing skills in Germany and elsewhere. Furthermore, in chapters 2.4, 2.5, 2.6, and 2.7, respectively, studies are reviewed which examine the effects of children’s language background, their gender, their cognitive background, and their social background on their reading and writing skills in German and English. Additional data are introduced in chapters 3 to 6. This new and hitherto unpublished study examines the overall effects of foreign language intensity with a large sample of fourth graders (N = 487), who attended one of four different elementary school programs differing in English intensity. Of special interest are not only effects of intensity but also effects of children’s language background, their gender, and their cognitive and social background, on their reading and writing skills in German and English. Chapter 3 presents the research questions; chapter 4 introduces the schools, the test materials and the subjects. The results of this study are illustrated in chapter 5 and discussed in detail in chapter 6, which also acknowledges the limitations of the study and, consequently, proposes ideas for further research. Chapter 7 is devoted to recommendations for teaching reading and writing in the FL elementary school classroom. Many of these recommendations include 12 1. Introduction ideas for the literacy-rich classroom, such as different types of scaffolding to facilitate students’ reading and writing output, as well as awareness-raising activities, which are embedded in authentic and relevant contexts (chapters 7.1 to 7.6). As spelling often is a neglected area in FL classrooms, chapter 7.7 deals with invented spelling, spelling activities, the role of spelling errors and teaching spelling rules, including recommendations on how to give feedback on student writing (chapter 7.8). Many of the recommendations for reading and writing activities presented in chapter 7 are not only aimed at teachers in FL classrooms, (in particular bilingual ones) but also at those teaching German-as-a-second / foreign language to students with family languages other than German, because in both contexts the focus is on fostering the target language while teaching subject content. A brief conclusion is presented in chapter 8, and references are listed in chapter 9. 13 1. Introduction 2. Literature review The focus of the following literature review is on children’s language back‐ ground (reviewed in chapter 2.4) and their institutional environment (i.e., program intensity, see chapters 2.2 and 2.3), because the present study deals with the linguistic development of minority and majority language children in elementary schools with different degrees of English intensity. However, other factors are also included. Chapter 2.5 reviews effects of gender on elementary schoolers’ development of German and English, and chapter 2.6 focuses on students’ cognitive skills, with special reference to nonverbal intelligence. Finally, chapter 2.7 deals with effects of students’ family and social environment, particularly with those relating to the parents’ educational background. As this study focuses on reading and writing, definitions and models of L1 (first language) and L2 (second / foreign language) reading and writing are provided at the beginning of chapter 2 (i.e., in chapter 2.1). 2.1 Definitions In this section, the terms “reading” and “writing” are defined, and reading and writing skills and processes are described in relation to L1 and L2 acquisition. Note that in this section, the term “L2” subsumes both second and foreign language acquisition. Minority language students’ foreign language reading and writing skills (which often constitute their L3) will be discussed in chapter 2.4. Literacy is traditionally defined as the ability to read and write, but this concept has evolved in meaning. The modern meaning of the term has been expanded to include the ability to use language, numbers, images, computers, and other basic means to understand, communicate, gain useful knowledge, solve mathematical problems and use the dominant symbol systems of a culture (e.g., UNESCO, 2012). However, in this book, the term “literacy” will refer only to reading and writing skills. 2.1.1 L1 reading Reading as such is nowadays understood as a comprehension process that occurs when the reader extracts and integrates information from the text and combines it with what is already known (e.g., Koda, 2005: 4). According to the OECD (2000: 21), reading literacy can be defined as the ability “to understand, use and reflect on written texts in order to achieve one’s goals, to develop one’s knowledge and potential, and to participate effectively in society”. These comprehension skills depend, among other things, on reading speed, which requires a high degree of short-term memory. Other determinants of reading competence include world knowledge, the ability to rapidly access lexical items, broad and in-depth vocabulary knowledge, reading motivation, a positive attitude towards reading, knowledge of text features, reading strategies, and other basic cognitive skills (Bundesministerium für Bildung und Forschung, BMBF, 2007). The reading process is characterized by an interaction of top-down and bottom-up processes (see e.g., Frisch, 2010; Thaler, 2012). In the latter, the reader recognizes meaning from the written text and moves from the individual parts to the whole, i.e., from letter, to syllable, to word, to phrase, to clause, to sentence, to paragraph and to the text (Grabe, 2009). In addition, meaning is derived via top-down processes with the reader at the center of the reading process: in order to infer meaning, s/ he uses general background knowledge, world knowledge and thematic knowledge, formal schemata (e.g., knowledge of text genres and text structures), content schemata (e.g., knowledge about the time in which a text was written), and finally scripts (i.e., knowledge about typical processes, e.g., how to check in at a hotel). These two processes continually interact: “The ‘construction’ of meaning that occurs in reading is a combination of ‘bottom-up’ processes […] and ‘top-down’ ones […]” (Ur, 1996: 141). L1 reading comprehension is generally preceded and affected by reading fluency. At a reasonable reading rate of approximately 250-300 words per minute, a reader is using very fast and efficient word recognition skills on the sublexical, lexical, and connected text levels, combining information from different sources while reading under rather intense time constraints (Grabe, 2010). That is to say, fluent reading includes both word-level reading skills and language processing/ comprehension skills (e.g., Jenkins, Fuchs, Van den Broek, Espin & Deno, 2003). In general, reading fluency is associated with the ability to read rapidly with ease and accuracy, and to read with appropriate expression and phrasing. It involves a long learning process, and text comprehension is an expected result of fluent reading (see, e.g., Grabe, 2009). Several developmental models of L1 reading have been put forward, most notably by Frith (1985) and Ehri (1995, see Kennedy, Dunphy, Dwyer, Hayes, McPhillips, Marsh, O’Connor & Shiel 2010 for a detailed review). Frith (1985) differentiated three stages, namely the logographic, alphabetic and orthographic stages. In the logographic stage, children use visual or graphic features to read words, while in the alphabetic stage, grapheme-phoneme relations are used to 16 2. Literature review process words. In the orthographic stage, children are finally able to use spelling patterns. Ehri (1995) proposed a similar developmental model which comprises four phases of reading development to identify the significant advances that occur as children learn to read. The four phases are pre-alphabetic, partial alphabetic, full alphabetic and consolidated alphabetic. During the pre-alpha‐ betic phase (which corresponds to Frith’s logographic stage) children rely on selected visual features (and not on letter-sound connections) to read words. Although they are essentially non-readers, they may ‘read’ environmental print from the contextual clues they notice. As children then learn the names and sounds of letters, they progress to the partial-alphabetic phase where they form connections between only some of the letters and sounds (often just the first and final letter sounds). During this phase they cannot segment sounds and will have difficulty decoding unfamiliar words, i.e., they do not have full knowledge of the alphabetic system yet. When children have reached the full-alphabetic phase, they learn sight words (i.e. words that are commonly used in reading and writing) and make connections between letters in written words and the corresponding sounds in speech; however, they use mainly grapheme-phoneme correspondences to identify words. Finally, the consolidated-alphabetic phase represents the children’s growing knowledge and use of specific orthographic patterns, as well as knowledge of morphological patterns and syllabic units. Such models (as proposed by Frith, Ehri and others) have also been criticized because the order of the stages is fixed and identical for all children; because no stage can be skipped so that proceeding to the next stage results only from mastery of the previous stage; because performance in the middle of each stage is homogenous; and because progression is characterized by qualitative changes, as processing strategies seem to differ radically from one stage to the next (e.g., Sprenger-Charolles & Casalis, 1996). 2.1.2 L2 reading Unlike L1 reading and writing, reading and writing in an L2 always involves the interaction of different languages. As Koda (2007: 1) points out, this “dual-language involvement implies continual interactions in between the two languages as well as incessant adjustments in accommodating the disparate demands each language imposes”. In other words, L2 reading and writing are inherently more complex than L1 reading and writing. In Grabe’s view (e.g., 2009; 2014: 10 ff.), differences between L1 and L2 reading center around the linguistic resources that are available to the L2 reader. For example, L2 readers have limited L2 language skills (i.e., knowledge of grammar, 17 2.1 Definitions vocabulary and discourse structure) when they begin reading. Furthermore, L2 students have less experience with reading exposure in the L2 and consequently less practice in L2 reading. In addition, L2 readers need to develop somewhat distinct cognitive processing which involves two language systems, using a joint strategy system, (e.g. Koda, 2005). Interference from the L1 (i.e., cognitive skills, strategies, goals and expectations) may affect L2 reading and either facilitate or hinder its outcomes. Finally, L2 readers rely on different combinations of general background knowledge when reading in the L2 because “world knowledge” (e.g., social and cultural assumptions in texts) may vary between L1 and L2 reading experiences. So far, L2 reading comprehension has been studied more frequently than L2 reading fluency, and there are far fewer studies dealing with reading fluency in the L2 than in the L1 (e.g., Grabe, 2010). Determinants of L2 reading fluency seem to be the same as those for L1 reading fluency and include automatic word recognition, a large recognition vocabulary, skilled grammatical processing, and automaticity and chunking (i.e., the detection of larger units). Several studies have shown L1 reading fluency skills to affect L2 fluency skills, and L2 reading comprehension to affect L2 reading fluency and vice versa, although the effects so far seem to be rather moderate (see e.g., Grabe, 2010: 73). Stages in L2 reading seem to parallel those in L1 reading. For example, Verhoeven & Van Leeuwe (2012) extended to L2 reading the Simple View of Reading Framework from Hoover and Gough (1990), which places reading comprehension as a product of word reading (decoding) and listening compre‐ hension. Verhoeven & Van Leeuwe (2012) found that the word decoding skills of L1 and L2 learners (in the long run) seem to develop at more or less equal rates, probably due to the fact that many L2 learners have already acquired essentials of the target language in the domains of phonology and orthography. The same seems to apply to reading fluency: for example, O’Brien & Wallot (2016) reported that the transition from fluent reading of words to texts emerges roughly at the same time for the L1 and the L2, namely in grade 4, i.e., around age 10. Following Frisch (2013), there are currently three models to account for the relationship between L1 and L2 reading, namely the Interdependence Hypothesis, the Threshold Hypothesis and the Transfer Hypothesis. According to the Interde‐ pendence Hypothesis (e.g., Cummins, 1991), the development of L2 reading is affected by the cognitive and general linguistic skills in the L1, and in particular by L1 reading skills. Many studies have shown that cognitive, linguistic and reading skills in the L1 predict L2 reading skills (e.g., Van der Leij, Bekebrede & Kotterink, 2010) and that there is a high correlation between L1 and L2 reading skills (e.g., Skutnabb-Kangas & Toukomaa, 1976). The Threshold Hypothesis 18 2. Literature review posits that “some minimal threshold of proficiency in the L2 must be attained in order for the reader’s first language reading skills to transfer to reading in the second language” (Carroll & Grabe, 2010: 223). Thus, a certain level of L2 proficiency is necessary before L1 reading strategies and skills can be utilized effectively in L2 reading. An important prerequisite is a learning context that promotes overall L2 language proficiency, at least for lower-level students. So far, however, empirical data are scarce regarding the nature of such a threshold. Finally, language proximity and language distance of the two involved systems are important components of the Transfer Hypothesis (e.g., Koda, 2008: Transfer Facilitation Model) to account for cross-linguistic transfer in reading. In Koda’s view, transfer is “an automated activation of well-established first-language competencies, which is triggered by second-language input” (Koda, 2008: 78). Transfer from L1 to L2 may also occur cross-modally (e.g., knowledge of L1 spoken language may be transferred to L2 reading development), for example, when L1 phonological awareness predicts L2 word recognition. Thus, transfer not only includes a set of rules but also any form-function relationship that the L2 users have acquired in their L1, which may be utilized in their L2 as well. 2.1.3 L1 writing Writing is a complex task which requires the coordination of fine motor skills and cognitive skills and reflects the social and cultural patterns of the writer’s time. Dyson & Freedman (1991: 58) define writing as “a process-oriented, goal-directed and problem-solving process, which involves the writer’s aware‐ ness of the composing process and the teacher’s or peer’s intervention at any time needed”. Writing relies on many of the same structures as speech within a language system, such as vocabulary, grammar and semantics, with the added dependency on a system of signs or symbols, often (as in the case of German and English) in the form of a formal alphabet. The writing process has been characterized in several cognitive models as a complex process with several feedback loops (e.g., Grabe & Kaplan, 1996). In simple terms, the text is planned, formulated and revised several times in accordance with the writing goal. A monitor controls for language (i.e., for linguistic-stylistic means) and content. Finally, the writer’s environment affects this process, in particular with respect to the writing task, the audience and the writer’s personal requirements (e.g., Thaler, 2012). Writing thus depends to a large extent on reading skills, because the process of composing a text always involves reading and rereading of the unfolding text (e.g., Myhill & Fisher, 2010). 19 2.1 Definitions Other fundamentals of successful writing include a positive attitude towards writing, topic knowledge, the ability to rapidly access lexical items, knowledge of text features, writing strategies, cognitive skills, visual-spatial performance and, in addition, spelling skills and strategies (see Myhill & Fisher, 2010). During their development, children progress from the level of copying familiar structures to recognizing and reproducing text structures. With respect to the elementary school context, evaluations of writing proficiency usually rely on components such as spelling, punctuation, vocabulary and grammar (e.g., Rathvon, 2004). There are many models which illustrate the development of L1 writing. The classical model by Hayes & Flower (1980) describes writing as consisting of three primary processes—planning, translating (i.e., the production of text), and reviewing—all operating under executive control within the constraints of the external task environment and the writer’s long term memory. Although the model has been revisited and revised over the years (Hayes, 1996, 2006), it has retained its cognitive character as well as its influence on the field. Berninger & Swanson (1994) revised the Hayes & Flower (1980) model to highlight the challenges that young or immature writers come to face during writing. According to the authors, two sub-processes at the translating level are particularly challenging for developing writers, namely transcription (e.g., handwriting and spelling) and text generation at the word, sentence and discourse level. As long as these processes (plus higher order executive processes such as planning and revising) are not automated, they compete for limited working memory resources during writing, and such competition for resources affects young writers’ products (e.g., De La Paz & McCutchen, 2016). Since spelling has been mentioned to constitute an important part in assessing young learners’ writing, one of the many stage models for spelling development (e.g., Bear & Templeton, 1998; Gentry, 2000) is presented in the following: Peregoy & Boyle (2013: 208 ff.) differentiate four developmental stages or levels, namely the prephonetic, phonetic, transitional, and conventional level. Prephonetic spelling consists of letters or letter-like forms which do not yet represent speech sounds, i.e., children do not yet understand the alphabetic principle that a letter or letter sequence represents a speech sound. Next, children move through the phonetic spelling level, which is characterized by one-to-one correspondence between sounds and letters. Transitional spellers then extend their knowledge beyond the phonetic aspects of spelling and begin to include conventional spellings that are not strictly phonetic and may be spelling 60 to 90 % of words correctly in their writing. Finally, at the conventional spelling level, most words are spelled conventionally and correctly. 20 2. Literature review 2.1.4 L2 writing Similarly to L2 reading, L2 writing differs from L1 writing. Learners have been found to be less efficient and slower in L2 than L1 writing (e.g., Cook & Bassetti, 2005). The reasons for these differences are manifold, ranging from differences in the writing systems of the L1 and the L2, to comprehension and memorization problems due to insufficient L2 word decoding. L2 texts are also often shorter and include more errors than L1 texts. Many of these errors are due to L1 transfer, which may be orthographic, grammatical, lexical or discourse-related (e.g., Leki, Cumming & Silva, 2008). Among others, predictors of successful L2 writing include L2 proficiency, the type of orthographies involved and (as in L1 writing) phonological awareness, rapid naming, and verbal memory (e.g., Geva, 2006). Other, less language-related factors relate to topic familiarity and mode, i.e., the type of writing involved (e.g., Hussein & Mohammad, 2011). Regarding the development of L2 writing, Peregoy & Boyle (2013: 210) point out that for young learners, it is very similar to L1 writing (with the prephonetic stage being skipped). In other words, children’s L2 writing also progresses from words and sentences to texts. There are, however, restrictions in children’s L2 text production due to their limited knowledge of L2 grammar, vocabulary and idiomatic expressions. Furthermore, L1 spelling strategies will transfer partially to L2 writing, relating to, for example, the capitalization of letters (Odlin, 1989, see also Steinlen & Piske, 2020). It has been emphasized by many authors that the more children read or are read to in the L2, the easier it will be for them to write (e.g., Krashen, 1982; Peregoy & Boyle, 2013). In the past, the analysis of L2 writing has often been restricted to the analysis of accuracy, whereas nowadays evaluations of elementary schoolers’ L2 texts mainly focus on the same components as that of their L1 texts, namely spelling, punctuation, vocabulary and grammar, which constitute the main sources of transfer (e.g., Pinter, 2006). For example, young German learners of English often use German spelling for English words (e.g., *<häpi> instead of <happy> and German words and German syntax in English sentences as in *<I have im October birthday> (e.g., BIG-Studie, 2015; Burmeister, 2010; Rymarczyk, 2010, 2016). Such interferences exist particularly for writers with low levels of proficiency in the L2, who often rely heavily on their first language resources (see more on foreign language spelling in the elementary school classroom in chapter 7.7). 21 2.1 Definitions 2.1.5 Some problems relating to phoneme-grapheme correspondences in German and English English and German belong to the West Germanic language family and share many similarities regarding phonology, spelling, morphology, lexis, syntax and the alphabet as their writing system, and are, therefore, characterized as typo‐ logically similar languages (e.g., Genesee & Jared, 2008; Frisch, 2013). However, contrastive analyses also recognize many differences in these areas (see König & Gast, 2009 for a detailed description of these linguistic differences). The English spelling system, therefore, poses many problems for beginning learners whose L1 is German (e.g., Frisch, 2013). For example, the German writing system is rather transparent (shallow) with relatively regular phoneme-grapheme correspondences (following the Orthographic Depth Hypothesis, e.g., Frost, 2005). In German, the letter combination <sch> usually corresponds to the sound / ʃ/ (e.g., Schuh, Tisch, Asche). English, though, exhibits many irregular phoneme-grapheme correspondences and it is, therefore, considered a deep (opaque) writing system. For example, the sound / ʃ/ may be spelt <sh> (shoe), <ti> (nation), <ch> (machine), <s> (sure), <ss> (issue), <ci> (social) or <ce> (ocean), to name just a few (e.g., Eckert & Barry, 2002). The different orthographic characteristics of German and English often lead to interlingual transfer in the spelling performance of students learning English. In particular, the deep orthography of the English language also influences the less complex German written language, and transfer effects are assumed (and found) because students apply specific spelling strategies of their L1 to the target language. Such strategies may include the phonemic route to spelling (where the L2 words are spelt like they sound, e.g. *<schur> instead of <sure>), the visual route to spelling (where the L2 words, particularly familiar and high-frequency ones, are retrieved as a whole, using visual imagery of the word, e.g. <the>) and the use of metalinguistic knowledge, which includes knowledge about the L2 and its irregularities and regularities, for example when two words share the same root but are pronounced differently (e.g., <sign> and <signal>, e.g., Frisch, 2013; James & Klein, 1994; Reichart Wallrabenstein, 2004). In sum, this chapter provided information with respect to the development of reading and writing skills in the L1 and in the L2. The examples above also point to the complex (and also crosslinguistic) relationship between reading and writing (see e.g., Schoonen, 2019 for a review), which will certainly affect the outcomes of any (foreign) language literacy test, including the ones presented here. 22 2. Literature review 2.2 Regular elementary school English programs in Germany In the school year 2019/ 20, over 2.9 million students attended one of the 15,431 (private or public) elementary schools in Germany (Statistisches Bundesamt, 2020b). Since 2004 foreign language (FL) teaching has been compulsory in elementary schools in Germany; however, there are differences across the Federal States in terms of the beginning year and the number of weekly lessons in the foreign language (e.g., Fleckenstein, Möller & Baumert, 2020; Hempel, Kötter & Rymarczyk, 2018). For example, the first foreign language is introduced in Baden-Württemberg in Year 1 (until 2019) but in Berlin-Brandenburg and Schleswig-Holstein in Year 3. In most schools, the FL is taught between one and two hours a week (e.g., Börner, Böttger, Kierepka & Lohmann, 2017; Fleckenstein et al., 2020; Hempel et al., 2018). In the following, elementary schools in which a FL (such as English) is taught as a separate subject are referred to as schools with a regular / mainstream / traditional / conventional FL program. Traditional FL teaching at school is characterized by a course-oriented approach, in which the foreign language itself is the subject of study. The primary goal of FL lessons is to master the new language as faultlessly as possible. Foreign language skills (including vocabulary and grammar) are usually taught and practiced with the help of textbooks and other materials, which present selected subject matters (Wolff, 1997). Thus, the acquisition of the FL does not happen naturally but through prefabricated and systematized foreign language input (e.g., Burmeister, 2006; Möller, Hohenstein, Fleckenstein, Köller & Baumert, 2017b). According to Kolb (2012: 32), there are five objectives of EFL (English as a for‐ eign language) teaching at elementary schools. These include attitudes towards language learning (to develop a “positive mind-set” (Schmid-Schönbein 2008: 37) for foreign language learning in general); intercultural learning (to promote an open and tolerant attitude towards other languages and cultures and prepare the children for intercultural encounters); language and cultural awareness (to raise children’s sensitivity for differences among languages and cultures); language learning competence (regarding, for example, learning strategies and techniques); and language competencies, i.e., basic communicative competence in the foreign language (relating to the five skills of listening, speaking, reading, writing and mediation). The language competencies acquired at elementary level correspond to the level A1 of the Common European Framework of Reference (Council of Europe, 2001, 2018) with a focus on listening and speaking (see Chapter 2.2.2). They are assessed in familiar communicative situations of language use in order to increase children’s self-confidence, employing 23 2.2 Regular elementary school English programs in Germany assessment tools such as pen and paper tests as well as observation sheets and language portfolios (e.g., Kolb, 2012). 2.2.1 Controversial views on EFL reading and writing in elementary school For many years, there has been a controversy regarding the introduction of FL reading and writing in elementary school. FL literacy skills played only a minor role in the curricula of the Federal States of Germany, and, consequently, literacy activities in the target language were often neglected (or not carried out at all) in elementary school FL classrooms. Critical views which have been put forward on introducing FL reading and writing during the elementary school years include the following arguments (see e.g., Doyé, 2008; Frisch, 2013; Legutke, Müller-Hartmann, Schocker-v. Ditfurth, 2009; Piske, 2010, 2017a, b; Treutlein, Landerl & Schöler, 2013 for more details): (1) Given the already limited time of exposure, it is irresponsible to familiarize learners with English spelling. The focus in elementary EFL is, therefore, on developing oral and not written communicative competences (i.e., the primacy of oral skills). (2) Due to the irregular grapheme-phoneme relationships in English, EFL reading and writing tasks put excessive demands on learners (particularly on weak ones), who already have to cope with literacy development in their first language or in the majority language. (3) Competences in FL reading and writing may be misused as an additional criterion to select learners for transition to different secondary school types (i.e. Haupt-/ Mittelschule, Realschule, Gymnasium). (4) Due to interferences, the introduction of the English writing system may negatively affect the acquisition of written (and oral) German. Many of the arguments listed above have been invalidated in a number of studies (see Piske, 2010 or Frisch, 2013, for a review). Nowadays, there is a strong tendency to offer a greater number of reading and writing activities the EFL classroom (see e.g., Burwitz-Melzer, 2010; Diehr & Rymarczyk 2010; Frisch, 2013; Legutke et al., 2009; Piske, 2010, 2017a, b; Rymarczyk, 2016): (1) The learners want to write in English and do so independently and quite naturally, without their teacher asking or supporting them to do so, for the same purposes as they use this skill in their L1. 24 2. Literature review (2) If prevented from writing, learners may get used to wrong spellings, which may be difficult to change (fossilization). Therefore, it is better not to withhold the correct spelling of words. (3) Not only are learners interested in writing, but they are also able to identify, understand and become aware of the different principles on which different writing systems are based, provided the explanations are age appropriate. (4) Integrating reading and writing supports learning, i.e., it appears to be easier for some L2 learners to identify word boundaries or to remember words or phrases if they have seen them in their orthographic form. (5) Writing activities increase the diversity of teaching methods in the EFL classroom. (6) Children will start to develop an awareness for individual sound segments such as vowels and consonants only when they learn to read and write alphabetic symbols. The controversy about whether to introduce English reading and writing in elementary school seems to have subsided in recent years. Nowadays the focus is rather on the question of how to introduce the English writing system. However, there is a dire need for studies relating different approaches to outcomes in L2 reading and writing. For the German context, Frisch (2013) examined effects of different methodological approaches (i.e., the Phonics and the Whole Word Method) on L2 reading, which is discussed in greater detail in chapter 7.1. 2.2.2 Curricula for English-as-a-subject relating to reading and writing As noted above, FL reading and writing used to play only a minor role in the elementary school curricula of the Federal States of Germany. Nowadays, though, the target level for functional communicative competences (including FL reading and writing) at the end of grade 4 is generally aimed at level A1 (Standing Conference of the Ministers of Education and Cultural Affairs, Kultus‐ ministerkonferenz, KMK, 2013a). This competence level describes an elementary use of language, i.e., the students can communicate in a simple manner, and understand and use everyday expressions and very simple sentences (cf. Council of Europe, 2001, 2018). With respect to literacy skills, level A1 students • can understand very short, simple texts a single phrase at a time, can pick up familiar names, words and basic phrases and can reread as required (Council of Europe, 2001: 68), 25 2.2 Regular elementary school English programs in Germany 2 „Neben Erfahrungen mit der mündlichen Ausdrucksform treten dabei in Klassen 3 und 4 erste Erfahrungen mit der Schriftlichkeit“ (Ministerium Baden-Württemberg für Kultus, Jugend und Sport, 2004: 68). In addition to experiences with oral expression, first experiences with writing will take place in grades 3 and 4 (own translation). 3 „Das Lesen und Schreiben ist somit dem Hören und Sprechen untergeordnet“ (Minis‐ terium für Bildung, Wissenschaft, Forschung und Kultur des Landes Schleswig-Holstein 2007: 4). Reading and writing is thus subordinate to listening and speaking (own translation). • can recognize familiar names, words and very basic phrases on simple notices in the most common everyday situations (ibid.: 70), • can write a short simple postcard (ibid.: 83), and • can copy familiar words and short phrases e.g., simple signs or instruc‐ tions, names of everyday objects, names of shops and set phrases used regularly (ibid.: 84). The situation of elementary school English in the three Federal States of Baden-Württemberg, Schleswig-Holstein and Berlin-Brandenburg, where the elementary schools in the present studies are located, may be characterized as follows: In Baden-Württemberg, English-as-a-subject was offered from grade 1 onwards (until 2019). According to the curriculum (2004) for English (which is relevant for the collection of data for the present study), reading and writing activities did not start before grade 3 2 , with the focus on oral skills in the first two grades; and competence levels were not specified (Ministerium Baden-Württemberg für Kultus, Jugend und Sport, 2004). From 2016 to 2019, English reading and writing activities were also included in grade 1 and 2; and level A1 was to be attaineded at the end of grade 4 (Ministerium Ba‐ den-Württemberg für Kultus, Jugend und Sport, 2016a). Since 2019, elementary schools in Baden-Württemberg offer English-as-a-subject only in grades 3 and 4 (Ministerium Baden-Württemberg für Kultus, Jugend und Sport, 2020). In Schleswig-Holstein, English-as-a-subject is introduced in grade 3, and reading and writing are subordinate to listening and speaking 3 ; level A1 is expected at the end of grade 4 (Ministerium für Bildung, Wissenschaft, Forschung und Kultur des Landes Schleswig-Holstein, 2007). Finally, English-as-a-subject is introduced in grade 3 in Berlin-Brandenburg. FL competence levels are not specified for elementary schools but level A2 should be attained in grade 8 (Senatsverwaltung Berlin für Bildung, Jugend und Sport, 2006). Depending on their level, students in Berlin at the end of grade 4 can read, copy and write very short English words or texts that they are familiar with, similar to level A1. 26 2. Literature review 2.2.3 English textbooks The use of literacy activities in the elementary school FL classroom has been examined by Hempel (2016). Based on interviews with almost 150 English elementary school teachers in Germany, she found that 98 % of these teachers used a textbook, usually Playway, Ginger, or Bumblebee. An analysis of these textbooks revealed that the activities in these books are mainly geared towards speaking (55 %) and listening (26 %); reading and writing activities are rather underrepresented (1 % and 9 %). This parallels the role of literacy activities in the different curricula of the Federal States of Germany for the subject English, in which reading and particularly writing still play only a minor role. In her analysis of English textbooks for elementary schools, Burwitz-Melzer (2010: 110) pointed out that textbooks nowadays include a greater number of tasked-based reading and writing activities than older versions. However, in her view, textbooks still do not sufficiently explain how to teach reading and writing; and they neither offer any literacy methodology based on any learning theory, nor any systematic strategies for reading and writing, which may support teaching and learning. Although certain levels of competence should be reached at the end of grade 4 (namely A1), textbooks do not provide sufficient guidelines as to how this goal could be reached, particularly with respect to the heterogeneous FL level in class. According to Burwitz-Melzer (2010) there is still considerable room for improvement regarding literacy instruction in English textbooks and in the respective guidelines for teachers of English as a subject. 2.2.4 Studies on FL reading in regular EFL programs In empirical research, an increasing number of studies have examined the influence of written language on foreign language learning in elementary schools in Germany, but the focus has been more on reading (comprehension) than on writing. These studies have been largely conducted in only one Federal State of Germany (e.g., EVENING for Nordrhine-Westfalia, or Rymarczyk, 2011; Steinlen & Piske, 2018 for Baden-Württemberg), with the exception of the BIG-Kreis (2015), which assessed students’ FL competence in different Federal States. Unfortunately, many findings were not related to the Common European Framework of Reference (CEFR, Council of Europe, 2001, 2018) and/ or statistically analysed. The EVENING study (Evaluation Englisch in der Grundschule) was conduc‐ ted with around 1,300 fourth grade classes in Nordrhine-Westfalia (NRW, Engel, Groot-Wilken & Thürmann, 2009), who had attended regular FL lessons since the beginning of grade 3. The reading test consisted of 25 multiple-choice and 27 2.2 Regular elementary school English programs in Germany matching tasks, which were specifically designed for the project, and assessed reading skills with regard to the children’s world knowledge, lexical knowledge and knowledge of chunks and phrasal verbs. In total, the children solved 66.8 % of all tasks. These results were better than expected, because English in its written form had only played a subordinate role in the curriculum in NRW at that time (2005-2007). This was also reflected in the low amount of reading activities carried out in the 88 hours of EFL lessons (Wilden et al., 2013). Elementary schoolers’ English skills in reading, writing, listening, and speak‐ ing were evaluated in the BIG-study with a sample of 2,148 fourth grade students from different Federal States of Germany (BIG-Kreis, 2015), using test material from the EVENING study. The 17 items on English reading comprehension included multiple choice tasks and tasks on reordering sentences to match pictures. On average, students answered 14 items correctly, corresponding to 82.4 %. However, the authors pointed out that the tasks may have been too easy, because the fourth graders in the BIG-study obtained scores on the reading comprehension tasks that were more than 15 % higher than those scores reported in the EVENING study (Engel et al., 2009). Smaller-scale studies on English reading skills in regular English classrooms in Germany include Rymarczyk (2011) and Steinlen & Piske (2018a). Rymarc‐ zyk (2011) examined English reading comprehension skills of 48 firstand third-graders in the FL classroom, reporting that even weaker learners, when confronted with English writing in the classroom, showed significant learning gains in reading. Due to the incongruent phoneme-grapheme relationship in English, the children in Rymarczyk’s study achieved better results on the silent-reading, word-picture matching task than on the reading-aloud task. Based on these results, Rymarczyk (2011) argued for the inclusion of written language in the first two elementary school years, as doing so may also present additional learning opportunities for weaker students (see also Wilden et al., 2013 and chapter 7.1). Steinlen & Piske (2018a) examined English reading comprehension skills of 73 students at the end of grade 4 who had attended a FL program since first grade. Using the subtests for reading from the Primary School Assessment Kit (PSAK, Little, Simpson & Catibusic, 2003), they found these children to perform in-between levels A1 and A2 regarding their English reading comprehension skills. Thus, the outcome for English reading exceeded the target level of A1 (KMK, 2013a). The authors argued that these results were possibly due to two factors. Apart from the children’s background (i.e., having age-appropriate cognitive abilities and coming from middle-class families), the high quality of the English lessons in the FL program may have been the decisive factor, since the teachers in the 28 2. Literature review FL program hold university degrees in teaching English-as-a-subject rather than teaching “fachfremd” (i.e., without any qualification for this subject which may negatively affect students’ performance, see also Fleckenstein et al., 2020). 2.2.5 Studies on FL writing in regular EFL programs As previously noted, only few studies have examined students’ English text production in the elementary school context in Germany: for example, the BIG-study (BIG-Kreis, 2015) also tested the 2,148 fourth graders on their English writing skills using tasks particularly designed for this study. The four writing tasks included copying words, writing single words to label pictures and filling out a gap text with English words and sentences where the children provided information about themselves. As expected, the children’s scores related to matching and copying single words and providing an English word to label a picture were highest; and between 54 % to 71 % of these tasks were solved accurately (or comprehensibly). Half of the children managed to provide adequate information in the gap text (the so-called “semi-creative task”) despite the fact that text production is usually not trained in the FL elementary school classroom. Errors in these gap texts pertained to the omission of words, incorrect plural endings or subject-verb agreement. Common misspellings in all tasks included the words <pencil> and <cheese>, which were words that the teachers had used in listening and speaking but not in writing activities. Omissions were noted for the words <bathroom, underground, scarf, Wednesday> as *<bathoom, undeground, scaf>; and transpositions for the words <white, blue, clothes, shoes> as *<wihte, bleu, colthes, shoes>. The authors concluded that the children form hypotheses and construct rules with reference to existing experience and knowledge, many based on German, as in *<shört, jelo, Freidi> instead of <shirt, yellow, Friday>. In sum, students’ English writing skills were of much lower quality than their reading skills. This may be attributable to the fact that, according to teacher interviews, writing activities were carried out only in half of their classes on a regular basis, and that almost 20 % of the teachers never let their students write in English. Steinlen & Piske (2018a) also examined English writing skills of the 73 fourth graders in the FL program. The results of the subtest for reading (PSAK, Little, Simpson & Catibusic, 2003) showed that these children generally performed at level A1 with respect to their English writing skills. This outcome is in line with the expectations by the Ministerium Baden-Württemberg für Kultus, Jugend und Sport (2004, 2016a). 29 2.2 Regular elementary school English programs in Germany As this short literature review has shown, there is a need to examine the development of FL literacy skills in regular programs, in particular because the Federal States Baden-Württemberg and North Rhine-Westfalia have recently decided to postpone the beginning of FL classes from grade 1 to grade 3 (e.g., Ministerium Baden-Württemberg für Kultus, Jugend und Sport, 2020). There have been two main reasons for this decision. First, a growing number of elementary school children show deficits in German reading, writing and arithmetic on various achievement tests. It is argued that by delaying the FL lessons, the schools gain two lessons per class level which they can use to better support these students. Second, some recent studies have indicated that in the long run (e.g., in grade 7 or 9), younger learners who started English-as-a-subject in grade 1 showed neither mediumnor long-term benefits over older learners who started in grade 3. As Fleckenstein et al. (2020) pointed out in their review, such results may also be due to the quality of FL teaching in elementary schools (with the current focus on implicit learning and not on FL literacy input); the CEFR level A1 being too low a goal; and the transition to secondary education having a strong focus on explicit learning, deductive teaching of vocabulary, grammar and metalinguistic abilities, as well as on a greater number of assessments of FL reading and writing skills. This is also one of the reasons why, in all of the studies reviewed above, the authors concluded that it is overdue to introduce systematic literacy instruction in the regular FL classroom in Germany (including, for example, an approach to teach English spelling, see also e.g., Börner et al., 2017; Burwitz-Melzer, 2010; Hempel et al., 2018). With such standards, more teachers would hopefully feel obliged to offer writing activities, which in turn would also give them more possibilities for diagnostics and differentiation in class. Students would then receive additional support for autonomous learning and more diverse opportunities to become aware of the structures of the English language, in particular with respect to FL phoneme-grapheme correspondences and the segmentation of the speech stream (BIG-Kreis, 2015, see also chapter 7). Unfortunately, the Ministries of Education of the sixteen Federal States of Germany have not yet been able to agree on a common procedure (e.g., Börner et al., 2017; Burwitz-Melzer, 2010; Hempel et al., 2018). 2.3 Bilingual programs in Germany Bilingual programs represent a more intensive way to acquire a new language within the school context. According to Wolff & Sudhoff (2015: 9), bilingual 30 2. Literature review teaching can generally be defined as a methodological concept in which learning school content is carried out in a language that does not correspond to the school language (usually the majority language). Of the 15,409 elementary schools in Germany, more than 2 % are currently offering a bilingual program, corresponding to more than 330 elementary schools. In these schools at least one content subject (e.g., science, music, physical education, math) is taught in the new language. The target language in German elementary schools is usually English or French (FMKS, 2014). The general aim of bilingual programs is to foster both FL learning and content learning (e.g., Coyle, Hood & Marsh, 2010). The Eurydice report (2006: 22) points to four additional objectives. Following Elsner & Keßler (2013), these include socio-economic objectives (to prepare students for life in a more internationalized society and to offer them better job prospects on the labor market); socio-cultural objectives (to teach students values of tolerance and respect vis-a-vis other cultures through the use of the target language); linguistic objectives (to enable students to develop language skills which emphasize effec‐ tive communication and to motivate them to learn languages by using them for real practical purposes); and finally educational objectives (i.e., subject-related knowledge and learning ability, to stimulate the assimilation of subject matter by means of a different and innovative approach). Next, chapters 2.3.1-2.3.4 describe general characteristics of bilingual pro‐ grams, and chapters 2.3.5-2.3.8 summarize curricular baselines for bilingual education, teacher education, supply of materials and various aspects of assess‐ ment in bilingual programs in Germany. Finally, the results of empirical studies on reading and writing skills in bilingual programs in Germany are provided in chapters 2.3.9-2.3.12. 2.3.1 Differences between bilingual programs and EFL programs The main differences between bilingual programs and regular FL lessons was outlined by Burmeister (2006), who compared regular FL teaching with intensive bilingual (immersion, IM) teaching in the elementary school context. In bilingual lessons, subject matter is taught using a foreign language. In regular FL lessons, the reverse is true: here, foreign language skills are taught with the help of subject matter. One could argue that the teaching materials for regular English lessons in elementary education also provide a variety of topics, e.g., units on animals, on the human body, or topics relating to cultural studies. The crucial difference between bilingual and regular FL programs, however, is based on the 31 2.3 Bilingual programs in Germany objective and thus on the respective function of language and subject matter as the subject of learning. Burmeister (2006) provided the following example: in regular FL teaching, the topic “animals” may have been chosen because it reflects the interests of elementary school children and is therefore motivating. Additionally, the topic may already constitute the following unit in the textbook and/ or perhaps it may just fit the topic taught in science (which is conducted in the majority language, in this case: German). However, the main reason for choosing this topic is that it provides linguistic input to train listening, speaking, reading or possibly writing skills in the FL. The topic thus acts as a vehicle for deliberately exercising and circulating selected linguistic structures and vocabulary. In contrast, the topic of “animals” would be chosen in the bilingual classroom because it is included as a topic in the curriculum for the subject “science” for grade 1. This topic, therefore, is genuinely based on the curriculum of a non-language subject, and the learners’ involvement with its content in the FL is authentic. The L2 acts as a vehicle to transport content; and the selection of linguistic means arises directly from the respective topic (Burmeister, 2006). In addition, dealing with topics in the non-language subject area goes beyond what is usually provided in regular FL teaching in terms of complexity and methodology. To illustrate the linguistic and cognitive challenges that the FL learners in the IM classroom face, Burmeister (2006) points to the curricular requirements for the subject “science” (which is taught in German in mainstream classes) for elementary schools in Schleswig-Holstein: the lessons in the subject “science” are aimed at developing “real life knowledge” and the initiation of “functional classification systems”, and the children at the end of grade 4 should be able “to draw conclusions”, “to think comprehensively”, “to jointly plan, execute and evaluate”, “to conduct research independently”, “to develop time concepts” and “to recognize and evaluate”, to mention just a few key compe‐ tences (e.g., Ministerium für Bildung, Wissenschaft, Forschung und Kultur des Landes Schleswig-Holstein, 1997: 93). In the IM classroom, the children learn and apply these complex skills in the FL (and not in the majority language with which most of the children are more acquainted). Thus, bilingual teaching is not just about vocabulary learning but about conceptual learning, and combines learning concepts and their linguistic expressions (e.g., Bonnet, Breidbach & Hallet, 2003). In sum, in language-driven (i.e., regular EFL) programs, content is used to learn the FL, i.e., FL learning is the priority, content learning is inciden‐ tal, the language objectives are determined by the FL curriculum (i.e., Eng‐ lish-as-a-subject), and students are evaluated on language skills/ proficiency. In 32 2. Literature review content-driven (bilingual) programs, the focus is on content being taught in the FL, i.e., content learning has priority, FL learning is secondary. Content objectives are determined by the curriculum of the specific subject (e.g., science), the teachers must select language objectives, and students are evaluated on content mastery (see Met, 1999: no page). 2.3.2 Different bilingual programs: CLIL vs. IM Throughout Europe, the umbrella term Content and Language Integrated Learn‐ ing or CLIL is used to refer to the educational option of teaching non-language subjects through a second language (L2). More specifically, this term pertains to “all types of provision in which a second language (a foreign, regional or minority language and/ or another official state language) is used to teach certain subjects in the curriculum other than the language lessons themselves” (Eury‐ dice, 2006: 8). In the Anglo-American context, however, the term “immersion” (IM) is being used. Both concepts, CLIL and IM, have in common that subject teaching takes place in a language that does not correspond to the students’ ambient language (i.e., English instead of German in Germany). The following core features are also shared by CLIL and IM programs (e.g., Nikula & Mård-Miettinen, 2014). First, both share the conviction that foreign/ second language competence should not be regarded as a separate skill but one intertwined with students’ cognitive, conceptual and social devel‐ opment, best supported by engaging students in meaningful and cognitively and academically challenging language use. Second, CLIL and IM subjects are timetabled as content and not as language lessons, which also means that teachers are typically content rather than language teachers. Third, CLIL and IM subjects complement foreign language instruction rather than serving as its replacement. However, Nikula & Mård-Miettinen (2014: 2) point out that that demarcating CLIL and IM may simply result in “dead ends” because CLIL and IM as terms have fuzzy boundaries, both when used in academic discourse and as everyday concepts. Nevertheless, IM and CLIL show many differences (see Nikula & Mård-Miettinen, 2014 for a detailed description). First, IM and CLIL differ according to the geographical and sociopolitical context and their dates of origin (starting in Canada in the 1960s and in Europe in the 1990s, respectively). Second, the new language often is a national language in immersion programs (e.g., French in Canada) but a transnational lingua franca in CLIL programs in Europe (e.g., English in Germany). For this reason, CLIL teachers are more commonly non-native speakers of the instructional language than immersion 33 2.3 Bilingual programs in Germany teachers, i.e., lingua franca users of English themselves. Third, the introduction of CLIL takes place more often in secondary levels, while it is commonplace for IM education to begin at kindergarten or elementary levels (see also Burmeister, 2013; Cenoz, Genesee & Gorter, 2014; Genesee, 1987). Finally, CLIL and IM programs may differ in their intensity: CLIL programs in Europe are often characterized by teaching only one or two subjects in the target language (e.g., history, geography or science), corresponding to 10-30 % of the teaching time. In Canada, however, IM programs may only be labeled as such when more than 50 % of the teaching time is conducted in the new language. This difference is illustrated below in chapter 2.3.3. 2.3.3 Intensity of different bilingual programs As mentioned above, bilingual programs (in Germany and elsewhere) differ according to their intensity, i.e., the quantity of FL input. Kersten (2019) provided an overview of content-based approaches in FL teaching, ranging from low to high intensity: Figure 1: Continuum of FL intensity in content-based approaches (adopted from Kersten 2019: 40). In Kersten’s (2019) view, content can - and also should - be part of any regular FL teaching, although in such programs the focus is not on content, but on the FL. To reiterate, bilingual programs (particularly in Europe) are conceived under the generic term “CLIL”, which encompasses all forms of teaching one or several subjects in a FL (e.g., Coyle et al., 2010; Mehisto, Marsh & Frigols, 2008; KMK, 2013). Less-intensive bilingual programs include individual bilingual modules or projects, where, for example, the topic “water” is taught in the FL over a limited period of time, usually lasting only a few days or weeks. “Bilingualer Sachfachuntericht” (or CLIL proper), which in Germany is most widely taught from secondary level 1 onwards, usually offers one to two subjects in the foreign language (for example history, geography), often lasting for 34 2. Literature review 4 In one-way immersion programs, the target language is either a lingua franca (e.g. English in Germany) or the students’ ambient language (e.g. French in Canada), which is chosen because it has high prestige or is a neighbouring language. The students often learn the alphabet in both languages (often first in L1, then in L2). In two-way immersion programs, the focus is on class composition, where students with two different L1s are taught together in the classroom in their respective L1s in some subjects and in their respective L2s in other subjects, so that the language of instruction is L1 for one half of the class and L2 for the other half (e.g., Möller, Fleckenstein, Hohenstein, Preusler, Paulick & Baumert, 2018). 5 Very briefly, one of the most prominent total IM programs (i.e., a one-way or French IM program) was first established in the 1960s when parents of English-speaking children living in a francophone community in Canada raised concerns about their children’s lack of ability to communicate in French (e.g., Lambert & Tucker, 1972). To meet this need, parents, teachers, and school principals developed a unique foreign language acquisition program that was distinguishable from other regular FL programs by the following characteristics: (a) the L2 was the medium of instruction, (b) the IM curriculum paralleled the local first language (e.g., French) curriculum, (c) overt support for the L1 existed, (d) the IM program aimed for additive bilingualism (cf. Lambert, 1975), (e) exposure to the L2 was largely confined to the classroom, (f) students entered with similar (and limited) levels of L2 proficiency, (g) teachers were bilingual, and (h) the classroom culture was that of the local L1 community ( Johnson & Swain, 1997: 6, see also Deventer, Machts, Gebauer & Möller, i. pr.). one year or several years. The most intensive forms of bilingual teaching are immersion (IM) programs. In partial IM programs, at least 50 % of the curriculum is taught in the FL (this also applies to two-way or dual IM programs 4 ). In full (or total) IM programs, teaching takes place 100 % in the FL and over a longer period, i.e., for many years. In Canada, total IM programs are often provided for young English learners in French IM programs, which turn into partial IM programs in later years in order to provide additional teaching in the students’ L1 English (e.g., Genesee, 1987) 5 . As a result, IM programs are the most intensive forms of bilingual education in the continuum of programs varying in FL intensity (e.g., Burmeister & Massler, 2010; Kersten, 2019; Mehisto et al., 2008, as illustrated in Figure 1) and have been found to be particularly effective in terms of FL competence, without neglecting subject competence (e.g., Genesee 1987, 2004; Pérez-Cañado, 2012; Wode 1995, 2009; Wesche, 2002, and also see below). In Germany, however, total IM programs (where the target language is used exclusively in all lessons and subjects in school) are not possible because, according to the Standing Conference of the Ministers of Education and Cultural Affairs (KMK, 2013b), the subject German must be taught in German. That means that in elementary schools in Germany, only a maximum of 70-80 % of the teaching time can be conducted in the new language, corresponding to partial IM programs (for a further review, see Piske, 2015). 35 2.3 Bilingual programs in Germany 6 In 50/ 50 programs (partial IM programs), 50 % of the teaching time is conducted in the FL (e.g. English) and 50 % in the school language (e.g. German in Germany). Similarly, 20 % of the lessons are taught in the FL and 80 % in the target language in 20/ 80 programs. In her review of IM programs, Wesche (2002) posited the question as to how much FL exposure would be needed to ensure adequate language development for learners in order to be able to maintain grade level learning in their other academic studies. She showed that in 50/ 50 IM programs with an early start (i.e., in grade 1 or kindergarten), a second language can be effectively taught, but that gains vis-a-vis early ‘total’ IM are lower, or roughly commensurate with the relative total time spent in the FL (Wesche, 2002). However, due to the restrictions outlined above, any comparisons involving total IM programs are not possible in Germany. As Wesche (2002) also pointed out, late-starting, low-intensity bilingual programs (i.e., 20/ 80 programs 6 starting in grade 7) may be quite successful in enhancing the FL proficiency of academically-inclined (Gymnasium) students (for Germany see e.g., Burmeister & Daniel, 2002; Köller, Leucht & Pant, 2012; Nold, Hartig, Hinz & Rossa, 2008; Rumlich, 2016; Wode, 1995, 2009). Such programs may also be successful at the lower secondary school level, as shown in studies on Realschule (Rischawy, 2016) and Hauptschule (Schwab, 2013) in Germany. Thus, bilingual programs can be implemented at every school level without compromising age-appropriate subject knowledge. Evidence for 20/ 80 programs with an early start is, unfortunately, scarce (see also Piske, 2020). Here, the FL is generally offered in only one subject, often science for the elementary school context. The findings of pilot studies indicate improvement in FL learning from grade 1 to 4, although the scores of the language tests did not always differ in a statistically significant way from year to year (e.g., Couve de Murville, Kersten, Maier, Ponto & Weitz, 2016; Steinlen & Gerdes, 2015). Unfortunately, academic achievement was not taken into consideration in these studies. In the following, “bilingual schools”, “bilingual programs” or “CLIL programs” are used as umbrella terms, which subsume bilingual programs with lower or higher intensity. The term “IM” (immersion) is ascribed to partial immersion programs where the FL is used more than 50 % of the teaching time. Abbreviations such as Bili-20, Bili-50 or Bili-70 are used to illustrate the intensity of any bilingual program, in which 20 %, 50 % and 70 %, respectively, are offered in the target language, with Bili-50 and Bili-70 programs corresponding to partial IM programs. 36 2. Literature review 2.3.4 Bilingual programs as “elitist” programs In Germany, bilingual programs are considered to be ‘elitist’ because they often are attended by students with particular personal, intellectual, or familial characteristics. This is partly due to the fact that schools do not always ran‐ domly assign students to a particular program, especially when there are more applicants than places (see e.g., Apsel, 2010; Genesee, 1987; Swain & Lapkin, 1982; Zaunbauer & Möller, 2007). Student-selection factors may include the age-appropriate knowledge of the L1, the ability to concentrate, perseverance, commitment and/ or communication abilities (e.g., Kersten, Fischer, Burmeister, Lommel, Schelletter, Steinlen & Thomas, 2010b). For example, employing a large longitudinal study comparing students in mainstream and IM elementary schools in Germany, Zaunbauer & Möller (2007, see also Zaunbauer & Möller, 2006, 2010; Gebauer, Zaunbauer & Möller, 2012, 2013) reported that in grade 1, IM students outperformed their peers in mainstream programs in a nonverbal intelligence test and pointed out that these cognitive differences may also point to prior selection effects (see also chapter 2.6). Elementary schools may discourage parents of struggling learners (e.g., dys‐ lexic children and children with auditory / perceptual / concentration problems) to attend an IM program, often reasoning that such a program would be too large of a burden for such children (Fischer, 2019). Other schools have opted for an alternative solution: when more students than can be accepted to the bilingual program have applied for admission, these students are randomly assigned to a bilingual or regular school program. For example, many IM schools, particularly in Canada, employ a lottery system, as in the case of the St. Lambert school (e.g., Genesee, 1987). Ideally, then, the IM students would be identical to regular students in all respects, except that they would be enrolled in different programs. Indeed, the results of nonverbal intelligence tests carried out in the St. Lambert school did not show any differences between first graders in the IM and in the regular program (e.g., Lambert & MacNamara, 1969; Lambert & Tucker, 1972; Genesee, 1987). The three schools where data have been drawn for this study deal very differently with the admission of students in their bilingual programs (see chapter 4.1 for more information). The Muhlius Schule in Kiel and the Platanus Schule in Berlin offer only one FL program, and that is a bilingual one. However, the Muhlius Schule is a public school, catering to a relatively high number of students with learning difficulties or associated problems that hinder learning (e.g., Steinlen & Gerdes, 2015), and there are no restrictions on admission. The Platanus Schule in Berlin is a private school, and the children are preselected for cognitive, linguistic and communicative skills (see Steinlen & Piske, 2018c). 37 2.3 Bilingual programs in Germany The Hügelschule in Tübingen, finally, has offered both a regular program and a partial immersion program since 2008/ 09 (e.g., Steinlen, 2016, 2017, 2018a, b; Steinlen & Piske, 2013, 2015, 2016a, b, 2018, 2020; Tamm, 2010). So far, a lottery system had to be employed only once. However, as teachers and the heads of the school point out, preselectional effects cannot be ruled out because the parents usually enroll their child in the bilingual program when they are very confident that s/ he would indeed be able to master it without any difficulties (see also Apsel, 2012; Zydatiß, 2009: 161 for similar impressions). Despite this, the results of a nonverbal intelligence test conducted in grade 1 in this school, with 256 students attending either a regular or a bilingual program, did not indicate any significant differences between the two groups (Yadollahi, Steinlen & Piske, 2020, see also Steinlen & Piske i. pr.). 2.3.5 Curricula for bilingual teaching in Germany In 2013, the Standing Conference of the Ministers of Education and Cultural Affairs (KMK, 2013b) published a report on bilingual instruction in Germany in secondary (but not in elementary) schools. In this context, bilingual instruction is defined similarly to CLIL (Eurydice, 2006), namely that “it refers to the teaching of a current subject other than foreign languages in more than one language” (KMK, 2013b: 3). The report emphasizes that the curriculum of the content subject constitutes the basis for bilingual instruction (KMK, 2013b: 8) and that the learning goal is content knowledge in the content subject in two languages. Furthermore, acquired knowledge and skills in the bilingual content subject should be the same as in the corresponding regular (i.e., monolingual German) content subject (KMK, 2013b: 8). As the only Federal State to do so, the Ministerium Baden-Württemberg für Kultus, Jugend und Sport (2008) published information on bilingual instruction at the elementary school level (“Niveaukonkretisierungen”, supplementing the curriculum for English 2004-2015). Referring explicitly to the subject MeNuK (Fächerverbund Mensch, Natur und Kultur, which combines science and culture) as an appropriate bilingual content subject, English competences for grade 4 are very briefly described in terms of lexical knowledge, world knowledge and writing skills. Unfortunately, these expectations do not deviate much from the standards set for English-as-a-subject lessons (see Ministerium Baden-Würt‐ temberg für Kultus, Jugend und Sport, 2004); that is, the close link between FL and content learning has not been taken into consideration. Administrative reg‐ ulations, approval requirements, scope and duration or teacher qualifications, 38 2. Literature review assessment and certification measures are also not provided by the Ministries of Education for Baden-Württemberg, Berlin or Schleswig-Holstein. 2.3.6 Teacher supply for bilingual programs in Germany In Germany, many teachers in bilingual elementary school programs hold a degree in bilingual teaching (Piske, 2015). According to the KMK (2013b), teachers in such programs (independent of the school form) ideally have obtained a teaching qualification for a content subject (e.g., math, science, geography) and for a modern foreign language (e.g., English, French). Regarding the schools in this study, most teachers at the bilingual program in the Hügelschule in Tübingen (Baden-Württemberg) hold a degree in bilingual teaching (“Europaleh‐ ramt” at the Pädagogische Hochschulen Freiburg or Karlsruhe) or a degree for English-as-a-subject and for another content subject. However, there are problems with teacher supply due to a lack of training courses in many Federal States. Therefore, it is also possible to teach in a bilingual program with a teaching qualification for a subject and a high level of FL competence, which should at least correspond to C1 (according to the CEFR, European Commission, 2001). For example, the teachers in the Muhlius Schule in Kiel (Schleswig-Holstein) have all spent at least three years or more in an English-speaking country, whereas the teachers at the Platanus Schule in Berlin are either native English speakers or have a high competence in English at the C2 level. In order to meet the demand for teachers in bilingual programs, the obvious solution is to employ teachers who are trained with a dual subject qualification, holding, for example, joint degrees in English and science for elementary schools. However, even in such cases, teachers are usually not provided with the specific English terminology for the respective subject (e.g., science or mathematics), and they usually do not have any experience in teaching science in the bilingual classroom. Usually, teachers working in schools with bilingual programs have pragmatically developed a certain teaching method of their own, frequently unsystematic, highly personal in style, and yet, just as frequently, highly successful. Still, the need to increase systematic bilingual-oriented teacher training, specifically aimed at the needs of the bilingual classroom (such as the “Europalehramt”), is more than obvious (e.g., Burmeister & Pasternak, 2004; Fischer, 2019; Kersten et al., 2010b; Möller, Fleckenstein, Hohenstein, Preusler, Paulick, Isabell & Baumert, 2018; Tel2L, 2000). 39 2.3 Bilingual programs in Germany 2.3.7 Supply of materials for bilingual programs Although the KMK (2013b) stated that the Federal States of Germany are responsible for teaching and learning resources for bilingual teaching, it is still a challenge to find appropriate materials for many teachers (e.g., Tel2L, 2000). Although it may be fairly easy to collect authentic teaching materials from the country of the target language, these are in most cases unsuitable for immediate use in the classroom as they are often much too demanding for FL learners in their original language version. Hence the teacher has to adapt the language of authentic materials to the learners’ level of achievement in the FL, which requires extra work. Nevertheless, authentic materials provide the intercultural touch which is also one of the core objectives of bilingual teaching. It therefore remains a delicate balance and is up to the individual judgement of the teacher to assess such authentic materials for use in the bilingual classroom (see also e.g., Böttger & Müller, 2020; Burmeister, 2006; Steiert & Massler, 2011). So far, only a few coursebooks for elementary school bilingual teaching have been published, because bilingual teaching takes place on a small scale in Germany, and because these bilingual programs vary in intensity and cover different subjects. Teachers may then hesitate to start bilingual teaching due to the limited choice of teaching materials available. This situation calls for universities and centers for teacher training to step in to develop and publish, in co-operation with experienced bilingual teachers, materials that provide an appropriate blending of authenticity and classroom needs for any age group (e.g., Böttger & Müller, 2020; Steiert & Massler, 2010; Tel2L, 2000). 2.3.8 Assessment in bilingual programs As noted above, CLIL classes follow the regular curriculum of the content subject, and students’ progress is therefore evaluated on a regular basis (KMK, 2013b). However, there is still a lack of official regulations for the assessment of learners’ attainment in CLIL teaching (e.g., Massler & Steiert, 2011). As the CLIL approach is relatively new in the European context, many teachers are insecure about how to assess CLIL learning. According to Massler & Steiert (2011), CLIL assessment differs from regular assessment in several ways because it needs to account for the goals and objectives of two different subjects, including knowledge, competences, skills, attitudes, and behavior, for both language and content. Particularly for the elementary school context, Steiert & Massler (2011) proposed that any assessment of CLIL learning should take into consideration both components - language and content - in an integrated manner (which may differ according to the intensity of the program). Such assessments ideally include the learners’ developmental stage 40 2. Literature review (as there may be a gap between the learners’ cognitive and linguistic competences) and both product and process (i.e., summative and formative assessment). Finally, it is important to make CLIL assessment transparent with respect to the question of which language will be assessed (i.e. the school language or the FL) and how much the FL will count. In some schools, for example, assessments in CLIL subjects are carried out not in the FL (e.g., English) but in the school language (e.g., German), in order to avoid a confound of FL knowledge and subject knowledge (e.g., Tel2L, 2000). More information on assessment, particularly regarding FL reading and writing skills in the bilingual elementary school context, is provided in chapter 7.8. 2.3.9 Studies on L1 and FL reading and writing in bilingual programs In Germany, bilingual programs have been examined in several large-scale studies focusing on one-way immersion and two-way immersion programs. The MOBI-project (“Monolinguales und bilinguales Lernen”) ran from 2005 to 2009 (e.g., Gebauer et al., 2012, 2013; Kuska, Zaunbauer & Möller, 2010; Zaunbauer & Möller, 2006, 2007, 2010; Zaunbauer, Bonerad & Möller, 2005; Zaunbauer, Gebauer & Möller, 2012, 2013) and involved over 650 elementary school students in Northern Germany attending either bilingual (immersion) programs (where all subjects were taught in English except for German-as-a-subject), i.e. one-way programs, or regular EFL programs (where English was taught as a subject), respectively, from grade 1 onwards. The focus of the MOBI-project was on the development of cognitive skills and language skills, such as German reading and writing, English vocabulary and reading, as well as the development of content knowledge in mathematics and science, which were assessed mainly with standardized tests. In grade 4, the data included around 300 students in immersion programs (e.g. Zaunbauer et al., 2012). The aim of the second project is to evaluate the two-way immersion programs of the “Staatliche Europa-Schulen Berlin” (SESB, e.g., Möller et al., 2017). Teaching takes place in German and in one of the respective partner languages, i.e., English, French, Greek, Italian, Polish, Portuguese, Russian, Spanish, or Turkish, and half of the teaching time is devoted to one of the two languages. The composition of the classes generally (but not exclusively) consists of 50 % of learners with the first language German and of 50 % of learners with the respec‐ tive partner language as L1. The teachers of the subjects taught in the partner language usually speak the partner language as L1. The focus of this project is on the development of reading skills, mathematics and science in German and the partner language. The results are also compared to the development 41 2.3 Bilingual programs in Germany of other students in monolingual programs. Altogether, 789 students in grade 4 participated in this study. In 2015, the Federal State of Bavaria initiated a project called “Lernen in zwei Sprachen - Bilinguale Grundschule Englisch” (Learning in two languages - bilingual elementary school English), with 21 elementary schools located all over Bavaria. The students in grades 1-4 were taught in two languages (i.e. German and English) in the subjects of mathematics, science, art, music and PE, corresponding to less intensive bilingual programs with less than 50 % conducted in the target language English. Instruction in the bilingual classes was based on the competency expectations formulated in the curriculum LehrplanPLUS Primary School. Teaching units or phases in the subjects were conducted in English on suitable topics and occasions. The respective teacher selected topics suitable for the English language, and the focus was on implicit teaching of the English language. Lessons in the bilingual classes were given according to the timetable, i.e. without additional time quotas. The final report included data from 430 fourth graders (Böttger & Müller, 2020). Since 2012, Steinlen and Piske have conducted studies on L1 and L2 language skills in three bilingual elementary schools in Germany. The main focus of these studies has been on the development of German and English by minority and majority language students. Most of the studies took place with students attending a partial immersion program offered by the Hügelschule (Tübingen, Germany). As another cohort at the Hügelschule has been added each year, the sample size increases by publication date, with more recent samples containing cohorts from earlier studies (e.g., Steinlen, 2016, 2017, 2018a, b; Steinlen & Piske, 2013, 2015, 2016a, b, 2018a,b, 2020, i. pr., Yadollahi et al., 2020). Other studies of this project dealt with a less intensive bilingual program at the Muhliusschule in Kiel (Steinlen & Gerdes, 2015; Steinlen, 2018a) and with an intensive partial immersion program at the Platanus Schule in Berlin (Steinlen & Piske, 2018c), where 20 % and 70 % of the teaching time, respectively, was conducted in English. Focusing on reading and writing, the results of these projects will be presented below. 2.3.10 L1 reading and writing in bilingual programs In studies on reading comprehension and reading fluency skills in the majority language German at the end of grade 4, children in elementary schools in Germany with partial IM programs (one-way and two-way) have been compared to those in mainstream programs. No significant differences between the two groups were reported, and the immersion groups achieved age-appropriate values (e.g., Baumert et al., 2017; Gebauer et al. 2012, 2013; Möller et al., 2018; 42 2. Literature review Steinlen, 2016; Steinlen & Piske, 2013, 2015, 2016a, 2018a; Zaunbauer et al., 2005, 2013, Zaunbauer & Möller, 2006, 2007, 2010, see also Baumert et al., 2012 for sixth grade students). Similar results regarding majority language children’s literacy skills in their L1 have been reported for the Canadian French IM context. L1 English children in IM programs generally obtained the same results in L1 English reading tests as L1 English children in non-IM programs (e.g., Genesee, 1978, 2004; Genesee & Jared, 2008; Genesee, Holobow, Lambert & Chartrand, 1989; Geva & Clifton, 1994; Lambert & Tucker, 1972; Rubin, Turner & Kantor, 1991; Swain & Lapkin, 1982; Turnbull, Lapkin & Hart, 2001, 2003). And again, similar results regarding the development of children’s respective L1s have also been reported for IM programs in other countries (e.g., Cheng, Kirby, Quiang & Wade-Woolley, 2010 for China; Björklund & Mård-Miettinen, 2011 for Finland; Lasagabaster, 2001 for the Basque Country; Jenniskens, Leest, Wolbers, Krikhaar, Teunissen de Graaff, Unsworth, & Coppens, 2018 for the Netherlands). As students’ reading performance is also affected by their basic cognitive abilities, their social status and the level of their parents’ education (e.g., Baumert et al., 2017), these factors have been controlled for in many studies. The results of the studies generally indicate that attending IM programs does not seem to negatively affect the development of L1 reading skills of majority language children. In studies on German writing skills (which mainly relate to spelling skills), el‐ ementary school students in partial IM programs and in less-intensive bilingual programs have been found to perform equally well as their peers in mainstream schools at the end of grade 4 (e.g., Böttger & Müller, 2020; Gebauer et al., 2012; Möller et al., 2018; Steinlen, 2018b; Steinlen & Piske, 2013, 2018a, 2020; Zaunbauer et al., 2005, 2013, Zaunbauer & Möller, 2006, 2007, 2010; to avoid possible selection effects many of these studies used basic cognitive abilities and socio-economic status as covariates). Similarly, in Canadian French IM programs, L1 English students’ English writing skills did not differ from their peers in regular programs (e.g., Genesee, 1987, 2004; Rubin et al., 1991; Turnbull et al., 2001). However, available evidence indicates that deficits initially appear in early full IM (Barik & Swain, 1976, but see Yadollahi, Steinlen & Piske, 2020) although these seem to balance out in one or two years after writing lessons in the L1 English have been introduced (Genesee, 2007; Rubin et al., 1991). In general, L1 literacy skills (be it reading or writing) of majority language children appear to develop age-appropriately in IM programs as well as in bilingual programs with lesser FL intensity. 43 2.3 Bilingual programs in Germany 2.3.11 FL skills in regular vs. bilingual programs There is an abundance of studies which compare FL reading and writing skills in IM and mainstream foreign language programs. The results are unambiguous: IM students (independent of the type of IM program, i.e., early or late, full or partial) outperform comparable non-IM peers in regular programs in any kind of FL language test (e.g., Genesee, 2004; Pérez-Cañado, 2012; Wesche, 2002). Regarding the development of FL reading skills by majority language children in IM programs in Germany, Zaunbauer et al. (2012) reported that at the end of grade 4, immersion students obtained higher scores in tests on English reading fluency and comprehension than their peers in mainstream programs. Steinlen & Piske (2018a), who examined 136 fourth graders in a regular and a partial IM program, reported similar results. Not surprisingly, the two groups differed with respect to their English reading and writing skills, with the partial IM students being at level A2/ B1 for English reading, and the students in the regular program at level A1. Some students in Zaunbauer et al. (2012) even scored within the range of native-speaker norms (see also Nold et al., 2008 for ninth graders in bilingual programs in Germany). Such native-like reading skills may emerge during the fourth year of immersion teaching but, as other authors noted, not earlier than that (e.g., Cashion & Eagan, 1990; Geva & Clifton 1994; Malicky, Fagan & Norman, 1988; Möller et al., 2017; see also Genesee & Jared, 2008 for more details). In general, the results for IM programs in Germany agree well with findings reported for French IM programs in Canada (e.g., Genesee, 1978; Genesee & Jared, 2008; Lambert & Tucker, 1972; Swain & Lapkin, 1982). FL reading and writing skills in less intensive bilingual elementary school programs were studied by Böttger & Müller (2020). 430 students in grade 4 completed an adapted version of the Cambridge English Test for Young Learners (Flyers & Movers), which assessed reading comprehension and writing skills in combined tasks. The findings indicated that “almost all students reach the language competence level A1 and many of them can even reach A2 level in different areas” (p. 40, own translation), although additional information (e.g., on the exact number of students reaching different levels and why; or on separate results for reading and writing) is, unfortunately, not provided. There is a general lack of studies for the elementary school context in Germany and elsewhere with respect to children’s writing skills in the target language. Many studies conducted in Germany are rather exploratory and do not include any statistical analyses (e.g., Böttger & Müller, 2020; Burmeister, 2010; Rymarczyk, 2010). Steinlen (2018b) and Steinlen & Piske (2018a, 2020) compared English writing skills by 136 fourth graders who either attended a regular or a partial IM program. They reported that students in regular programs 44 2. Literature review were at level A1, whereas students in a partial IM program were at level A2, with significant differences between the groups, even when controlling for socio-economic status, family language and cognitive ability. Outside of Germany, it has also been shown that IM students generally outperform their peers in mainstream programs (e.g., Bae, 2007; Genesee, 1987, 2004; Wesche, 2002). Nevertheless, native-like writing skills in the FL as measured in terms of fluency, complexity (both lexical and grammatical), and accuracy are usually not achieved in IM programs (e.g., Wesche, 2002). IM students apparently produce similar errors in FL writing as their peers in mainstream programs (e.g., with respect to L1 transfer, e.g., Burmeister, 2010); however, their texts are longer and more complex in structure and wording (e.g., Bae, 2007; Rymarczyk, 2010; Swain, 1975). Most studies in this context have, however, been carried out with secondary school students (e.g., Nold et al., 2008; Zydatiss, 2009, for Germany; Bulon, Hendrik, Meunier & Van Goethem, 2017 for Belgium; Lyster, 2007; Swain & Lapkin, 1995 for Canada; Roquet & Pérez-Vidal, 2017 for Spain) with a focus on longer texts as well as on the use of grammatical structure and idiomatic and pragmatic expressions. In general, the results strongly indicate that FL literacy skills develop partic‐ ularly well in IM programs. This is probably due to the fact that a much higher number of writing activities are carried out in FL subject lessons than in regular FL lessons; and that the students not only receive more written input, but also have to produce more written output. Chapter 7 will point out many activities that are currently employed in FL bilingual (but also in regular FL) classrooms to foster FL literacy skills. 2.3.12 FL reading and writing in bilingual programs with different FL intensity So far, a within-group comparison of FL reading and writing skills in bilingual programs with different FL intensity has not been carried out in Germany. For Canada, such a comparison included only immersion programs with different intensity. Comparing total vs. partial IM (i.e., 50/ 50) programs with an early start, the findings demonstrated that the gains (e.g., in reading) vis-a-vis early total IM will be lower, or roughly commensurate with the relative total time spent in the FL (Wesche, 2002). Late-starting low-intensity bilingual programs (i.e., 20/ 80 programs) may be quite successful in enhancing FL English profi‐ ciency of academically-inclined (Gymnasium) students (Wesche, 2002; see also Burmeister & Daniel, 2002; Wode, 1995, 2009). Evidence for 20/ 80 programs with an early start is, unfortunately, scarce. The findings of smaller-scale studies 45 2.3 Bilingual programs in Germany indicate improvement in FL learning (regarding FL reading, writing, grammar and vocabulary comprehension) from grade 1 to 4, but not always significant differences from year to year (e.g., Böttger & Müller, 2020; Couve et al., 2016; Steinlen & Gerdes, 2015). There seem to be very few studies which compared low vs. high FL intensity programs, and these either relate to FL receptive vocabulary of elementary school children or to older learners. For example, Couve et al. (2016) compared 61 students in three different bilingual classes in Germany and assessed their per‐ formance for English receptive vocabulary, using the British English Vocabulary Scale. The students, who were tested in grade 1 and 2 (at T1 and T2), attended one of three programs where either 20 %, 25 % or 30-40 % of the teaching time were conducted in English. The results yielded significant differences between the 25 % and the 30-40 % program and between the 20 % and the 30-40 % program but not between the 20 % and the 25 % program at both T1 and T2. Similarly, in terms of FL vocabulary gains within a year, significant differences between T1 and T2 were found for the 25 % and the 30-40 % program but not for the 20 % program. Couve et al. (2016) attributed these group differences to class composition, with an uneven distribution of children in the three classes who either had previous knowledge of the target language and/ or higher metalinguistic awareness due to a multilingual background. Turning to older learners in different IM programs, Reeder, Buntain & Takakuwa (1999) examined almost 80 English students in grade 6, who attended either a 50/ 50 or an 80/ 20 French IM program in Canada. They were tested on FL French reading as well as on descriptive and narrative French writing. The results indicated significant between-group differences for French reading and narrative writing (but not for French descriptive writing), thus underscoring the positive effect of FL intensity for the development of FL literacy skills. In sum, studies on bilingual programs indicated a) that students’ L1 reading and writing skills develop age-appropriately and b) that FL competence in reading and writing depends on the intensity of the FL program on the one hand and on the start of the FL program on the other hand. 2.4 Multilingual background This chapter focuses on the effects of language background on elementary schoolers’ reading and writing skills in German and English in programs with different FL intensity. First, some terms revolving around a multilingual background are defined (section 2.4.1), then results of studies on minority 46 2. Literature review 7 „Eine Person hat einen Migrationshintergrund, wenn sie selbst oder mindestens ein Elternteil die deutsche Staatsangehörigkeit nicht durch Geburt besitzt“ (Statistisches Bundesamt 2020: 4). A person has a migration background when s/ he or at least one parent is not a German citizen by birth (own translation). 8 Accordingly, these changes in defining a “migrant background” are also reflected in the Steinlen studies 2013-2020. 9 Zu den Menschen mit Migrationshintergrund zählen „alle nach 1949 auf das heutige Gebiet der Bundesrepublik Deutschland Zugewanderten, sowie alle in Deutschland geborenen Ausländer und alle in Deutschland als Deutsche Geborenen mit zumindest einem zugewanderten oder als Ausländer in Deutschland geborenen Elternteil“ (Statis‐ tisches Bundesamt 2012: 6). People with a migration background include all those who immigrated to the territory of the Federal Republic of Germany after 1949, as well as all foreigners born in Germany and all those born in Germany as Germans with at least one immigrant or foreign-born parent (own translation). language students’ literacy skills in the school language (L2 German) and in the FL (L3 English) are presented (sections 2.4.2 and 2.4.3), including models on L3 acquisition. 2.4.1 Facts and figures According to the Statistisches Bundesamt (2020: 4), 26 % of the total population and around 38 % of all children under the age of 10 years nowadays hold an immigrant status in Germany. Such a status is assigned when a person or one of his/ her parents does not have German citizenship. 7 This definition has actually changed in recent years: 8 Until 2012, for example, an individual was regarded as an immigrant if s/ he (or at least one of her/ his parents) migrated to Germany after 1949 (Statistisches Bundesamt, 2012: 7). 9 In Germany, most of these immigrants come from diverse linguistic and ethnic backgrounds, which often coincide with a relatively low socio-economic status (e.g., Bos & Pietsch, 2006; Chudaske, 2012; DJI, 2020; Hesse, Göbel & Hartig, 2008; Kessler & Paulick, 2010; Schwippert, Wendt & Tarelli, 2012). According to Genesee & Fortune (2014: 196), children from an “ethnically and linguistically diverse background […] including minority ethnic groups” are often considered ‘at-risk’ students. This is also the case in Germany, where an immigrant status in conjunction with a first language (L1) background different from the majority language German has repeatedly been described as a risk factor for children’s adequate acquisition of German, and where limited skills in German have repeatedly been found to hinder participation in education and society (e.g., Baumert & Schümer, 2001; Chudaske, 2012; DJI, 2020; Dubowy, Ebert, von Maurice & Weinert, 2008; Möller et al., 2017; Stanat, 2006). 47 2.4 Multilingual background Regarding terminology, immigrant (or migration) background and multilingual background are often not clearly distinguished. In English publications, the terms “minority language children” (e.g., Bild & Swain, 1989; Genesee, 1987; Genesee & Jared 2008; Steinlen, 2016, 2017) or “multilingual children” (e.g., Bérubé, Daniel & Marinova-Todd, 2012) are most often used. That is, the focus is on the linguistic background of the children, albeit focusing on different aspects. The term “multi‐ lingual” characterizes the use of more than one language (e.g., Cenoz, 2013) and often relates to the number of languages which an individual uses (e.g., Turkish, German, English). The term “minority language” (i.e., a language that is spoken by a small group of people in the country) focuses on the children’s family language/ s (e.g., Turkish) which does not correspond to the ambient (majority) language (in this case German, e.g., Van Dongera, Van der Meer & Sterk, 2017). The term “heritage language” is also often used, defined as the speaker’s “language of origin” (Arnett & Mady, 2013: xv, see also Brehmer & Mehlhorn, 2018), often used with a cultural connotation. In other words, a multilingual child in Germany may grow up with Turkish at home as the heritage language (which is incidentally a minority language in Germany), learning German as a second language in kindergarten (as the majority language) and English as the first foreign language (not the ambient language) in school. In studies from German-speaking countries, the term “migration back‐ ground” (e.g., Chudaske, 2012; Keßler & Paulick, 2010; Piske, 2017; Siegert, 2008; Steinlen & Piske, 2013) is very often used, although such children do not necessarily have to have a multilingual background. More recent publications in this context (i.e., Hopp, Kieseier, Vogelbacher & Thoma, 2018; Wilden & Porsch, 2015; Steinlen, 2018a, b; Steinlen & Piske, 2018a, 2020; Yadollahi et al., 2020) rather use the term “multilingual” students, focusing on the languages a child grows up with. In this respect, a “monolingual” child in Germany grows up with German only, and a “multilingual” child with another family language (e.g., Arabic, Russian, Turkish, etc.). In some studies, the latter group is subdivided into “bilingual” children (who speak a language other than German as their heritage language at home and also German) and “heritage” language children (who only speak the heritage language at home, e.g., Hopp, 2018). Such a distinction, however, is not part of this study and must be left to future analyses of the present data set. In the following, the terms “minority language background” and “multilingual background” are used synonymously, focusing on the children’s linguistic background (i.e., on the language/ s spoken at home and in school) and not on the children’s citizenship or their parents’ birthplace. Chapters 2.4.2 and 2.4.3 focus on studies dealing with the development of the literacy skills of minority language children in German (the school language) on 48 2. Literature review 10 The abbreviations for the research projects read as follows (in alphabetical order): DESI (Deutsch Englisch Schülerleistungen International, German English Achievements of Students International), DFG (Deutsche Forschungsgemeinschaft, German Research Foundation), IQB (Institut zur Qualitätsentwicklung im Bildungswesen, Institute for Quality Development in Education), KEIMS (Kompetenzentwicklung in multilingua‐ len Schulklassen, Development of Academic Competence in Multilingual Classes), KESS (Kompetenzen und Einstellungen von Schülerinnen und Schülern am Ende der Jahrgangsstufe 4 in Hamburger Grundschulen, Fourth Graders’ Competences and Attitudes in Hamburg), PIRLS (Progress in International Reading Literacy Study), PISA (Programme for International Student Assessment). the one hand, and English (the FL target language) on the other hand; at either a regular program or a bilingual program at the elementary school level. 2.4.2 L2 literacy (German) 2.4.2.1 Minority language children’s German literacy skills in regular programs There is an abundance of studies examining German literacy skills of minority language children in mainstream elementary schools in Germany. The results clearly indicate that these children generally score lower than majority language children do (for reading, e.g., Bos & Pietsch, 2006: KESS 10 ; Chudaske, 2012: KEIMS; Dollmann, 2010: DFG-Köln; Haag, Kocaj, Jansen & Kuhl, 2017: IQB; Hussmann, Stubbe & Kasper, 2017: PIRLS; for writing, e.g., Chudaske, 2012: KEIMS; Haag et al., 2017: IQB; May, 2006: KESS). The same holds true for secondary schools in Germany. Here, majority language ninth graders have been found to perform better than their minority language peers in tests on German reading comprehension (e.g., Hesse et al., 2008: DESI; Haag, Böhme, Rjosk & Stanat, 2016: IQB, see also OECD, 2019: PISA 2018 for the international context) and on German writing (Haag et al., 2016: IQB). There are two explanations that have repeatedly been offered for the finding that minority language students attending mainstream programs performed worse than majority language students in studies examining literacy skills in the majority language German. First, there is a mismatch between the minority lan‐ guage children’s L1 and German as the main language of instruction. Second, a minority / immigrant background is often associated with a low socio-economic status, and children from such a context have been found to be less exposed to print at home than comparable children with a higher socio-economic status (e.g., Bos & Pietsch, 2006; Chudaske, 2012; DJI, 2020; Haag et al., 2016; Hesse, Göbel & Hartig 2008; Mullis, Martin, Foy & Hooper, 2017; OECD, 2019; 49 2.4 Multilingual background Schwippert et al., 2012 and see chapter 2.7). Note that both explanations may account for minority language students’ academic achievements in the school language. 2.4.2.2 Minority language children’s German literacy skills in bilingual programs For the bilingual (immersion) context in elementary schools in Germany, Stein‐ len & Piske examined in several studies how well minority language children’s reading and writing skills in the majority (school) language German develop. They used two standardized tests on German reading (i.e., Würzburger Leise Leseprobe-R: Schneider, Blanke, Faust & Küspert, 2011; and Ein Lesetest für Erstbis Sechstklässler: Lenard & Schneider, 2006) and one on German writing (Ham‐ burger Schreibprobe, May, 2010). The results of their studies (with an increased sample size from 2013 to 2020) did not show any differences between majority and minority language children attending a partial IM program. Furthermore, both groups achieved age-appropriate results on all tests. Possible explanations for this finding include the minority language learners’ social status (with parents from both groups having a rather middle class and an academically oriented background). Furthermore, high quality German-as-a-subject lessons may have also contributed to the results (Steinlen, 2016, 2018b; Steinlen & Piske, 2013, 2015, 2016a, 2018a, 2020, i. pr.; Yadollahi et al. 2020). Turning to two-way IM programs, Möller et al. (2017) examined 769 fourth graders in various bilingual schools in Berlin (Germany) with different target languages, showing different results. In this context, minority language children in grade 4 generally performed worse than their majority language peers in the German reading test, in particular in programs offering Polish and Turkish. The authors pointed to possible pre-selection and class composition effects, which may have negatively affected the results, without providing more detailed explanations. However, in two-way English-Spanish IM programs in the United States, elementary and secondary school students with Spanish as the L1 performed as well as their English monolingual peers on English reading tests, particularly in more advanced grades (e.g., Lindholm-Leary & Hernández, 2011; Marian, Shook & Schroeder, 2013; Thomas & Collier, 2012, and see also Baumert et al., 2012; Duarte, 2011 for similar results in other two-way IM programs in Germany). It is important to point out here that the inconsistent results concerning the development of the school language by minority language students in IM programs are due to differences in the subjects’ age and/ or social background, to the different literacy skills being tested (reading vs. writing) and/ or to program differences (one-way vs. two-way IM programs). For the IM context, there is 50 2. Literature review clearly a need to examine the development of minority language children’s reading and writing skills in the majority language in more detail. 2.4.3 L3 literacy (English) Among other things, one focus of the present book is to explore minority language children’s reading and writing skills in the FL English, for whom English very often is the L3 (or even the L4 or L5). Unfortunately, clear definitions of L3 reading or L3 writing per se and their characteristics have not been put forward yet. This gap may be due to the fact that L3 acquisition research is a relatively recent field which has been established alongside the fields of bilingualism and second and foreign language acquisition only in the past twenty years (e.g., Rauch et al., 2010). Extending Cummins’ Interdependence Hypothesis (e.g., Cummins, 1991), it is assumed that language learners who have already acquired two languages are able to draw on competences in their first language (L1) as well as in their second language (L2) to acquire a third language (L3, e.g., Rauch et al., 2010). In addition, linguistic competencies of later acquired languages (e.g., L3, L4) may also transfer back to previously acquired languages (L2, L1). As an example, Rauch et al. (2010) referred to grammatical knowledge learned in the FL classroom which may positively affect language-analytical competences in the children’s L1. Such competences may then also be useful for grammatical activities in German-as-a-subject class (with German being the minority language children’s L2 in Germany). Similarly, reading literacy in the L1 may be influenced by a student’s knowledge of FL reading strategies (e.g., skimming, scanning, etc.). Even though the processes used in third language acquisition may be very similar to those used by L2 learners, it is generally assumed that the additional language may complicate cognitive operations of linguistic processes. Within the field of L3 acquisition, some approaches have been proposed which may account for transfer patterns in L3 production, although none of them explicitly refer to reading and writing in a third language. These approaches and how they may transfer to L3 literacy skills will be discussed below. 2.4.3.1 General models of L3 acquisition With respect to L3 acquisition in general, the question of which language is activated in L3 production (particularly regarding lexis) has raised a lot of attention. Williams & Hammarberg (1998) proposed four interacting factors which determine which language will most often become co-activated in L3 51 2.4 Multilingual background production. These factors include a) recency (the extent to which the learner has used the language recently), b) proficiency (the learner’s level of proficiency in the background language/ s), c) typological similarity between the background language/ s and the L3, and d) status (i.e., if the background language is an L1 or L2 for the learner). It has been hypothesized that the language that reaches the highest total score on these four scales (recency, proficiency, similarity and status) is the one which the learner will activate most often. Other contributing factors have also been suggested, such as whether a language has been acquired in natural or formal contexts (Ringbom, 1987), whether it has been used actively (Heine, 2004), at what age a language has been acquired (Cenoz, 2001), and the learner’s emotional attitude towards activating a certain language (Hammarberg & Williams, 2009). Some studies on L3 acquisition have indeed suggested an influence of speech typology and language distance so that, for example, learners with a Romance background may find it easier to study French as a target language than learners with non-Romance backgrounds (e.g., Bild & Swain, 1989). Also, the prestige that the learner’s L1 and L2 enjoy in society and, of course, L2 skills and frequency of use seem to influence the activation of a language in L3 production (e.g., Angelovska & Hahn, 2014; Williams & Hammarberg, 1998). Based on Hufeisen’s Factor Model (e.g., 2018), Hammarberg (2018) proposed a variable model of the actual situations of language use for L3 reception and L3 production. The processes of use and acquisition take place as the individuals interact with their current linguistic environment, receiving and/ or producing utterances in this context. The model (see Figure 2 below), therefore, includes factors such as the users’ linguistic knowledge (i.e., the current knowledge of one or more L1s, one or more familiar L2s, and the current state of L3 proficiency) and their contextual knowledge (e.g., awareness of the preceding discourse, understanding of situation, world knowledge pertinent to situation). Similar to Hufeisen’s Factor Model, Hammarberg’s model also takes into account individual factors which include neurophysiological factors (e.g., age, general language learning capacities), learner-external factors (e.g., learning environment/ s, types and amount of input), affective factors (e.g., motivation and anxiety), cognitive factors (e.g., language and metalinguistic awareness) and specific L2 learning factors (e.g., L2 experiences, strategies). 52 2. Literature review Figure 2: L3 use by a multilingual speaker (adapted from Hammarberg 2018: 138). Regarding L4 acquisition, Hufeisen (e.g., 2018) argues that beginning to learn a fourth language adds another language to an individual learner’s language/ s repertoire in quantitative terms more so than in qualitative terms. That is, the language-specific factor now includes all languages that the user is familiar with, with all other factors remaining the same. Hufeisen’s Factor Model 2.1 (2018) relating to L4 acquisition is depicted in Figure 3. 53 2.4 Multilingual background Figure 3: Factor Model 2.1 for L4 acquisition (adapted from Hufeisen, 2018: 186). Some models on L3 acquisition have recently been proposed which are specif‐ ically geared at L3 grammar (see Hopp, 2018 for a detailed account). These models are mentioned because grammatical knowledge also affects reading and writing outcomes, independent of the acquisition type. Most of these models deal with grammatical transfer in adult L3 acquisition and focus on the initial state of sequential L3 acquisition. It is postulated that the initial grammar for the developing L3 is based on a copy of either the L1 or the L2 grammar. For example, some studies have proposed that the L1 constitutes the default language from which transfer originates (Privileged L1 Transfer Hypothesis, e.g. Hermas, 2010, 2014b; Jin, 2009; Na Ranong & Leung, 2009, all in Hopp, 2018). Other studies found the L2 to act as the default source of grammatical transfer, for it represents the most recently acquired non-native or foreign language (L2 Status Factor Model, e.g., Bardel & Falk, 2007). However, this model makes predictions only for adult L3 learners but does not apply to child L3 learners because it is grounded in the maturational difference between early and late L3 acquisition. Children, in contrast to adults, acquire all languages before maturational constraints allegedly restrict the employment of the procedural memory system. According to Bardel & Falk (2012), the L2 can take on the role of a (second) L1 for child learners, so that transfer from either the L1 or the L2 may be compatible with the L2 Status Factor Model. A different approach is taken in the Typological Primacy Model (e.g., Rothman, 2015), which claims that grammatical transfer is 54 2. Literature review determined by similarities in phonology, morphology, syntax and lexis between languages. Depending on the most similar language, either the L1 or the L2 grammar transfers as a whole at the L3 initial state. Flynn et al. (2004), in turn, argue that transfer is property-specific in that learners transfer similar L1 or L2 properties to L3 properties, so that they facilitate L3 acquisition (Cumulative Enhancement Model). Finally, in her Scalpel Model, Slabakova (2017) focuses on the stages beyond the initial state of L3 morpho-syntactic learning, arguing that the activated grammatical possibilities of the L1-plus-L2 combined grammar act with a scalpel-like precision, rather than as a blunt object, to extract the enhancing, or facilitative, options of L1 or L2 parameter values. It is not clear to what extent such models of L3 grammar transfer also apply to L3 literacy skills, although they have been used to account for grammatical mistakes in L3 adult writing, as shown by Angelovska & Hahn (2012). Therefore, it remains an important field of study to determine the following aspects: how long does the “initial state” last? In other words, are, for example, fourth graders in IM programs still in the initial state? Which of the children’s languages (i.e. their L1 or their L2 German) constitute the basis for any transfer patterns in L3 English? Are the respective languages transferred “whole-sale” or only by specific properties? 2.4.3.2 L3 Reading in regular FL programs Many studies relate to L3 listening, vocabulary, or grammar in the regular EFL elementary school context (e.g., Elsner, 2007; Hopp, Kieseier, Vogelbacher, Köser & Thoma, 2017; Hopp, 2018; May, 2006), but only a few studies have examined reading and writing development in a third language. L3 English reading skills in mainstream programs were examined by Kessler & Paulick (2010), based on data of the EVENING study with 1748 students in grade 3 and 4 (see chapter 2.2.4). They found that majority language (i.e., German) fourth graders achieved slightly better results than their minority language peers (unfortunately, no statistical analyses were supplied). Moreover, Turkish-speaking children were reported to have been outperformed by children with L1 Russian, French, Spanish or Polish, who obtained scores similar to those of German children. In interviews, some English teachers pointed out that many minority language children exhibited outstanding abilities, were highly motivated and outspoken (in contrast to their behavior in other subjects), and often experimented with the new language. Kessler & Paulick (2010) concluded that minority language children per se do not automatically constitute an at-risk group in early FL teaching. 55 2.4 Multilingual background Similar results were reported in studies conducted in Switzerland. For exam‐ ple, Haenni-Hoti (2008) examined almost 600 Swiss-German third graders, with a quarter of them having a minority language background. She found that FL English reading skills correlated with German reading skills, but that language background did not predict the outcome of the English reading test. In their project “Ganz In - Mit Ganztag mehr Zukunft. Das neue Ganztags‐ gymnasium in NRW” (The New All-Day Secondary School in North Rhine-West‐ falia), Wilden & Porsch (2015) examined over 5.000 students with a minority and a majority language background at the beginning of grade 5 regarding their reading skills in English. They compared two groups, namely students with only two years and those with three and a half years of English experience from elementary school. Apart from the fact that all students’ English reading skills improved as a function of exposure to the target language, the results did not show any influence of language background when the students’ gender and socio-economic status were controlled for. Different results were reported by Steinlen & Piske (2018a) who compared 73 minority and majority language fourth graders (3 cohorts) in a regular program regarding their English reading skills. Although both groups reached level A1, the minority language children scored lower than their majority language peers, with significant differences between groups (but see Yadollahi et al., 2020 for different results for first graders). The authors attributed this finding to lower competences in German because this group of minority language children also obtained considerably lower scores in the standardized test for German reading. These findings are in line with results reported for English listening comprehension tests in regular school programs (e.g., Elsner, 2007; May, 2006). For older students, the large-scale study DESI (Deutsch Englisch Schüler‐ leistungen International) did not report any significant differences between minority and majority language students with respect to their performance in English reading comprehension tests in grade 9 (Hesse et al., 2008; see also Rauch et al., 2010). In a more recent study (IQB-Bildungsvergleich 2015), minority language students were subdivided into two subgroups: Haag et al. (2016) reported that ninth graders with one parent born outside of Germany performed as well as their peers with a German background. However, students whose parents were both born abroad (particularly those with a Turkish or Arabic background) obtained lower scores in English reading tests than the majority language students and minority language students with one parent born outside of Germany, although these differences were smaller when the students’ social background was controlled for. 56 2. Literature review 11 In Griva & Chostelidou’s paper (2013) metacognitive strategies for writing refer to the awareness of the writing process and include planning, selected attention and reviewing. Cognitive strategies while writing a text pertained, for example, to translations from the L1 or L2, composing a text sentence by sentence and re/ drafting. In order to overcome limitations in writing, students also employed compensation strategies, e.g., by using synonyms, by switching to the L1 or L2 or by asking for help. 2.4.3.3 L3 Writing in regular FL programs Few studies have examined minority language students’ L3 writing skills at elementary school age level in regular FL programs in Germany. For example, Yadollahi et al. (2020) examined six cohorts with 128 students at the end of grade 1, and reported that minority language children performed as well as majority language peers in the English reading and writing test, and that the differences between the groups were not significant. In the study by Steinlen & Piske (2018a) with 73 children in grade 4, minority language children scored significantly lower in the English writing test than their majority language peers, although both groups reached level A1. It is possible that the inconsistent results in Steinlen & Piske (2018) and Yadollahi et al. (2020) whose students all attended the same regular school program may be due to effects based on either input (with first graders being less exposed to English than fourth graders), the students’ age (7 vs. 10 years) or the sample size (73 students in grade 4 vs. 128 students in grade 1). For secondary schools, the DESI study included a task on English text production for ninth graders. A figure in Hesse et al. (2008: 218), which illustrates effects of language background on various English tests, points to differences between students with a majority and a minority background. Unfortunately, the authors only discussed language background effects for English reading and listening skills but not for English text production. International studies on minority students’ English writing abilities in regular FL elementary school classrooms are also scarce. For example, Griva & Choste‐ lidou (2013) examined 32 bilingual students in Greece, aged between 10 and 12, from Albanian, Russian and Georgian backgrounds, regarding their writing in the third language English. They found that success in L3 English writing depended on competence in L1 and L2; and that metacognitive (but not cognitive or compensatory) strategies 11 acquired in L1 and L2 were transferred to the L3. The less-skilled writers in particular had problems with gaining control of the ‘basics’ of L3 writing (i.e. spelling, vocabulary and grammar), and their major concern was to recall and to use the appropriate vocabulary and the correct spelling. In contrast, more skilled L3 writers’ think-aloud protocols revealed 57 2.4 Multilingual background general metacognitive procedural knowledge and control which they also used in their L1 and L2 texts (see Cenoz, 2003). For the secondary school context, Cenoz & Gorter (2011) examined writing skills of 165 bilingual Basque/ Spanish 15-year-old students who had started to learn the L3 English in grade 1. The texts in Basque, Spanish and English were evaluated according to content, structure, vocabulary, grammar, and orthography. The results showed that these aspects were cross-linguistically interconnected. In other words, students with a high score in orthography or vocabulary in one of the languages also obtained good results in the other languages and vice versa. Cenoz & Gorter (2011) argued in favor of a “flexible bilingual pedagogy” (p. 367) which should be employed in schools. According to the authors, boundaries between languages should be permeable as the traditional strategy of separating languages and using only the target language in a class does not allow multilinguals to use their full resources in communication. 2.4.3.4 L3 reading in bilingual programs The development of L3 literacy skills in bilingual schools, especially by young learners, constitutes the focus of Steinlen & Piske’s project. They examined minority language children in a partial IM school in Tübingen (Germany) where 50 % of the teaching time is conducted in English (see section 4.1 for additional information). Starting with one cohort (Steinlen & Piske, 2013) and adding a new cohort each year (e.g., Yadollahi et al., 2020 with six cohorts), the results of the English reading tests are consistent over the years. Minority language students in this program performed as well as their majority language peers, without any significant differences between the two groups. This result was found independent of the grade, the sample size (one cohort vs. six cohorts), or the English reading test format, i.e., Primary School Assessment Kit - Reading (Little et al., 2003), Test of Reading Comprehension (Mossenson, Stephanou, Forster & Masters, 2003) and Test of Silent Word Reading Fluency (Mather, Hammill, Allen & Roberts, 2004). Generally, minority language students obtained level A2/ B1 at the end of grade 4 (Steinlen & Piske, 2016a, 2018a; Steinlen, 2018b), and the children obtained scores that were comparable to those of monolingual English children who were one year younger regarding reading comprehension and two and a half years younger for reading fluency (e.g., Piske, 2020; Steinlen, 2016a, 2017, 2018a; Steinlen & Piske, 2013, 2014, 2015, 2016b). For older learners, in their review on the development of reading skills in IM programs in Canada, Genesee & Jared (2008) also summarized findings from eighth graders from middle partial IM programs in Toronto (e.g., Bild 58 2. Literature review & Swain, 1989; Swain & Lapkin, 1991; Swain, Lapkin, Rowen & Hart, 1990). Generally, minority language students in such programs performed as well as their majority language peers. They obtained particularly good test scores when they had also developed literacy skills in their family language. For Spanish-English two-way IM programs in the U.S., it was reported that minority and majority language students did not differ significantly regarding their performance in English reading tests (e.g., Lindholm-Leary & Hernández, 2011; Marian et al., 2013; Thomas & Collier, 2012). However, as Genesee & Jared (2008) emphasized, there is still little research on the impact of L1 literacy experiences on FL (L3) reading skills. In their view, longitudinal studies are warranted in order to study reading development of young minority language students in IM (and any other bilingual) programs in more detail. 2.4.3.5 L3 writing in bilingual programs L3 writing skills in elementary schools in Germany were examined by Rymarczyk (2010), with 107 children in four IM classes and one regular FL class, all at the beginning of grade 3. Of these children, 34 had a minority language background. The children’s task was to write down as many English words as they could. The results indicated that minority language children generally produced more FL words than their majority language peers (unfortunately, statistical analyses were not provided). Rymarczyk suggested that minority language children may benefit from prior experience with other languages (i.e., higher metalinguistic awareness and less inhibition to commit mistakes) and from general cognitive advantages (e.g., in terms of divergent thinking). Furthermore, Rymarczyk (2010) noted that minority language children were particularly successful when their parents showed an interest in education. Rymarczyk (2010) concluded that neither majority nor minority language children were overburdened with the English writing system, and that children generally should be confronted with FL literacy experiences as early as possible in order to avoid invented rules and fossilization (see also chapter 7.7.1). In their studies, Steinlen & Piske used the Primary School Assessment Kit (Little et al., 2003) to assess English writing skills of students with different language background who all attended a partial IM program in Tübingen. As for English reading, the authors did not find any significant differences between majority and minority language children whose English writing skills were at level A2 at the end of grade 4 (Steinlen, 2018b; Steinlen & Piske, 2016a, 2018a, 2020). Non-significant group differences were also reported for English writing skills by minority and majority language learners in the same program at the end of grade 1 (Yadollahi et al., 2020). The authors argue that this finding may be 59 2.4 Multilingual background influenced by family variables (e.g., socio-economic status, parental education and support), the minority language students’ degree of language awareness, and the quality of teaching in the partial IM program. Studies relating to L3 writing have also been conducted with older minority language students in secondary school, for example in the Basque Country. Sagasta Errasti (2003) examined 155 students between the ages of 12 and 16 years who had started English-as-a-subject lessons at the age of 8, and later attended a Basque-Spanish IM program. The students were asked to write an informal letter and a recipe in Basque, Spanish, and English. Evaluations of the essays assessed content, organization, language use, vocabulary, and orthography on the one hand; and fluency, grammatical complexity, lexical complexity, and accuracy on the other hand. Sagasta Errasti’s (2003) results confirmed Cummins’ (1991) Threshold Hypothesis, emphasizing the influence of the level of language competence in each language. Those students who developed the highest degree of bilingual competence in Basque and Spanish also obtained the best scores in English. In addition, Sanchez Perez (2015) examined L3 English writing of 721 Basque-Spanish students who had been learning German as the first FL through CLIL, and English as the second FL through formal instruction in school (i.e., in regular FL English lessons). The subjects’ ages ranged from 9 to 16 years. The English data were collected by means of a written picture storytelling task, and were assessed according to fluency, grammatical complexity, and accuracy. The results indicated that the assessment measures for English were highly correlated. For example, the higher the scores in English grammatical complexity, the higher the scores in English fluency and/ or English accuracy, paralleling findings from L2 acquisition (e.g., Archibald & Jeffrey, 2000). Sanchez Perez (2015) pointed out that additional studies are required to provide specific details about the interaction between different FL (i.e., the interaction of fluency, grammatical complexity, and accuracy in German and English). For the Canadian immersion context, Hart et al. (1988) analyzed data on French proficiency in grade 8. The measures included, among others, an open-ended writing task in French. The authors found that the minority language students in their sample significantly outperformed English majority language students on this task (see also Bild & Swain, 1989 for an analysis of a subset of these data). In a subsequent analysis, Swain et al. (1990, see also Swain and Lapkin, 1991; Swain et al., 1991) investigated the role of heritage (L1) language proficiency and literacy on the acquisition of French as an L3 by minority language students in this middle IM program. Their results indicated a positive effect of L1 literacy knowledge on L3 literacy skills, independent of L1 oral/ aural skills, general level of L1 proficiency and the linguistic/ historical relationship between the L1 and the L3. More 60 2. Literature review specifically, Swain et al. (1990) argued in favor of ‘linguistic interdependence’ of the languages because they found minority language students’ literacy knowledge in their L1 to enhance fluency and length of their L3 texts. However, ‘linguistic interdependence’, in their view, does not extend to written L3 grammatical or spelling errors, which represent only so-called ‘surface level/ language specific’ aspects of language, because in their data Swain et al. (1990) did not find any significant differences regarding the number of such errors in minority and majority language students’ French texts. The French writing skills of students aged 11 to 12 were examined by Mady (2014, 2017), who compared Canadian-born English-speaking students; Cana‐ dian-born multilingual (immigrant minority language) students; and newly-ar‐ rived multilingual immigrant students (predominantly from South Asia and the Middle East) in early French IM programs. The comparison showed that the Canadian-born multilingual group scored on a par with the Canadian-born English-speaking group. The newly arrived multilingual immigrant group outperformed both the Canadian-born English-speaking and Canadian-born multilingual group. As subsequent analyses showed, a higher metalinguistic awareness could not account for this result, because the immigrant group scored significantly lower on a test of metalinguistic awareness than the Canadian-born multilingual group (Mady, 2017). However, this group showed significantly higher integrative motivation and willingness to communicate than the Cana‐ dian-born English-speaking group and the Canadian-born multilingual group; and higher instrumental motivation than the Canadian-born English-speaking group (Mady, 2014). The findings of the studies presented here confirm Somers’ conclusion (2017), namely that minority language students’ participation in bilingual (IM) programs did not seem to lead to any academic or linguistic disadvantages for this group. Furthermore, he noted that “CLIL is popular among immigrant minority language students as it allows them to acquire linguistic, social, economic, and symbolic capital. CLIL is a strategy for upward mobility, allowing them to better access national and international language communities and a globalized job market” (Somers 2017: 513). Finally, Somers called for further research, particularly with regard to effects of the languages in the bilingual program on minority students’ home languages, their academic achievements in content subjects, and the development of their motivation, self-concept and identity in bilingual programs. 61 2.4 Multilingual background 2.5 Gender A common stereotype is that girls master language learning tasks faster and more easily than boys (see e.g., Klann-Delius, 2005 for a review). In the following, results regarding gender differences in language learning in elementary schools are presented, whereas children’s acquisition of German and English before entering elementary school or by adolescents and adults will only play a minor role. Different terms are used relating to the performance of girls and boys (e.g., Böttger, 2016; Elsner & Lohe, 2016). For example, the term “sex” refers to the biological/ physiological factum. When the focus is on the product of cultural and social processes, “gender” is the appropriate term. Following Schmenk (2002) or Böttger (2017), the term gender, defined as a social construct, is used in this book. However, both dimensions (sex and gender) are often closely related because, as the Aktionsrat Bildung (2009: 14) argues, female and masculine traits, skills and lifestyles are partly based on the biological differences between the sexes, such as the ability to give birth to children (see Uhl, 2016 for a detailed discussion). 2.5.1 L1 literacy skills in regular programs by gender Many individual variables may influence the development of L1 reading and writing skills, and gender is one of the variables that has often been examined. The findings are unanimous regarding boys’ and girls’ achievements in German reading and writing tests: girls show better results than boys. In large-scale studies examining German reading performance in regular ele‐ mentary school programs in grade 4, girls regularly outperformed boys (e.g., PIRLS-2016: Hussmann et al., 2017; see also IQB: Schipolowsky, Wittig, Weirich & Böhme, 2017). The same result has been reported for older learners in secondary schools in Germany (e.g., DESI-Konsortium, 2008; PISA-2015: OECD, 2019; IQB: Böhme, Sebald, Weirich & Stanat, 2016). Similar findings have been obtained in other countries for both regular elementary and secondary school programs (e.g., Mullis et al., 2017 for PIRLS-2016; and OECD, 2019 for PISA-2018). Girls also seem to have a head start in German written language learning (e.g., Richter & Brügelmann, 1994). For example, in the KESS study, girls performed better than boys in the Hamburger Schreibprobe in all aspects of spelling and punctuation; furthermore, they committed significantly fewer errors caused by lack of attention or control (May, 2006). Such effects have also been attested in other studies outside of Germany, which examined L1 writing skills in the mainstream elementary school context, (e.g., Kennedy, Dunphy, Dwyer, Hayes, McPhillips, Marsh, O’Connor & Shiel, 2012), as well as in studies on secondary 62 2. Literature review school students’ L1 German writing skills (e.g., Drucks, 2011; Hartig & Jude, 2008: DESI; Böhme et al., 2016: IQB). Several reasons have been discussed as to why girls outperform boys on German literacy tests. These range from girls generally having higher verbal intelligence than boys; being faster in acquiring language (resulting in better vo‐ cabulary and grammar skills); showing higher scores on tests of concentration; achieving higher scores on tests of letter and word knowledge at the beginning of grade 1; and often being already able to read before school entry (see e.g., Richter, 1996, for a review; see also Bos, Tarelli, Bremerich-Vos & Schwippert, 2012), among other possible reasons. Several studies also indicated that girls displayed many positive school-related traits and behaviors (e.g., intrinsic motivation, appreciation of school, self-discipline, completing homework) to a greater extent than boys, which may partially explain their better competences in reading (e.g., McElvany, Kessels, Schwabe & Kasper, 2017). Altogether, studies in general attested girls to have a higher motivation to read and a higher self‐perception of competence while reading. Furthermore, in grade 4, girls already assigned more importance to reading than boys, and their reading habits differed from boys in the sense that girls more often read for pleasure outside school (see e.g., McElvany et al., 2017, for a review). As these studies clearly indicate, girls seem to be advantaged when assessed with literacy tests in their L1, independent of the female students’ age, the school type and the test format being used. 2.5.2 L1 literacy skills in bilingual programs by gender There are only a few studies on gender and L1 (here: German) achievements in the bilingual elementary school context, and the results are inconsistent. In their sample of 769 fourth graders attending different bilingual programs with different target languages (i.e., English, French, Greek, Italian, Polish, Portuguese, Russian, Spanish and Turkish) at the SESB school in Berlin, Baumert et al. (2017) found girls outperforming boys in German reading comprehension tests. However, in the Steinlen & Piske (2016a) study, 94 fourth graders in a partial IM program did not show any significant gender-related differences in a standardized German reading comprehension test. Both groups obtained age-appropriate results (see also Steinlen, 2016, with a subsample, also employing Ein Leseverständnistest für Erstbis Sechstklässler, Lenhard & Schneider, 2006). The same applied to German writing skills, using the standardized Hambur‐ ger Schreibprobe (May, 2010) with the same fourth grade students (Steinlen & Piske, 2016a): girls and boys also achieved equally good (and age-appropriate) 63 2.5 Gender results. Similar results (i.e., no gender-specific differences for L1 reading) were reported for two-way IM programs in the United States, where students were tested at the end of grade 11 (e.g., Thomas & Collier, 2002). Results on gender, unfortunately, have not been reported in studies dealing with L1 reading and writing within the Canadian IM context. Thus, the few studies that have been conducted do not indicate a clear trend for gender effects in L1 reading and writing tasks in the bilingual classroom. In some studies, girls perform better than boys, while in others there are no significant gender effects. The scarcity of studies, using different sample sizes at different ages in different bilingual programs assessed by different test formats, points to a dire need for further studies in this field. 2.5.3 FL literacy skills by gender In 2002, Schmenk (in a very detailed review) pointed out that learners’ gender per se has no systematic influence on FL learning (Schmenk, 2002: 118). Never‐ theless, an increasing number of books and articles comparing the performance of boys and girls in foreign language classes have been published in recent years (e.g., Böttger, 2014, 2016, 2017; Elsner & Lohe, 2016; Fuchs, 2015). This is also reflected in media reports in Germany (e.g., “Mädchen lernen Fremdsprachen besser als Jungen”; Christina Merkel in the Nürnberger Zeitung on February 10, 2015). Such headlines may influence how teachers promote and challenge their students in the foreign language classroom, for example, in teacher-learner interactions, among other things (Schmenk, 2002: 89 ff). 2.5.3.1 FL literacy skills in mainstream programs by gender The results of studies examining gender-specific effects on FL reading and writing (particularly regarding the elementary school context) are consistent: girls achieved better grades in English (BIG-Kreis, 2015) and better results on English listening tests than boys (e.g., May, 2006; see also Burstall, 1975; Duursma, Romero-Contreras, Szuber, Proctor & Snow, 2007; Lynn & Wilson, 1993). Studies for secondary schools also reported gender differences in favor of girls for English listening and reading comprehension in grades 5 and 7 ( Jaekel, Schurig, Florian & Ritter, 2017), listening comprehension and C-tests in grade 8 (Nikolova & Ivanov, 2010) and English reading comprehension, a C-test, and text production in grade 9 (Hartig & Jude, 2008). A variety of reasons were proposed to account for these findings. For example, some studies indicated that girls showed more motivation to learn the new language, their attitude to the new language was more positive, their willingness to work was greater and 64 2. Literature review their parents’ support stronger (e.g., Kissau & Turnbull, 2008). In addition, girls interacted more frequently with teachers, were more active in peer interactions in the classroom, and seemed to benefit more from gender-specific activities in terms of teaching materials and their variety than boys (e.g., Böttger, 2016, 2017; Decke-Cornill, 2007; Fuchs, 2015; and the review by Schmenk, 2002). One exception is reported from the Netherlands, where boys in grade 8 outperformed girls on tests of English listening, reading and verbal knowledge (Edelenbos & Vinjé, 2000). Based on their observations, the authors suggested that this finding is due to the fact that boys spent more time in the Internet than girls in their free time and subsequently read and used the English language more frequently. Yet another result was found by Steinlen (2018a). She examined 110 students in grade 4 who attended a regular program. In this study, girls and boys did not differ significantly on English tests of reading and writing; the same applied to tests of English receptive grammar and vocabulary. The author suggested that the teachers in this regular program may have ensured that girls and boys participated equally in FL interactions and were encouraged to use learning strategies such as cognitive, metacognitive and social strategies. In addition, these teachers may also have used creative, experimental, kinesthetic and performative activities, which engaged girls and boys equally (e.g. games, role-plays, projects, videos). Similar results were reported in two studies on gender-specific differences in FL literacy skills in mainstream programs: Uhl, Schwanke & Piske (2020) tested 264 elementary school children’s French skills in Germany after their first year of learning and did not note any significant differences between girls and boys in a tests on French reading fluency. Finally, Uhl (2016) examined 40 ninth graders regarding their FL writing skills in French, and she also did not find any significant gender-specific differences. Even though the studies on regular teaching programs showed inconsistent results, girls outperformed boys in FL reading and writing tasks in most studies. However, L2 writing research in particular has not gone very far yet in exploring the interaction between L2 writing and gender (see also Leki et al., 2008). 2.5.3.2 FL literacy skills in bilingual programs by gender There are even fewer studies for the bilingual school context, and the results are as inconclusive as the ones obtained in the regular school context. For example, a study on 94 elementary school partial immersion program indicated no significant differences between girls and boys in grades 3 and 4 on an English reading test (e.g., Steinlen & Piske, 2016a, see also Steinlen, 2016, with a smaller sample size but different English reading tests). The same result was 65 2.5 Gender also reported for a larger sample of 123 students of the same partial IM program (including the 94 students from Steinlen & Piske, 2016a) and an additional 137 students from a less intensive bilingual program where only 20 % of the teaching time was conducted in English (Steinlen, 2018a, with the same results for receptive grammar and vocabulary). The same finding was also noted for students from secondary bilingual education programs regarding their FL reading skills (e.g., Canga Alsonso, 2015; Thomas & Collier, 2002; Zumhasch, 2010; Zydatiss, 2007). The authors suggested similar reasons as the ones mentioned above, namely, teachers ensuring that girls and boys alike participate equally in foreign language interactions, encour‐ aging them both to use learning strategies such as cognitive, metacognitive and social strategies or to use creative, experimental, kinesthetic and performative activities in class (e.g. Böttger 2016, 2017). However, a different result was reported for the nine two-way IM schools in Berlin. Baumert et al. (2017) and Fleckenstein et al. (2017) reported that girls outperformed boys on the reading tests of the partner languages, which applied to students in grade 4 as well as in grade 9. However, the authors pointed out that these gender-difference were smaller than the ones obtained for reading comprehension tests in German. Unfortunately, no explanation was provided for this finding. The effect of gender on FL writing skills in bilingual classrooms has only been examined by Roquet et al. (2016). They tested 100 Spanish-Basque bilingual students in grade 7/ 8 on short English writing tasks and assessed subskills such as task fulfilment, organization, grammar, and vocabulary. The subjects had all received FL English lessons since age 6 but continued secondary education either in a regular or in a bilingual CLIL program. The authors found significant differences between boys and girls in the CLIL program for task fulfilment and organization, but not for grammar and vocabulary. Because girls did not outperform boys in the regular program with respect to these subskills, Roquet et al. (2016) concluded that the CLIL context would not necessarily lead to a diminishment of gender differences relating to FL writing competencies (as proposed by Lasagabaster, 2008). So far, the results of the few studies on gender effects in FL reading and writing skills in the bilingual classroom are inconsistent. Lasagabaster (2008) pointed out that in bilingual programs, motivational differences between boys and girls may be compensated for by the fact that boys are indirectly encouraged by the instrumental necessity of the new language to advance their language development. In addition, topics in bilingual classes may also often be more relevant and authentic to girls and boys alike, in contrast to the content of 66 2. Literature review regular textbooks, which often appeal to girls rather than boys and may influence their motivation accordingly (e.g., Roquet, Llopis & Perez Vidal, 2016). More systematic approaches are, however, needed to examine the impact of gender (and gender-associated traits) on FL reading and writing skills in more detail. 2.5.3.3 Model of gender differences in FL learning Because so many studies found girls to be the more successful foreign language learners, López Rúa (2006: 112) designed a model to account for these findings (see also Uhl, 2016, for slight modifications of this model), which is depicted in Figure 4. Success in FL learning, thus, depends on the interaction of neurological, cognitive, affective, social and educational factors: Figure 4: The influence of gender and other variables in FL learning (adopted from López Rúa, 2006: 112). According to López Rúa (2006), it is likely that girls will be more successful than boys in FL learning because girls possess “superior verbal intelligence, high 67 2.5 Gender aptitude, more motivation, a social role of modelling behaviour and support‐ ing communication, and the assumption of tasks requiring verbal interaction (teaching, child-caring)” (López Rúa, 2006: 112). However, several researchers concluded that successful language learning is a combination of different variables and not gender itself (e.g., Uhl, 2016). Such a view may help to explain why there is no actual consensus regarding gender’s impact on FL performance; because not only do ‘language proficiency’ and ‘gender’ have to be taken into consideration, but so do a variety of additional factors. 2.5.4 The interaction of gender and minority language background in language learning The studies presented so far have dealt with the influence of either gender or language background on the achievements in the majority language (e.g., German) or in the FL (e.g., English). However, some other studies also point to an interaction between the two variables. For example, female migrants have been found to be more likely to acquire higher education than male migrants (e.g., Siegert, 2008). Such a combination of variables has not been studied to a great extent for the school context. Regarding older learners in a regular school context, reanalysis of the PISA data 2009 indicated that in grade 9, boys with a migrant background tended to perform lower on tests of German reading (see Gottburgsen & Gross, 2012, Gross & Gottburgsen, 2013). This group also obtained the lowest values on questionnaires on other factors such as school leaving, educational aspiration, and parental support (Segeritz, Stanat & Walter, 2010). Such an approach, however, has not yet been taken for the regular elementary school context. For the bilingual preschool context, Steinlen, Kersten & Piske (2019) examined the interaction of gender and language background in a study with 124 five-year old children who attended one of seven German-English bilingual preschools in Germany. The results indicated neither any significant effects of such an interaction for the standardized German Sprachentwicklungstest für dreibis fünfjährige Kinder (Grimm, 2001) nor for two receptive FL English tests of grammar and vocabulary. Thus, the authors pointed out that there is a need for research on the relationship between multilingualism, immigration history and social background, to clarify why and under what conditions bilingual preschools may be successful in promoting heterogeneous learner groups by, for example, including variables such as gender and language background. Steinlen (2018a) conducted a similar study with 370 students in grade 4 who attended either a regular English program, a less intensive bilingual program 68 2. Literature review or a partial IM program. The results on FL English tests of reading and writing (as well as of receptive grammar and vocabulary) did not show any significant effects of the interaction between gender and language background for any FL program. That is, regardless of the intensity of the FL input, the differences between multilingual girls and multilingual boys, between monolingual and multilingual girls or between monolingual and multilingual boys were not significant. In regression analyses, the FL teaching program turned out to be the only significant predictor to account for any variance. As this literature review has shown, there is a clear lack of studies examining the interaction of gender and language background (or migration background) in various school contexts. A future study could, for example, examine interac‐ tional effects of gender by language background regarding reading and writing skills in elementary schools in the majority language (here: German). Based on the findings summarized in 2.3 and 2.4, it is likely that the results in regular elementary school programs would be similar to those reported for regular secondary schools (e.g., Gottburgsen & Gross, 2012, Gross & Gottburgsen, 2013), thus yielding significant differences between multilingual girls and multilingual boys, between monolingual and multilingual girls or between monolingual and multilingual boys. Such interactional effects would, however, not be predicted for the bilingual elementary school context, because prior studies did not report any significant differences between girls and boys or between minority and majority language children (e.g., Steinlen & Piske, 2013, 2015, 2016a, 2018a, 2020; Steinlen, 2016, 2018b; Yadollahi et al., 2020). 2.6 Cognitive variables (with a focus on nonverbal intelligence) The individual cognitive abilities of each child play a very important role in the acquisition of languages. For reading and writing, these include, among others, working memory (i.e., how well learners can remember something), short-term phonological memory (i.e., the ability to remember linguistic sounds), phono‐ logical awareness (i.e., the ability to be aware of the differences in linguistic sounds), and verbal and nonverbal intelligence (e.g., Berninger, Cartwright, Yates, Swanson, & Abbot, 1996). Many studies have examined the relationship between different cognitive abilities and learners’ language skills, finding bi-di‐ rectional effects: language learning affects cognitive abilities, and conversely, cognitive abilities also affect language learning. These findings are obvious, as 69 2.6 Cognitive variables (with a focus on nonverbal intelligence) linguistic and cognitive abilities develop in interdependence (see Kersten, 2019, for a review). The interaction between linguistic and cognitive abilities is important when comparing students in bilingual and mainstream programs. In many studies it has been noted that students in bilingual (especially IM) programs receive better scores on cognitive tests than their peers in mainstream programs (e.g., Baumert et al., 2017; Gebauer et al., 2012; Zaunbauer & Möller, 2007, 2010; Zaunbauer et al., 2005 for Germany, and see Deventer et al. (i. pr.) for a meta-analysis). One may therefore expect students in bilingual programs to outperform students in mainstream programs on any language test. This is the reason why cognitive abilities are usually entered as covariates when comparing students’ language proficiency in mainstream and bilingual programs (e.g., Baumert et al., 2017; Gebauer et al., 2012; Zaunbauer & Möller, 2007, 2010; Zaunbauer et al., 2005). As a proxy for cognitive abilities, many studies (e.g., Bialystok, Peets & Moreno, 2014; Edele, Kempert & Schotte, 2018; Gebauer et al., 2012, 2013; Genesee & Jared, 2008; Genesee, 1987; Jared, Levy, Cormier & Wade-Woolley, 2011; Zaunbauer & Möller, 2007, 2010; Zaunbauer et al., 2005) use a nonverbal intelligence test, which is an approach that is also followed in Steinlen’s studies (Steinlen, 2016, 2017, 2018a, b; Steinlen & Gerdes, 2015; Steinlen & Piske, 2013, 2015, 2016a, b, 2018a; Yadollahi et al., 2020), including the present one. Nonverbal intelligence generally describes thinking skills and problem-solv‐ ing abilities that do not fundamentally require verbal language production or comprehension, and involves the manipulation or problem solving of visual information. The amount of internalized, abstract, or conceptual reasoning or motor skills that are required to complete a task may vary (e.g., Kuschner, 2013). For example, in the different versions of the Progressive Matrices (Raven, 1976a, b), which have also been employed in the present study, the subject’s task is to find a missing piece in a geometrical pattern, using one of six or eight possible alternatives. Nonverbal intelligence as a variable was chosen because approximately 30-60% of the children attending the elementary schools in this study had a minority language background, and it was not clear in the beginning of the project how well-developed these children’s German skills actually were in grade 1, thus ruling out any intelligence test based on language. 2.6.1 Nonverbal intelligence and bilingual children Among the first to employ nonverbal intelligence tests were Peal & Lambert (1962), who examined the relationship between bilingualism and different measures of intelligence of children at age 10; they found bilingual French-Eng‐ 70 2. Literature review lish children to outperform monolingual English children on many tests in‐ cluding tests of nonverbal intelligence. The authors concluded that bilingual French-English children may have developed more flexibility in thinking as they regularly switch from one language to the other. Several years later, Hakuta & Diaz (1985) administered the Raven Progressive Matrices (among other tests) at the kindergarten level and in grade 1 to bilingual Spanish-English children in a bilingual English-Spanish program, and to monolingual English children in a mainstream English program in the United States. They found the degree of bilingualism to be related to nonverbal cognitive ability and suggested that “bilinguals’ greater sensitivity to linguistic structure and detail is then transferred and generalized to different verbal and nonverbal tasks” (Hakuta & Diaz, 1985: 340). The development of nonverbal intelligence in two groups of students in mainstream and bilingual (immersion) elementary school programs was one of the foci of a study by Zaunbauer & Möller (2010). Although both groups obtained test scores that differed significantly at the end of grades 1 and 2, nonverbal intelligence, nevertheless, increased within that year at the same rate. As this developmental pattern is similar for the two groups, Zaunbauer & Möller (2010), therefore, disagreed with the recurring claim that bilingualism may be conducive for the development of cognitive abilities (see Blom, Boerma, Bosma, Cornips & Everaert, 2017; Hakuta & Diaz, 1985), at least with respect to nonverbal intelligence. As Cummins pointed out in 1981, there may be a certain threshold of second language proficiency necessary before cognitive benefits will develop. More research is needed in order to determine how much L1 and L2/ FL input is actually necessary in order to improve cognitive skills, in particular regarding nonverbal intelligence. 2.6.2 Literacy skills by nonverbal intelligence Because nonverbal intelligence has been mainly used as a background variable in most studies, the number of studies that directly examine the relationship between nonverbal intelligence and literacy skills is rather small. Nonverbal intelligence has been found to be a predictor for L1 reading (Gebauer et al., 2012; Morvay, 2015; Naglieri, 2001; Stanovich, Cunningham & Feeman, 1984; Zaunbauer & Möller, 2010). In addition, students with higher nonverbal intelligence achieved better scores in L1 writing tests than their peers with lower nonverbal intelligence (Zaunbauer & Möller, 2010). Similar results, but on nonverbal intelligence effects on FL reading, were reported by Morvay (2015). She tested 65 Hungarian 12 th graders (who had been 71 2.6 Cognitive variables (with a focus on nonverbal intelligence) studying English-as-a-subject in school for eight years) on FL English reading comprehension as well as on nonverbal intelligence and found a significant correlation between the two, albeit not a particularly strong one. This relation‐ ship between (FL) literacy skills and nonverbal intelligence in elementary school is, however, not the focus of the present study and is examined elsewhere in more detail (Steinlen & Piske, i. pr.). In conclusion, any study on language development should include cognitive data when comparing groups. Nonverbal intelligence as a covariate has been used in many recent studies. Future studies should also examine the effects of other cognitive variables on academic and language development in schools with different FL intensity, for example working memory, short-term phono‐ logical memory, phonological awareness, executive functioning, and verbal intelligence, which are, however, not within the scope of the present study. 2.7 Social background With regard to (foreign) language learning, children’s social environment is another important area of influence, particularly in terms of their family’s back‐ ground. The socio-economic status (SES) is an aggregate measure intended to capture social standing, which is often estimated by identifying an individual’s income, educational attainment, and job status, i.e., his/ her social, cultural and economic resources. In childhood, family SES can be estimated by measuring these variables through parent reports of household income and education (e.g., Sheridan, Sarsour, Jutte, D’Esposito & Boyce, 2012, Avvizati, 2020). Cultural resources (for example the number of books in a household) are often also included as variables (see also Kersten, 2019). There is a general consensus that a disadvantage in the educational system due to social status arises early in life (see e.g., Sachverständigenrat Deutscher Stiftungen für Integration und Migration, 2016: 7 f.). This is often intensified during school time by the interaction of several factors, such as the development of family competences, the educational system and/ or by cost-benefit consid‐ erations in educational decisions. The theory of elementary and secondary effects of social origin deals with these factors (Boudon 1974: 29-30) and is still considered to be one of the most influential structural-individualistic models for explaining social inequality. Elementary social disparity effects relate to children at home being stimulated to learn to different extents and being supported to varying degrees. For Germany, it has been reported that less educated parents are less likely to read to their young children (e.g., Biedinger, 2009) and that they 72 2. Literature review motivate them less often to learn for school (e.g., McElvany, Becker & Lüdtke, 2009). Thus, different learning opportunities at home and family socialization not only affect the learning conditions at the start of school, but also the development of students’ competencies throughout their school years (see e.g., the results for L1 reading in PIRLS-2016: Hussmann et al., 2017, and in PISA-2018: OECD, 2019). Secondary social disparity effects include educational decisions in families, which may strongly depend on social and ethnic origin. Ditton, Krüsken & Schauenberg (2005), for example, showed for Germany that when choosing a particular school or when transitioning from elementary to secondary school, parents - and also older children - weigh the costs and benefits of the different educational pathways against each other. It is not only important how well a child performs at school and what expectations the parents have of their children’s education and later professional status, but also how much parents can and want to invest in their children’s education (e.g., a longer journey to school, tuition fees, support with homework, etc.). Finally, the parents themselves play an important role, for example when choosing a school based on whether they are more distant from or close to the school culture. Thus, educational inequality may increase at points of educational transitions (e.g., Sachverständigenrat Deutscher Stiftungen für Integration und Migration, 2016: 8). The results of many studies so far have also indicated a relationship between SES and cognition. For example, more cognitively-stimulating activities and more complex linguistic stimulation with children are used more often in fam‐ ilies with high SES than in families with low SES (Sheridan et al., 2012, see also chapter 2.6). Studies also found effects of SES on the child’s cognitive abilities, such as storing information in working memory, planning, solving problems and switching between different tasks (e.g., Kishiyama, Boyce, Jimenez, Perry & Knight, 2009; Sheridan et al., 2012). Indeed, confounds between SES and cognitive abilities have been reported in many studies. Consequently, these variables are usually used as covariates in statistical analyses to control for such confounds when comparing different groups of students (e.g., Baumert et al., 2017; Bialystok et al., 2014; Gebauer et al., 2012; Reeder et al., 1999; Rolff, Leucht & Rösner, 2008; Steinlen, 2016, 2017, 2018a, b; Steinlen & Piske, 2016a, b; 2018a; Swain & Lapkin, 1982; Thomas & Collier, 2002; Wilden & Porsch, 2015; Zaunbauer & Möller, 2007, 2010; Zaunbauer et al., 2005), and this approach is also taken in the present study (see chapter 5). Many studies have also noted that bilingual (IM) programs in Germany mainly attract parents with a higher SES and educational background, and are therefore often characterized as “elitist programs” (e.g., Apsel, 2012; Baumert 73 2.7 Social background et al., 2012; Genesee, 1987; Kuska et al., 2010; Möller et al., 2017; Möller et al., 2018; Zaunbauer & Möller, 2007; Zydatiss, 2007, see also chapter 2.3.4). For the outcomes of the present study, it is expected that the social and educational selectivity of bilingual (IM) programs may also be accompanied by a performance-based selection. 2.7.1 L1 literacy skills by SES Many studies in Germany and elsewhere have shown that the development of students’ academic competence depends to a large degree on their parents’ SES. For example, students from families with high SES achieved higher levels of reading competence in German compared to students with a lower SES background (e.g., Krüger, Budde, Kramer & Rabe-Kleberg, 2010, for a review on students’ performance in German tests, ranging from kindergarten to university). This applies to elementary school (e.g., Haag et al., 2017; Hussmann et al., 2017: PIRLS-2016) as well as to secondary school (e.g., Rolff et al. 2008; Stanat et al. 2016). The same finding has been reported in international studies on reading tests for regular elementary and secondary school programs (e.g., Mullis et al., 2017: PIRLS-2016; OECD, 2019: PISA-2018). Similarly, a high SES may also positively affect elementary schoolers’ writing skills in the majority language (e.g., May, 2006; Zöller & Roos, 2009 for German), and such effects have also been attested in other studies outside of Germany which examined L1 writing skills in mainstream elementary schools (e.g., Abbott, Berninger & Fayol, 2010; Ehri, 2003; Kennedy et al., 2012). However, as Haag et al. (2017) pointed out, writing in general and orthography in particular seem to be less affected by SES than reading. This may be due to the fact that parents at home generally carry out more reading activities than writing activities. Therefore, writing skills are primarily trained at school (and not at home) and are less susceptible to SES and the frequency of literacy activities engaged in by parents and children (but see Zöller & Roos, 2009, who showed similar effects of SES for reading and writing). 2.7.2 FL literacy skills by SES Inconsistent results have been reported regarding possible effects of SES on FL skills for both regular and bilingual programs. For regular programs, some studies reported a higher SES to positively affect students’ FL performance (e.g., Gardner, Masgoret & Tremblay, 1999; Maluch et al., 2015; May, 2006; Sorenson Duncan & Paradis, 2018; Wilden & Porsch, 2015). However, Rolff et al. (2008) 74 2. Literature review showed in the DESI study that parental attitudes towards the use of the FL language at home influenced the results of FL C-tests in grade 9 more positively than SES, parental educational background or cultural capital (as indexed by the number of books at home). The authors also pointed out that the development of FL skills was less affected by SES than the development of the majority language German. For the bilingual (immersion) context, the interaction of SES on FL compe‐ tence is even less clear. On the one hand, Kim, Hutchison & Winsler (2013) reported that in 90/ 10 programs in the U.S., high SES English students performed better on Spanish reading tests than their English peers with a low SES. On the other hand, Baumert et al. (2017) found that in two-way immersion programs in Berlin, SES and parental educational background did not affect students’ reading skills in the partner (FL) language, in contrast to their reading skills in the majority language German (see also Rolff et al., 2008; Fleckenstein et al., 2017 for ninth graders). Likewise, Zaunbauer et al. (2012) did not find any significant influence of the covariate SES on fourth graders’ English vocabulary, reading fluency or reading comprehension. Similar results were also reported from Canada (e.g., Genesee, 1987, 2004; Genesee & Fortune, 2014; Holobow, Genesee & Lambert, 1991) and the United States (e.g., Lindholm-Leary, 2001; Lindholm-Leary & Borsato, 2006; see also Thomas & Collier, 2017). Note that the studies reviewed so far have examined the impact of SES on FL reading. The impact of SES on FL writing has apparently not been explored in this context yet. 2.7.3 The interaction of SES and language background on literacy skills In Germany and elsewhere, an immigrant status, where individuals come from diverse linguistic and ethnic backgrounds, has often been found to coincide with a relatively low SES (e.g., Sachverständigenrat Deutscher Stiftungen für Integration und Migration, 2016). In many studies, individuals with a low SES and an immigrant background achieved lower educational qualifications and lower linguistic (and mathematical) competencies (see e.g., Gottburgsen & Gross, 2012, for a review). For the school context, Jungbauer-Gans & Gross (2011) noted that students with an immigrant background are particularly likely to attend schools with a low proportion of parents having a high educational background, as well as a high proportion of migrants with low SES. Such children are often considered ‘at-risk’ students (e.g., Genesee & Fortune, 2014: 196). The interaction between 75 2.7 Social background low SES and immigrant status has also been shown in many studies on German reading skills in regular elementary schools (e.g., Bos & Pietsch, 2006; Chudaske, 2012; Haag et al., 2017; Hussmann et al., 2017; Schwippert et al., 2012), and, of course, in regular secondary school programs (e.g., Haag et al., 2016; Hesse et al., 2008; OECD, 2019). However, effects of language background (or immigrant status, see chapter 2.4.1 for a critical discussion of these terms) often disappeared (or at least decreased) once the data on minority language students had been controlled for SES (e.g., Stanat et al.; 2016; Wilden & Porsch, 2015). In Germany, recall that bilingual school programs have often been charac‐ terized as a (highly) selective school form, in particular attracting immigrant families with higher SES, as Baumert et al. (2017) point out. This also applies to the partial IM program in an elementary school in Tübingen, where the SES of minority and majority language students (as operationalized by parents’ selfestimated wealth and/ or parental educational background) did not differ significantly. In this particular 50/ 50 program, parents had a relatively high educational background (corresponding to a university entrance certificate) and considered their economic background as slightly above average in terms of wealth. This may have been one of the reasons why minority language children scored as well as majority language children on German and English tests of reading and writing (Steinlen, 2016, 2017, 2018a, b; Steinlen & Piske, 2013, 2015, 2016a, b, 2018a, b, 2020). However, such effects have not been studied for bilingual programs with higher or lesser FL intensity. As this review clearly showed, any study on majority and minority students’ language skills should include information on their socio-economic background, especially on their social, cultural and economic resources as well as on cogni‐ tively and linguistically stimulating activities at home, in order to disentangle possible confounds of SES, language and migration background in tests on language/ academic achievement in any school program. 76 2. Literature review 3. Research questions In the following, a study is presented which examines the effect of FL intensity on German and English reading and writing skills of elementary school students in grade 4. This topic has not been addressed in previous studies because those have rather focused on the development of German and English reading and writing skills in a partial IM Bili-50 program, where half of the teaching time was conducted in English (Steinlen & Piske 2013; Steinlen, 2016, 2017, 2018a). In Steinlen (2018b), program differences due to the intensity of FL input were mentioned in a footnote, and Steinlen & Piske (2018b) restricted their analysis to students’ literacy skills in a regular program and a Bili-50 program. The present study compares four elementary school programs differing in FL intensity. In the EFL program, English is taught as a subject for two lessons per week. One subject, i.e., science, is taught in English in the so-called Bili-20 program. In the Bili-50 program, the teachers teach several subjects (i.e., MeNuK: Mensch - Natur - Kultur, corresponding to science, arts, music, and PE) in English. Finally, almost all subjects in the Bili-70 program are offered in English, most of them are taught by native speakers of English, and the children obligatorily attend extra-curricular activities which are also carried out in English. While prior studies have compared students in regular and bilingual (partial IM) programs, studies comparing bilingual programs with different degrees of intensity (particularly on the lower end of intensity) are still rare. The focus of the present analysis is on literacy (i.e. reading and writing) skills in German and English in grade 4. Whether factors such as the children’s gender, cognitive and social background, or their language background affect reading and writing skills in German and English, is also examined. Based on the literature review, the following research questions are ad‐ dressed: (1) Are there any between-group effects on German and English reading and writing in programs with different degrees of FL intensity at the end of grade 4? (2) Are there any within-group effects on German and English reading and writing based on children’s gender, language background or cognitive and social background at the end of grade 4? Between-group effects are expected for English reading and writing skills, due to the different degrees of FL intensity of each program (e.g., Genesee, 2004; Möller et al., 2017; Pérez-Cañado, 2012; Wesche, 2002; Zaunbauer et al., 2012, and see chapter 2.3.11 and 2.3.12 for a review). For German reading and writing skills, between-program differences are not expected because it has been reported in many studies that the development of the majority language (often the children’s L1) is not negatively affected in bilingual programs (e.g., Baumert et al., 2017; Gebauer et al. 2012, 2013; Genesee, 1987, 2004; Genesee & Jared, 2008; Lambert & Tucker, 1972; Zaunbauer et al., 2005, 2013, Zaunbauer & Möller, 2006, 2007, 2010, see chapter 2.3.10 for a review). For gender, within-group effects are expected for both the German and the English tests because in many studies, girls outperformed boys (e.g., BIG-Kreis, 2015; Hussmann et al., 2017; Kennedy et al., 2012; Schipolowsky et al., 2017, see chapter 2.5 for a review). Similarly, effects of language background (i.e., minority vs. majority language background, see chapter 2.4 for a review) are expected for majority language students outperforming their minority language peers on German tests (e.g., Bos & Pietsch, 2006; Chudaske, 2012; Dollmann, 2010; Elsner, 2007; Haag et al., 2017; Hussmann et al., 2017; May, 2006) but not on English tests (e.g., Steinlen & Piske, 2013, 2014, 2015, 2016, 2018, 2019; Steinlen, 2016, 2017, 2018). Finally, children’s cognitive and social backgrounds are expected to affect their reading and writing skills in German and English both within a program as well as across programs, as higher SES and better cognitive skills result in better outcomes on any academic test (see chapters 2.6 and 2.7 for reviews). Taking into consideration various background factors, the purpose of the present study is to extend the previous studies by Steinlen (e.g., 2016, 2017, 2018 a, b; Steinlen & Piske, 2013, 2015, 2016, 2018a, 2020) in terms of sample size and number of FL programs in order to provide a more concise picture of the (foreign) language level that students in various FL programs can attain at the end of grade 4. 78 3. Research questions 4. Method This chapter presents the schools, the test materials and the subjects. Chapter 4.1 introduces the three schools, which participated in this project, which offered four FL programs in total. The test material is illustrated in chapter 4.2. Control measures include a cognitive test as well as a parent questionnaire, which provided information on the children’s social background. Reading and writing tests in German and English are also presented in chapter 4.2. More information about the subjects is provided in chapter 4.3. 4.1 The schools and their programs The data regarding the regular foreign language program (EFL) and the partial immersion program (Bili-50) were collected at an elementary school in the south of Germany (Hügelschule in Tübingen, Baden-Württemberg). It is a district school that has offered both a music program and a bilingual program since 2008/ 09, with one cohort per year (e.g., Steinlen, 2016, 2017, 2018a, b; Steinlen & Piske, 2013, 2015, 2016a, b, 2018a, b, 2020, Tamm, 2010, Yadollahi et al., 2020). Testing in this school took place between 2012 and 2019. A whole-day program in this school started in 2015, with lessons taking place until 3: 30 p. m.; prior to that school ended at 12: 30 p. m. The other two schools offering a less intensive and a more intensive bilingual program, respectively, are presented in chapters 4.1.3 and 4.1.4. 4.1.1 The Bili-50 program In the bilingual track of the Hügelschule all subjects are taught in English, except for German-as-a-subject religious education and mathematics. That is, 50 % of the teaching time is conducted in English. Thus, the program can be characterized as a partial IM program (following Canadian terminology), labelled Bili-50. Many of the teachers, who all have a German background, obtained their de‐ gree as part of a European Studies Program (“Europalehramt”) at the Universities of Education in Freiburg or Karlsruhe. They only speak English in class, although technical terms are always provided in English and German. The teachers follow the curricular guidelines for the subjects, independent of the language that is used in the classroom (e.g., Ministerium Baden-Württemberg für Kultus, Jugend und Sport, 2016b). For MeNuK (Fächerverbund Mensch, Natur und Kultur, combining science and culture), the Ministerium Baden-Württemberg für Kultus, Jugend und Sport (2008) described English competences at the end of grade 4 briefly in terms of lexical knowledge, world knowledge and writing skills, although these competences are similar to the standards for English-as-a-subject lessons (cf. Ministerium Baden-Württemberg für Kultus, Jugend und Sport, 2004, 2016a). As in other German elementary schools, the children in this Bili-50 program receive initial literacy instruction in German first, from grade 1 onwards. Reading and writing skills in English are acquired incidentally and, according to the teachers, are not taught systematically, although the English writing system is present from the start. In the middle of grade 1, students start to read and write single English words. Different literacy activities focusing on the content of the respective subjects being taught in English are carried out from grade 2 onwards (e.g., silent reading, reading aloud in groups and in class, reading comprehension checks on texts, etc.). Students start to write longer English texts from grade 3 onwards, but there is no specific focus on spelling errors. A few notes on spelling: in the initial phase of the Bili-50 program, the teachers were not sure to what extent they should address spelling issues in class, as this is not part of the curriculum (e.g., Ministerium Baden-Württemberg für Kultus, Jugend und Sport, 2004). However, experience soon showed that older students in grades 3 and 4 asked for certain spellings and expected specific answers. The teachers also felt that specific activities targeting certain aspects of spelling seemed to help the children internalize the correct spellings in question. For this reason, certain aspects of spelling are now included in the school curriculum, including sight words (i.e. high frequency words in texts). The topics are always embedded in the context of the teaching of a particular subject in the FL and are offered either at the request of the students or because the teacher notices that without addressing a particular spelling problem there is a risk of fossilization (i.e., the wrong language becomes habit and cannot be easily corrected, e.g., Ellis, 2008). 4.1.2 The EFL-program In the music track of the same school in Tübingen (Germany), which offers additional lessons on flute and percussion, English was taught as a subject from grade 1 onwards with two hours per week, adhering to the curriculum of the Federal State of Baden-Württemberg (Ministerium Baden-Württemberg für 80 4. Method Kultus, Jugend und Sport, 2004, 2016a). This program is abbreviated EFL (Early Foreign Language). As a side note, English-as-a-subject has started at grade 3 from 2018 onwards (Ministerium Baden-Württemberg für Kultus, Jugend und Sport, 2020). This, however, does not affect the data presented in this study, which includes cohorts of fourth graders from 2012 to 2019 who had attended the EFL program from grade 1. In contrast to many other elementary schools, the English teachers (who have a German background) all hold relevant qualifications (that is, they do not teach English “fachfremd”). Some of the teachers have also been working in the bilingual program (e.g., Steinlen & Piske, 2018a). The curriculum of the Federal State of Baden-Württemberg (Ministerium Baden-Württemberg für Kultus, Jugend und Sport, 2004, 2016a) for Eng‐ lish-as-a-subject is adhered to. That is, students are to obtain level A1 as defined in the Common European Framework of Reference (Council of Europe, 2001, 2018) in English reading, writing, listening and speaking at the end of elementary school in grade 4, with a focus on oral skills in the first two grades. To achieve this aim, the teachers in the EFL program work with the “Sally” series (e.g., Bredenbröcker, Brune & Elsner, 2015), which is often also supplemented with additional material specially designed for their class. 4.1.3 The Bili-20 program The second school is an elementary district school in the north of Germany (Muhlius Schule in Kiel, Schleswig-Holstein). In this school, only the subject science (Heimat- und Sachunterricht) is taught in English from grade 1 onwards (Steinlen & Gerdes, 2015), with two hours a week in grade 1 and 2 (i.e., corresponding to roughly 10 % of the teaching time), and five hours in grades 3 and 4 (20 % of the teaching time). During the time of testing (2013 to 2015), the lessons in the Muhlius Schule took place from 8: 00 a. m. to 12: 30 p.m, and two cohorts per year attended the school. In the following, this less intensive bilingual program is abbreviated as Bili-20. In the Federal State of Schleswig-Holstein, regular English lessons start with two hours per week in grade 3, and level A1 is expected at the end of grade 4, although reading and writing are subordinate to listening and speaking (Ministerium für Bildung, Wissenschaft, Forschung und Kultur des Landes Schleswig-Holstein, 2007). In this particular school, these English lessons are, in some cases, devoted to science as well, especially in grades 3 and 4. The curriculum of Schleswig-Holstein for the subject science was generally adhered to (Ministerium für Bildung, Wissenschaft, Forschung und Kultur des Landes 81 4.1 The schools and their programs Schleswig-Holstein, 1997). However, not all topics were taught in English by all teachers (for example the topic “Schleswig-Holstein” and the obligatory test for riding bicycles in grade 4). When teachers taught the subject science in English, they mainly used material that they designed themselves. For the English-as-subject lessons, they relied on the “Sally” series (e.g., Bredenbröcker et al. 2015). At the time of testing, the teachers in the Bili-20 program had spent at least three years or more in an English-speaking country. 4.1.4 The Bili-70 program The fourth teaching program is offered by a private elementary school in Berlin (Platanus Schule). Here, several subjects (i.e. math, science, music, arts, etc.) are taught in English, often by English native speakers from different English-speaking countries (e.g., Australia, United Kingdom, and the United States) or by teachers with a high competence in English at the C2 level. Usually one or two students in each class have an English-speaking background (however, these students are excluded from the following analyses). At the time of data collection (2016-2018), two cohorts per year attended the program. Team teaching is an important aspect of many lessons, i.e., some lessons are staffed by English-English, German-English (e.g., math, science, English, PE) or by German-German teams (for German-as-a-subject). In addition, obligatory extra-curricular activities are carried out by native English preschool teachers, English teachers and German-speaking staff, i.e. students attend school from 8: 00 a. m. to 4: 00 p. m. Thus, at least 70 % (if not more) of the teaching time is conducted in English (e.g., Steinlen & Piske, 2018c). Accordingly, the program is called Bili-70. The Platanus Schule in Berlin has developed a concept of foreign-language literacy. The students in grade 1 learn the alphabet in German first. From the second half of the first grade onwards, the Phonics program “Letters and Sounds” is used to train English sounds and sound combinations and their graphemic representations (e.g., Department for Education and Skills, 2007). This concept, which has been implemented in the school program, also involves a strong focus on the use of English books and materials with different levels of difficulty, topics and genres. Thus, FL literacy activities are carried out on a regular basis for all students, independent of their grade. 82 4. Method 4.2 Test material The students were tested approximately six weeks before the end of the school year. At the same time, the teachers handed out the parent questionnaires and collected them. Participation in the project was voluntary, and parents gave written consent for their children to take part in the study. 4.2.1 Control measures: Family background and cognitive background In the questionnaire, the parents provided information on their child’s age, her/ his country of birth, the language/ s used at home and their educational background, with “1” corresponding to no school certificate and “6” to a university entrance certificate (following Zaunbauer et al., 2012). Altogether, 345 parents filled out the parental questionnaire, which corresponds to a response rate of 70.8 % (i.e., 71.4 % for the EFL, 53.3 % for the Bili-20, 76.6 % for the Bili-50, and 63.3 % for the Bili-70 program). Generally, a 75 %-80 % response rate is considered acceptable (e.g., Draugalis, Coons & Plaza, 2008). According to the teachers, non-responses were usually due to skepticism about surveys in general and/ or parental time constraints, which may also account for the fact that not all questions were answered by all respondents. More information on educational background was available for mothers than for fathers, which is why in the following, education pertains to maternal education. A cognitive variable was examined as a control variable in this study because test scores pertaining to linguistic or academic achievements may be affected by children’s cognitive skills (e.g., Bleakley & Chin, 2004; Chudaske, 2012; Gamsjäger & Sauer, 1996 and see chapter 2.6). The focus was on general non‐ verbal intelligence, which was assessed by using the SPM (Standard Progressive Matrices, Raven, 1976) in grade 4. Using one of six or eight possible alternatives, the children’s task was to complete an incomplete geometrical pattern. The test consisted of 48 items, which were presented in four sets of 12, in increasing order of difficulty within each set. 20-30 minutes were allocated for the test. The publishers (Heller et al., 1998) reported the internal consistencies to lie at r=0.81 for German elementary school children, and the SPM to be a good indicator for Spearman’s g-factor (which generally yields satisfying correlations with school performance tests). 83 4.2 Test material 4.2.2 English tests The following English tests were used to determine English reading and writing skills. The PSAK (Primary School Assessment Kit, Little et al., 2003) was originally designed for assessing immigrant children’s English language skills in Ireland, that is, for children who learn English as an additional language. The PSAK is comprised of four parts: writing, reading, speaking and listening. The test was chosen because grading is carried out in relation to the three levels used in the English Language Proficiency Benchmarks - A1, A2 and B1 (Council of Europe, 2001). The lowest level is A1, i.e., for students with little existing English language proficiency, and the highest level is B1. Little et al. (2003) noted that full integration into mainstream programs (i.e., with monolingual English students and teachers) is possible when students are capable of performing at B1 in all assessment tasks. All sets begin at level A1 and continue up to level B1. When this project started, it was hoped that these levels would cover FL competences by both EFL students (due to the expectations set by the KMK, 2013a) and by students in intensive bilingual programs alike. The PSAK was also chosen because of its colorful design, which, according to many students in the project, motivated them to work through the booklet. The reading part of the PSAK (called PSAK-R) consisted of nine subtests. In A1.1 children drew a line from each word to the matching picture (word-level); in A1.2 the children read small phrases (i.e., a green pencil) and drew a circle around the matching picture (sentence level). Afterwards, children looked at a picture, read sentences related to the picture and chose “yes” or “no” for each sentence (A1.3 and A2.1). In A2.2 and A2.3, children saw six pictures which they had to match with 8 sentences (sentence level). B1.1 was a text comprehension task with multiple choice questions. B1.2 related to grammar, and children read a text and circled alternative words provided in boxes. The final task (B1.3) required children to answer additional questions on the two previous texts, ticking off possible alternatives (text level). On the whole, 45 points could be obtained on the reading test, with 45-39 points corresponding to level B1, 38-22 points to A2 and scores below 21 points to A1. The first two sub-tests of the writing part (PSAK-W) were based on picture prompts, i.e., matching words and pictures by copying the appropriate word from a box and filling out gap texts (A1 and A2). The focus here was on vocabulary and spelling. The subtest B1 required free writing. The children’s task was to write five sentences on activities they had carried out the day before (the topic “Yesterday”). The texts were then evaluated for vocabulary, grammar, spelling, and punctuation. A total of 39 points could be obtained. 13 points or more were required for level A1, 22 for A2, and 35 for B1. 84 4. Method As test quality criteria are not available for the PSAK, Steinlen & Piske (2016a) computed some measures based on a sample of 94 children in grade 4 and found split half reliability values of .85 and .80 (Cronbach’s alpha), yielding a satisfying result for internal consistency. Furthermore, the PSAK-R in grade 4 correlated moderately (between .56 and .57) with other English reading tests (i.e., with the Test of Reading Comprehension, Mossenson et al. 2003 and the Test of Silent Word Reading Fluency, Mather et al., 2004, respectively, which were conducted with the same children at the same time). Due to the fact that additional writing test formats in English, teacher evaluations, or children’s grades for English-as-a-subject or other subjects were not available, correlations with the PSAK-W and other English tests could not be computed. Based on the sample of 94 children (see Steinlen, 2018b), the PSAK-W exhibited split half reliability values of .68 and .42. The second low value mainly relates to the free writing part (B1), paralleling findings from other studies on reliability values of essay tests which also found low inter-rater reliability correlations (e.g., Greenberg, 1992). 4.2.3 German tests German reading comprehension was assessed by means of the ELFE (Ein Leset‐ est für Erstbis Sechstklässler; Lenard & Schneider, 2006). This test measured reading comprehension at word, sentence and text level. At word level, one of four words had to be assigned to a picture. In the sentence comprehension task, students selected one of four alternative words to complete a sentence. At text level, children answered questions on short texts. Altogether, 10-16 minutes were allocated for this test. Based on data collected from 4,893 German students (Lenard & Schneider, 2006), internal consistency reliability estimates of .92 to .97 were reported, depending on the subtest. Furthermore, moderate to high correlations were noted with respect to other reading test formats, teacher evaluations and grades for German (.45 to .71). As in other studies examining minority language children’s performance in German (e.g., Chudaske, 2012) or majority language immersion students’ performance in German (Zaunbauer & Möller 2005, 2007; Zaunbauer et al., 2012), spelling skills were tested using the standardized test HSP (Hamburger Schreibprobe, May, 2002). As a measure of spelling ability, the number of correctly spelled words was computed. Internal consistency reliability estimates of .93 to .98 were reported (the normative sample consisted of 23,000 students from grades 1-9) and high correlations with other reading test formats, teacher evaluations and grades for German-as-a-subject (.72 to .85). 42 points could be 85 4.2 Test material reached in grade 4; 16 words and five sentences had to be written down. For the sake of brevity, the HSP will be referred to as a writing test, despite the fact that the focus is on spelling and punctuation only. Note that the German and English test formats for reading and writing are not equivalent. For example, although the HSP and the PSAK-W are both writing tests, they use different methods, i.e., the HSP focuses on orthography whereas the PSAK-W is not only concerned with spelling and punctuation, but also with grammar and vocabulary. Similarly, the two reading comprehension tasks differ in their design, in particular with respect to sentence comprehension. The PSAK-R employs pictures at a sentence-based level, whereas the ELFE does not. These differences may have ramifications for the outcomes of this study. 4.3 Subjects Altogether, a total of 487 children participated in the present study. However, the data sets for each individual child are not always complete. The reasons are manifold, ranging from sickness and school-related activities to parent questionnaires that were not (completely) filled out. In the Hügelschule (Tübingen), the data of the first cohort of students in grades 1-4 were collected in 2012, and each year, another cohort was added, resulting in eight cohorts (as of 2019 and on-going). Thus, the sample size increases by publication date, with more recent samples containing cohorts from earlier studies (Steinlen, 2016, 2017, 2018a, b; Steinlen & Piske, 2013, 2015, 2016a, b, 2018a, b, 2020). Children in the Muhlius Schule (Kiel) were tested from 2012-2015, resulting in three cohorts with two classes each. Testing took place three times (2016-2018) in the Platanus Schule (Berlin), with two classes in two cohorts. The following paragraphs provide information on the number of girls and boys in each program, their ages, as well as their language background, i.e., which language/ s the children spoke at home. To facilitate analyses, language background was coded as a binary variable (majority/ minority language back‐ ground). 140 students (50 % girls and 50 % boys) of the EFL program participated in the tests. At the time of testing, the children were on average 10.6 years old (SD = 7.9 months). 57 % of the children used other languages and German at home, and 43 % of the children only spoke German at home. 37 girls and 23 boys had a monolingual German background, and 33 girls and 47 boys a multilingual background. Among others, parents named Arabic, Croatian, 86 4. Method 12 Of the three schools, the Hügelschule is the only school which offers Turkish-as-a-sub‐ ject to children with a Turkish background, with one lesson per week from grade 1 onwards. To our knowledge, the other minority language children were not formally instructed in their respective family languages. Dutch, Greek, Italian, Kurdish, Romanian, Russian, Spanish, Tamil, Turkish, Twi and Vietnamese as family languages. In the Bili-20 program, 137 children (38 % girls and 62 % boys) took the tests. On average, at the time of testing the fourth graders were 10.4 years old (SD = 8.3 months). 45 of the children used other languages as well as German at home (33 %), and 92 children spoke only German at home (67 %). 38 girls and 54 boys had a monolingual German background, and 13 girls and 32 boys a multilingual background. Some of the family languages that were mentioned were Albanian, Arabic, Kotokoli, Kurdish, Latvian, Mandarin, Pashtu, Persian, Russian, Spanish, Turkish, Twi and Urdu. A total of 150 children from the Bili-50 program participated in the tests, half of them girls and the other half boys. On average, they were 10.3 years old at the time of testing at the end of grade 4 (SD = 5.8 months). 81 of the children used other languages and German at home (54 %), and 69 of the children spoke exclusively German at home (46 %). 29 girls and 40 boys had a monolingual German (majority language) background, and 46 girls and 35 boys a multilingual (minority language) background. The family languages included Arabic, French, Greek, Kiswahili, Kurdish, Persian, Russian, Serbo-Croatian, Spanish, and Turkish. 12 Finally, 90 fourth graders (52 % girls and 48 % boys) attending the Bili-70 program took the tests. They were, on average, 10.0 years old at the time of testing (SD = 4.0 months). 65 of these children (35 girls and 30 boys) had a monolingual German background, and 25 children (12 girls and 13 boys) had a multilingual background. Family languages included Arabic, Czech, Danish, Dutch, French, Greek, Gujrathi, Hindi, Lithuanian, Mandarin, Mongolian, Per‐ sian, Polish, Portuguese, Russian, Swedish, Serbian, Spanish, Thai, Turkish, Twi, Ukrainian and Vietnamese. In some cases, English was also spoken at home, in addition to the family language. The minority language children had mostly been born in Germany, and they all used their family language plus German at home. The parent questionnaire, unfortunately, did not ask for information concerning the use of the family language and the use of German prior to the children’s enrollment at school. It is, therefore, not clear whether the minority language children had learned German as an L1 or an L2. In informal interviews, however, most parents stated that the family language was their children’s L1, with German being acquired 87 4.3 Subjects in preschool (at age 3) at the latest. The foreign language English is, therefore, the children’s L3 (in some cases even the L4). 88 4. Method 5. Results In this study, SPSS version 25 (2017) was used to compute statistical analy‐ ses. Descriptive results for each school program are reported first, including ANOVAs for within-group differences for each dependent variable (Table 1). This is followed by additional statistical analyses (covariance analyses, regres‐ sion analyses) to examine between-group differences for the language tests. Note that the number of participants per test (N) varies due to sickness or school-related reasons. 5.1 Within-group comparisons Table 1 illustrates the results of each German and English language test as well as the results of the children’s nonverbal intelligence test SPM and of their mothers’ educational background. For each dependent variable, ANOVAs were calculated to examine the effect of children’s language background (i.e., minority vs. majority). In all measures (except for the HSP), the Bili-70 students achieved the highest scores, the children in the EFL and the Bili-20 programs the lowest scores, while the children in the Bili-50 program performed in-between (see Table 3 for statistical analyses to support this impression). Norm values (i.e., age-appropriate values) were obtained for the nonverbal intelligence test SPM by the Bili-70 and Bili-50 group, whereas the scores for Bili-20 and EFL were slightly below average. For the German writing test HSP, age-appropriate values were found for Bili-70, Bili-50 and Bili-20, while the results were slightly below average for EFL. All four groups performed slightly above average in the German reading test ELFE. With regard to the English reading and writing tests, the following levels (according to the CEFR, Council of Europe, 2001, 2018) were reached. The English reading skills of Bili-70 were at level B1, of Bili-50 at the high end of A2, and of the Bili-20 and EFL group at the low end of level A2. English writing skills corresponded to upper A2 level for Bili-70, solid A2 for Bili-50, and level A1 for Bili-20 and EFL. Measure [max. points] Program General results Language background F-values / p values η 2 Majority Minority Education [6]Bili-70 5,8 (0,5) 5,7 (0,5) 5,9 (0,3) F (1, 52) = 0.288, p=0.954 0.041 N = 321 Bili-50 5,3 (1,1) 5,6 (0,8) 5,0 (1,3) F (1, 107) = 7.875, p=0.013*0.066 Bili-20 4,8 (1,2) 4,8 (1,1) 4,8 (1,5) F (1, 67) = 0.015, p=0.904 0.071 EFL 4,8 (1,2) 5,1 (1,1) 4,5 (1,2) F (1, 86) = 5.276, p=0.024* 0.080 SPM [48] Bili-70 41,2 (6,4) 41,9 (5,9) 39,6 (7,3) F (1, 78) = 2.066, p=0155 0.298 N = 405 Bili-50 39,6 (6,9) 41,5 (6,5) 37,9 (6,8) F (1, 108) = 8.095, p=0.005*0.263 Norm: 41-43 Bili-20 37,4 (7,1) 37,6 (7,6) 36,6 (5,6) F (1, 102) = 0.375, p=0.551 0.272 EFL 36,8 (7,4) 38,9 (6,5) 35,2 (7,7) F (1, 109) = 7.330, p=0.008*0.330 HSP [42] Bili-70 29,2 (8,8) 31,0 (8,1) 24,6 (8,9) F (1, 74) = 8.733, p=0.004* 0.504 N = 418 Bili-50 32,4 (7,5) 33,1 (6,5) 31,8 (8,2) F (1, 127) = 0,908, p=0.325 0.212 Norm: 30-33 Bili-20 29,5 (8,8) 29,9 (8,2) 29,3 (1,9) F (1, 101) = 0.479, p=0.491 0.262 EFL 26,5 (8,0) 28,7 (6,7) 24,7 (8,5) F (1, 108) = 7.021, p=0.009*0.399 ELFE [120] Bili-70 96,3 (22,1) 100,5 (19,7) 87,6 (23,6) F (1, 77) = 6.017, p=0.016* 0.502 N = 398 Bili-50 92,0 (20,3) 96,0 (15,5) 88,4 (23,5) F (1, 105) = 3,860, p=0.052 0.546 Norm: 88 Bili-20 87,9 (19,5) 89,4 (20,3) 86,0 (16,9) F (1, 93) = 1.386, p=0.242 0.519 EFL 92,6 (22,9) 99,7 (22,5) 87,5 (21,9) F (1, 114) = 8.462, p=0.004*0.559 PSAK-R [45] Bili-70 42,8 (1,8) 43,0 (1,4) 42,6 (2,5) F (1, 76) = 0.529, p=0.469 0.091 N = 401 Bili-50 38,55 (5,3) 38,8 (4,2) 38,3 (6,1) F (1, 103) = 0,205, p=0.651 0.205 A1 = 13+ Bili-20 27,2 (8,7) 26,4 (8,1) 29,1 (10,1) F (1, 103) = 1.943, p=0.166 0.271 A2 = 28+ B1 = 43+ EFL 29,2 (6,1) 30,9 (5,4) 28,1 (6,3) F (1, 111) = 5.862, p=0.017*0.242 PSAK-W [39] Bili-70 30,2 (5,0) 30,9 (5,0) 28,6 (4,6) F (1, 76) = 3.728, p=0.057 0.505 N = 400 Bili-50 26,7 (5,6) 27,0 (5,9) 26,4 (5,3) F (1, 102) = 0,310, p=0.579 0.245 A1 = 13+ Bili-20 13,6 (7,1) 13,0 (6,6) 15,3 (8,4) F (1, 103) = 2.261, p=0.136 0.221 A2 = 22+ B1 = 35+ EFL 17,4 (4,4) 17,0 (4,7) 17,6 (4,3) F (1, 109) = 0,439, p=0.509 0.230 Table 1: Descriptive statistics for the variable language background in the four pro‐ grams EFL, Bili-20, Bili-50, and Bili-70, regarding the children’s maternal educational background, the nonverbal intelligence test SPM, the German reading test ELFE, the German writing test HSP, the English reading test PSAK-R and the English writing test PSAK-W. The results are reported in raw scores, standard deviations in round brackets, and the maximum score of each test in square brackets. The results for ANOVAs (Bonferroni) include effect sizes eta square (η 2 ). For easier reference, statistically significant results (p<0.05) are marked by an asterisk. Note: A1, A2 and B1 relate to the levels described by the Common European Framework of Reference (Council of Europe, 2001). A comparison of the students’ language background did not yield any signif‐ icant differences between minority and majority language children for the English writing test PSAK-W, independent of the program they attended (see Table 1). For the English reading test PSAK-R, no significant differences 90 5. Results between the two groups of majority and minority language students were noted for Bili-70, Bili-50, and Bili-20. However, majority language students in the EFL program achieved better results than their minority language peers on the English tests. For the German reading and writing tests ELFE and HSP, no significant differences between the majority and minority language groups were noted for Bili-50 and Bili-20, but the majority language students outperformed their minority language peers in the Bili-70 and EFL program on both German tests. 5.2 Between-group comparisons Table 2 shows statistical comparisons of the control variables regarding maternal educational background and children’s nonverbal intelligence across the four FL programs. The results indicate that the four programs differed significantly with respect to these two background variables. The effect sizes are, however, rather small and medium, respectively, i.e. below .01 and .06 (anything >.14 is considered a large effect, as described, for example, by Field, 2013). Measure η 2 Education F (3, 317) = 11.023, *p=0.000 0.011 SPM F (3, 401) = 7.786, *p=0.000 0.075 Table 2: Results of ANOVAs (Bonferroni) of the control variables maternal educational background (“Education”) and nonverbal intelligence (“SPM”) in the four FL programs. As post-hoc tests revealed, significant differences for both maternal education and for nonverbal intelligence (SPM) were found between EFL and Bili-50, between EFL and Bili-70, and between Bili-20 and Bili-70 (p<0.05); the other group comparisons (i.e. EFL vs. Bili-20 and Bili-50, and Bili-20 vs. Bili-50) were not significant (p>0.05). Due to these between-group differences, covariance analyses were calculated with maternal educational background and the children’s nonverbal intelligence as covariates (following e.g., Zaunbauer & Möller, 2007, see also Swain & Lapkin, 1982). The results of the four language tests are illustrated in Table 3. 91 5.2 Between-group comparisons Measure η2 HSP F (3, 239) = 3.973, *p=0.009 0.048 ELFE F (3, 224) = 1.748, p=0.158 0.023 PSAK-R F (3, 221) = 58.954, *p=0.000 0.444 PSAK-W F (3, 215) = 86.096, *p=0.000 0.546 Table 3: Results of covariance analyses of the four language tests conducted in grade 4 in the four FL programs. As Table 3 shows, program as a variable significantly affects almost all language tests (except for the German reading test ELFE) when the effect of the covariates SPM and maternal education are kept constant. Thus, the children indeed obtained significantly different results regarding their English and German reading and writing skills depending on the program they attended. Strong effect sizes are noted for the English reading and writing test, whereas the effect for the German test HSP is rather small. Program-dependent effects are, however, not noted for the German reading test ELFE, in which the children scored similarly in all programs. Additionally, five regression models were run a) to determine the respective influence of different factors on the overall data set, and b) to include the factors as predictors in different models (Table 4). Only complete data sets were used. Because the factor “program” consisted of more than two categories, dummy variables (2-4) were created for the regression analyses (model 5), in which the EFL program served as reference. 92 5. Results Parameter ELFE (N= 229) HSP (N= 244) Model 1 2 3 4 5 1 2 3 4 5 Lg. background Gender -0.182* -0.184* -0.089 -0.118 -0.083 -0.114 -0.085 -0.114 -0.080 -0.168* -0.170* -0.135* -0.096 -0.131* -0.080 -0.142* -0.086 -0.153* SPM 0.336* 0.328* 0.341* 0.318* 0.282* 0.274* Education 0.026 0.028 0.121 0.125 Program Dummy2 Program Dummy3 Program Dummy4 -0.079 -0.146 -0.015 0.131 0.216* -0.012 R 2 corr 0.029 0.033 0.138 0.135 0.142 0.024 0.039 0.131 0.141 0.176 93 5.2 Between-group comparisons Parameter PSAK-R (N= 225) PSAK-W (N= 220) Model 1 2 3 4 5 1 2 3 4 5 Lg. background Gender -0.050 -0.050 0.033 0.008 0.033 0.033 0.018 -0.023 0.016 0.038 0.040 -0.064 0.095 -0.068 0.117 -0.080 0.020 -0.064 SPM 0.332* 0.263* 0.131* 0.298* 0.229* 0.073 Education 0.230* 0.067 0.228* 0.054 Program Dummy2 Program Dummy3 Program Dummy4 -0.092 0.447* 0.562* -0.216 0.442* 0.558* R 2 corr -0.002 -0.005 0.099 0.142 0.517 -0.003 -0.003 0.079 0.121 0.591 Table 4: Linear regression analyses for linguistic, cognitive and social factors predicting German reading (ELFE), German writing (HSP), English reading (PSAK-R) and English writing (PSAK-W) at the end of grade 4; OLS regression with the standardized regression coefficients β (* p<0.05). Program dummy 2: EFL vs. Bili-20, Program dummy 3: EFL vs. Bili-50, Program dummy 4: EFL vs. Bili-70. 94 5. Results As expected, the teaching program was the most important predictor for the results of the English tests, accounting for 51.7 % and 59.1 % of the variance. Less pronounced effects were noted for the German tests, where the program factor accounted for only 14.2 % and 17.6 %, respectively. The effect of language background was not significant for the English tests, even when it was entered as the sole predictor (as in model 1, see also Table 1). It served as a predictor for the German tests when entered separately (see Table 1) or together with gender (model 2) but lost its influence when nonverbal intelligence (SPM, model 3), maternal education (model 4) and the school program (model 5) were added as additional predictors. Gender only served as a predictor for the German writing test HSP (model 2). For the HSP, gender actually remained a significant predictor even when adding the other factors (models 3-5). It did not, however, turn out to be a significant predictor for the other three tests. In model 4, cognitive basic skills (operationalized by the SPM) were included, which appeared to be a significant predictor in all language tests, with the exception of the English writing test PSAK-W. An effect of this factor has already been suggested in Table 2, which indicated significant between-program differences of nonverbal intelligence. In model 5, the school program (using dummy variables) was added as a final predictor. This factor most significantly contributed to the variance, particularly for the English tests, where it showed much stronger effects than the other four factors. However, the school program predicted the results of the German tests to a much lesser extent. Here, the effect of nonverbal intelligence (SPM) accounted for more variance than the school program. Because of the use of dummy variables for the factor program, between-pro‐ gram effects could be shown, at least for a comparison between the EFL program (serving as reference) and the more intensive bilingual programs. For the English tests, the results indicated a significant difference between the EFL and BILI-50 and Bili-70, but not between EFL and Bili-20. No between-group differences were noted for most of the German tests, with the exception of the HSP, where EFL and Bili-50 showed significant differences. In sum, the teaching program turned out to be the most significant predictor for the results of the English tests, while nonverbal intelligence was the most significant predictor for the results of the German tests. The results of the language tests were, however, not predicted by the children’s language background, in particular when other factors were added as possible predictors. 95 5.2 Between-group comparisons 6. Discussion The present study examined the effect of FL input intensity on German and English reading and writing skills of fourth graders in four FL programs. The purpose of this chapter is to discuss the results for the four teaching programs (chapter 6.1), as well as learner-internal variables such as language background (6.2), gender (6.3), and cognitive and social background (6.4). Finally, limitations and additional ideas for further research will be presented in chapter 6.5. 6.1 Effectivity of the FL programs A large number of studies have compared language proficiency of students in bilingual vs. non-bilingual (including immersion) programs (e.g., Couve et al., 2016; Genesee, 1987, 2004; Genesee & Jared, 2008; Lambert & Tucker, 1972; Möller et al., 2017; Pérez-Cañado, 2012; Rymarczyk, 2010; Steinlen & Gerdes, 2015; Steinlen & Piske, 2018; Swain, 1975; Swain & Lapkin, 1982; Wesche, 2002; Zaunbauer et al., 2012). However, this is the first time that the effect of FL intensity on reading and writing in different FL (bilingual) programs at the elementary school level has been compared in more detail. Chapter 6.1.1 is devoted to the outcomes of the majority language German reading and writing tests, while chapter 6.2 will present the findings of the FL English tests. 6.1.1 German Regarding bilingual programs, parents, politicians and teachers have been found to worry that children in such a school context would be disadvantaged in their development of literacy skills and other content-related subject knowledge because a FL (and not only the majority language) is used as the language of instruction (see e.g., Harley, Hart & Lapkin, 1986). In the present study, reading and writing skills of German as the official language in Germany were examined because academic achievement in the German school system would not be possible without adequate literacy competences, independent of the program that students attend. In general, the results of regression analyses did not reveal teaching program as a significant predictor (see Table 4). The final model, which included the teaching program, language background, gender, SES and SPM, accounted for only 14.2 % of variance for the German reading test and 17.6 % for the German writing test. Other factors such as mother’s educational background (for ELFE) and gender and nonverbal intelligence (for the HSP) seemed to be more important predictors than the teaching program. The most important result relates to the level of competence in German reading and writing. For the ELFE, the fourth graders in general obtained scores that are age-appropriate (or even better, see Table 1), i.e., corresponding to norm values (e.g., Lenhard & Schneider, 2006). In addition, no significant differences were noted between the four groups (Table 3). Similar results have been noted in other studies in Germany on bilingual programs (e.g., Baumert et al., 2017; Böttger & Müller, 2020; Gebauer et al., 2012, 2013; Steinlen, 2016, 2018a; Steinlen & Piske, 2013, 2015, 2016a, 2018a; Zaunbauer et al., 2005, 2013, Zaunbauer & Möller, 2006, 2007, 2010) and elsewhere (e.g., Björklund & Mård-Miettinen, 2011; Cheng et al., 2010; Genesee, 1987, 2004; Genesee & Jared, 2008; Genesee et al., 1989; Geva & Clifton, 1994; Jenniskens et al., 2018; Lambert & Tucker, 1972; Lasagabaster, 2001; Möller et al., 2018; Rubin et al., 1991; Swain & Lapkin, 1982; Thomas & Collier, 2017; Turnbull et al., 2001, 2003). Thus, attending bilingual programs does not seem to negatively affect the development of reading skills in the majority language. A similar result applies to the German writing test HSP. The children’s results in the three bilingual programs (i.e., Bili-20, Bili-50 and Bili-70) did not deviate from the norm (see Table 1). Therefore, the writing skills in the majority language (here: German) appear to develop age-appropriately in bilingual programs (independent of the FL intensity), replicating findings from other studies (e.g., Böttger & Müller, 2020; Gebauer et al., 2012; Genesee, 1987, 2004; Rubin et al., 1991; Steinlen & Piske, 2013, 2015, 2016a, 2018a, 2020; Steinlen, 2018b; Turnbull et al., 2001; Zaunbauer et al., 2005, 2013; Zaunbauer & Möller, 2006, 2007, 2010). It could also have been expected that the students in the bilingual programs would perform above the norm in the German reading and writing tests. Taking into account Cummins’ Threshold Hypothesis and Interdependence Hypothesis (e.g., 1979), such positive effects are likely when a high level is achieved both in the L1 and in the FL (e.g., Baker, 2001; Gonzalez, 1999). However, this is not the case in our data (i.e., the students generally obtained norm values), and similar results have also been reported from the Canadian and German immersion context (e.g., Swain & Lapkin, 1982; Turnbull et al., 2001; Zaunbauer & Möller, 2007). It may be speculated that the fourth graders in the present sample have apparently not been advanced enough in their FL to transfer any positive effects back to their L1. It would be of interest to examine whether our students would 98 6. Discussion perform above the norm in German reading and writing tests when they are older and have attained a higher level of competence in English, and how much exposure to English would be necessary for such positive effects to kick in. In this context, the results of the EFL children are also worth mentioning. They obtained age-appropriate values on the standardized German reading test, which were, however, slightly lower than the scores found for the three bilingual programs. Regarding the German writing test, EFL students performed below norm values and lower than students in the three bilingual programs (Table 1), significantly contributing to between-program differences (Table 3). Considering that the Hügelschule has always catered to students with a mixed social and language background (e.g., Tamm, 2010), this result is not surprising (see chapter 6.2 and 6.4.2 for additional information). In conclusion, this study replicates findings from other studies which gen‐ erally reported age-appropriate development in literacy skills of the majority language (most often the children’s L1). Thus, the development of children’s reading and writing skills does not seem to be negatively affected by attending a program with higher FL intensity, despite the fact that not all content subjects are taught in the majority language. 6.1.2 English For English reading and writing, the results of the present study parallel other findings (e.g., Bae, 2007; Böttger & Müller, 2020; Genesee, 1978, 2004; Genesee & Jared, 2008; Lambert & Tucker, 1972; Möller et al., 2017, 2018; Nold et al., 2008; Pérez-Cañado, 2012; Reeder et al., 1999; Rymarczyk, 2010; Steinlen, 2016, 2017, 208a, b; Steinlen & Piske, 2013, 2015, 2016a, b, 2018a, b, 2020; Swain, 1975; Swain & Lapkin, 1982; Wesche, 2002; Yadollahi et al., 2020; Zaunbauer et al., 2012; Zydatiss, 2009) and are unambiguous: the higher the intensity of the FL program, the better the students’ results in FL English reading and writing tasks (see Tables 1 and 4). In the present sample, the students in the Bili-70 program obtained the highest scores, the students in the EFL and BILI-20 programs the lowest scores, with the Bili-50 group in between. Regression analyses of the present study (see Table 4) clearly suggested the teaching program to be the most important predictor for the outcomes of English reading and writing skills, accounting for 51.7 % and 59.1 % of the variance. Thus, the teaching program turned out to be a stronger predictor than learner-internal variables such as language background, gender, SES, and/ or cognitive background (see chapters 6.2-6.4 for a discussion of these factors). 99 6.1 Effectivity of the FL programs Which level of competence did the students in the different programs attain at the end of grade 4? On the English reading test, the results of the Bili-70 group corresponded to level B1 (according to the levels set by the Primary School Assessment Kit (Little et al., 2003) in conjunction with the Common European Framework of Reference for Languages (Council of Europe, 2001, 2018); the Bili-50 group to level A.2.3 (with a great number of students at B1 level, see also Steinlen, 2016; Steinlen & Piske, 2018a); and, finally the Bili-20 and EFL students being at a low A2 level. For English writing, the Bili-70 group scored at the upper end of level A2, Bili-50 at a comfortable A2 level and the Bili-20 and EFL groups at level A1. These competences go way beyond the reading and writing skills that children (particularly in mainstream programs) are expected to acquire after four years of English lessons (in Germany usually consisting of two 45-minute lessons per week), where learners should attain level A1 (e.g., KMK, 2013; Ministerium Baden-Württemberg für Kultus, Jugend und Sport, 2004, 2016a; Ministerium für Bildung, Wissenschaft, Forschung und Kultur des Landes Schleswig-Holstein, 2007; Senatsverwaltung Berlin für Bildung, Jugend und Sport, 2006). 6.1.2.1 The two partial IM programs The results for the partial IM programs are particularly encouraging. For reading and writing, the levels B1 and upper A2, respectively, are not expected in Ger‐ many until the end of year 9 of lower secondary school (Hauptschulabschluss, KMK, 2004). Thus the fourth graders in the Bili-50 and Bili-70 programs were not only able to understand short simple sentences (A1) but also simple texts (A2/ B1, depending on the complexity of the text). More specifically, B1 level students can read straightforward factual texts on subjects related to their field and interest with a satisfactory level of comprehension, whereas at A2, they can understand short, simple texts on familiar matters of a concrete type which consist of high-frequency and everyday language (e.g., Council of Europe, 2001: 69). Regarding writing skills at level A2, learners are expected to write short, basic sentences which describe immediate needs, personal events, familiar places, hobbies or work; to use the most frequent connectors (e.g., and, but, because) to link sentences in order to write a story (e.g., Council of Europe, 2001: 240); and to write short, familiar words with reasonable phonetic accuracy, while not always conforming to standard spelling (ibid.: 118). The findings of the present study suggest that FL literacy skills can develop particularly positively in IM programs. This is probably due to the fact that a much higher number of reading and writing activities are carried out in English subject classes than in regular 100 6. Discussion English classes, and that the students not only receive more written input, but also have to produce more written output. As mentioned previously, the students in the Bili-70 program reached level B1 for reading and A2 for writing in English, i.e. the best results in this sample. A similar finding was obtained in a standardized test on English reading fluency (TOSWRF, Mather et al., 2004), which was conducted with the same group of students (Steinlen & Piske, 2018c). The results indicated that these fourth graders were only one year behind those of monolingual English students of the same age from Australia, paralleling findings reported by Zaunbauer et al. (2012) for the same test with 37 students attending a Bili-70 program in Northern Germany. How can these findings be accounted for? In this particular bilingual school, many teachers were native speakers of English who had previously also worked as teachers for content subjects (such as science or math) in other English-speak‐ ing countries. Carrying out English reading and writing tasks in content subjects was for them, therefore, a natural part of their job, the only difference being that their students now were not native English speakers but learners of English. As pointed out in chapter 4.1.4, the school also implemented a concept of FL literacy where a program based on the Phonics Method is carried out from the second half of the first grade onwards. Thus, students are trained in using English sounds and sound combinations and their graphemic representations on a regular basis, often in small groups in lessons with team teaching. FL teaching in this school also involves a strong focus on the use of English books and materials with different levels of difficulty, on different topics in different genres. Thus, each classroom has its own small library (supplementing the school library) from which the students frequently take home books for the weekend, and teaching regularly involves book projects where students are encouraged to present their favourite books. More information about literacy activities in the FL classroom may be found in chapter 7. In the Bili-50 group, students reached upper A2 level for English reading (some students even obtained level B1) and a solid A2 for English writing. In previous studies on this partial IM program, which employed subsamples of the present sample, similar findings were reported (e.g., Steinlen, 2016, 2017, 2018a, b; Steinlen & Piske, 2013, 2015, 2016a, b, 2018a, b, 2020). In addition, two other English reading tests were also used with these groups, namely standardized tests from Australia on reading comprehension (TORCH, Mossensen et al., 2003) and on reading fluency (TOSWRF, Mather et al., 2004). A comparison with norm values by native English students at the same age showed that Bili-50 students in grade 4 obtained scores that were comparable to those of monolingual English children who were one year (reading comprehension) or two and a half years 101 6.1 Effectivity of the FL programs younger (reading fluency), respectively (e.g., Steinlen, 2016; Steinlen & Piske, 2015, 2016a). However, it remained unclear why these students performed better in the test on comprehension than on fluency. The authors mentioned test formatting as one explanation (i.e., that the children had been more familiar with fill-in-the-gap texts than with reading unrelated words not separated by spaces), and also suggested that comprehension may lead fluency in beginning FL acquisition as a second explanation, but called for additional research in this context. Regarding English writing skills of a large subsample of the Bili-50 group, Steinlen & Piske (2020, see also Steinlen, 2018b) found, not surprisingly, A2 to be the level that the fourth graders reached in the test PSAK-W. In addition, Steinlen & Piske (2020) also examined spelling errors found in students’ texts on the topic “Yesterday”, geared at level B1, which is the first study in the German FL elementary context of its kind. Each text consisted, on average, of five sentences and 31 word tokens (corresponding to 22 word types). Five spelling errors per text were found, i.e. two errors concerning upper and lower case, one omission error, one substitution error and one word with multiple errors. An error rate of 13 % is within the 60-90 % rate of correctly spelled words in the transition phase as reported by Peregoy & Boyle (2013: 210). Spelling errors were noted on the intralingual and interlingual levels alike. For example, errors regarding upper and lower case were due to transfer from German, while developmental errors (which may also be found in native English writers) often concerned omissions of grapheme doublettes. However, ample research is still needed to determine how the number of spelling errors and their types change over time, also taking into consideration the learners’ age and their exposure to the FL. Despite the fact that this particular school did not provide a concept to sys‐ tematically foster FL literacy skills in the IM classroom, the encouraging results of the English literacy tests in the Bili-50 program clearly pointed to successful teaching strategies. As the teachers reported in personal communications, the learners in this program regularly have to carry out a variety of reading and writing tasks in the content subjects (e.g., MeNuK); that is, they are exposed to and have to produce a variety of texts of different genres and formats. Some examples of such FL literacy activities are found in chapter 7, which are aimed to help FL learners improve their reading and writing skills. 6.1.2.2 The less intensive bilingual program One surprising result of this study pertains to the Bili-20 program (see also Steinlen & Gerdes, 2015), the students of which performed as well as the EFL group on the English reading and writing tests, i.e. at level A1/ A2 for English 102 6. Discussion 13 One-way ANOVAs (Bonferroni) of the Steinlen (2018a) data indicated significant differences between the EFL and Bili-20 group regarding English receptive vocabulary (F (2, 287) = 68.607, p<0.05), English receptive grammar (F (2, 292) = 33.447, p<0.05), English reading fluency (F (2, 294) = 12.046, p<0.05), English reading comprehension (F (2, 274) = 67.818, p<0.05), and English writing (F (2, 273) = 118.638, p<0.05). Note that these results have not been controlled for SES or cognitive background variables. reading and writing (with reading at lower A2 level). This language level is achieved despite the fact that the Bili-20 group had received more English input within the four years of elementary school, namely two science lessons per week in grades 1 and 2, five science lessons in grades 3 and 4 and additionally two English-as-a-subject lessons from grade 3 onwards; whereas the EFL children had only two lessons of English-as-a-subject starting in grade 1. How can this result be accounted for? The first and most important argument relates to the quantity of literacy activities in the FL classroom. As noted in chapter 4.1, FL reading and writing did not constitute the focus in the Bili-20 program because in Schleswig-Holstein, the primacy of FL speaking and listening is key, with reading and writing activities playing only a minor role (Ministerium für Bildung, Wissenschaft, Forschung und Kultur des Landes Schleswig-Holstein, 2007). Thus, it is very likely that the children in this program would outperform students in regular EFL programs in FL listening and speaking tasks because these skills were particularly promoted in the Bili-20 program. For example, Böttger & Müller (2020) noted that in low-intensity bilingual programs in Bavaria, students in grade 4 generally attained level A2 in English listening tasks, thus exceeding the expectations of the KMK (2013a) who envision level A1 at the end of elementary school. FL oral skills, however, were beyond the scope of the present project but should be addressed in future studies, especially when comparing regular EFL and low-intensity bilingual programs. The second argument points to the intensity of the FL input. Informal interviews with teachers in the Bili-20 program revealed that English was not used consistently in either the science lessons or the English lessons. For example, the science units on Schleswig-Holstein or on bicycles were usually taught in German. It seems that German had also been used as the language of instruction in other topics related to science or English-as-a-subject. This is also reflected in the data on English receptive vocabulary and grammar. No significant differences between the EFL and the Bili-20 group were noted, paralleling the findings for English reading and writing, as a re-analysis of the Steinlen (2018a) data indicated. 13 Thus, it is always important to consider not only the number of content subjects taught in the FL, but also the quality thereof. 103 6.1 Effectivity of the FL programs Quantity alone may present a misleading picture of the actual extent of FL input that the students receive. There is still a growing need in Germany to study the effectiveness of low intensity FL programs in more detail, because a growing number of elementary schools nowadays are offering such programs (although exact figures so far are, unfortunately, lacking). Many schools are not able to provide a high intensity bilingual (aka immersion) program due to personnel or financial resources. It seems to be easier to implement low-intensity FL programs in elementary schools, as the example of the projects “Lernen in zwei Sprachen - Bilinguale Grundschule Englisch (Bilingual Elementary Schools English, Böttger & Müller, 2020) and “Bilinguale Grundschule Französisch” (Bilingual Elementary Schools French) in Bavaria show (Stiftung Bildungspakt Bayern). In this context, Böttger & Müller’s (2020) data on fourth grade students in low-intensity bilingual pro‐ grams with English as the target language are particularly interesting because they support the findings of the present study, in that „fast alle Schülerinnen und Schüler das Sprachkompetenzniveau A1 erreichen und viele von ihnen in verschiedenen Bereichen sogar A2-Niveau schaffen können“ (… almost all students reach the language competence level A1 and many of them can even reach A2 level in different areas, ibid.: 40) in English tests on reading and writing. In conclusion, it seems that FL reading skills in particular benefit from less intensive bilingual programs because many activities that students carry out in content subjects such as science would not be possible without the help of reading (e.g., reading texts for information, reading instructions for experiments or other activities, answering questions of a quiz, etc.), whereas FL writing skills in low-intensity bilingual programs seem to develop at a similar rate as those in regular programs. 6.1.2.3 The regular EFL program Turning to the results of the EFL group, the students in this program reached almost level A2 for English reading and A1 for English writing. Thus, the outcome for writing is in line with the expectations by the Ministerium Baden-Württemberg für Kultus, Jugend und Sport (2004, 2016a) but exceeds them for English reading (see also Steinlen, 2018a; Steinlen & Piske, 2018a). Unfortunately, in many other studies on EFL students’ English competences in elementary school, the test outcomes were not related to the Common European Framework of Reference (Council of Europe, 2001, 2018). This applies, for example, to the EVENING study (Engel, Groot-Wilken & Thürmann, 2009; Wilden et al., 2013), the BIG study (BIG-Kreis, 2015) and to other studies conducted in Germany on English reading and writing (e.g., Frisch, 2013; 104 6. Discussion Rymarczyk, 2011), which would have facilitated any comparison between the studies. The positive results of the EFL program in the present study are very likely attributable to several factors which relate to a high quality of the English lessons in this particular EFL program. First, the English teachers hold university degrees of teaching English-as-a-subject and are not teaching “fachfremd” - which often negatively affects students’ performance (see also Fleckenstein et al., 2020); some of them even work in the Bili-50 program of the same school. Second, English lessons in this particular school (which also offers a partial IM program) may enjoy more prestige than in other elementary schools. For example, English lessons which cannot take place (due to the teacher’s sickness or other school-related absences) may be taken over by another qualified English teacher. In other schools, the students may, in such a case, be distributed among different classes, or any teacher colleague who has an open spot in their schedule but whose expertise is not English may serve as a substitute for the English lesson. Personal communication with the English teachers of the EFL program also revealed that they took care to use English as a medium of communication in class (“as much English as possible - as little German as necessary”). They also tried to provide many opportunities for EFL students to read and write English texts from grade 3 onwards (following the guidelines of the Ministerium Baden-Württemberg für Kultus, Jugend und Sport, 2004). Some of these FL literacy activities, which seem to have contributed to the positive results of this group, are presented in chapter 7. In conclusion, the results of the present study on the development of English literacy skills in programs with different FL intensity are, indeed, very promising - the students obtained levels that exceeded the expectations based on curric‐ ular guidelines by the three Ministries of Education in Baden-Württemberg, Berlin-Brandenburg and Schleswig-Holstein, in particular for English reading. However, many teachers in these programs pointed out that their FL students more or less acquire English literacy skills in passing, which is why they advocate an instructional approach to systematically train FL literacy skills in the classroom (see also Börner et al., 2017; Burwitz-Melzer, 2010; Hempel et al., 2018). In this context, this study did not discern how German and English reading and writing instruction per se may have related to the test results (see e.g., Jared et al., 2011). Therefore, further research is warranted to explore the effects of teaching approach in the FL classroom in more detail. In addition, many studies within the IM context found students’ positive attitude and motivation to contribute to the achievement of additive bilingualism (e.g., Gardner & Lambert, 1972; Lambert & Tucker, 1972), which may well apply to 105 6.1 Effectivity of the FL programs the acquisition of literacy skills in German and English. The same applies to students’ meta-linguistic skills, working memory, FL spoken language skills, as well as their FL vocabulary and FL syntactic skills, which may have had an impact on the test results (e.g., Grabe, 2010). These and other factors need to be addressed in future studies, for example, by using additional tests and questionnaires in which students provide information about their reading habits at home and in school in either language, and about their attitudes towards the teaching approach. 6.2 Language background The most important finding of the present study is that the minority and majority language children examined here generally performed equally well in both English and German reading and writing tests. Thus, the variable majority/ minority language background did not seem to exert any influence on the reading and writing skills in the majority language German or on the FL English, independent of the intensity of the FL program. 6.2.1 German Regarding German reading and writing skills, the present study yielded different results for the different FL programs on the effect of language background (see (Table 1, 4). For the Bili-20 and Bili-50 program, there were no significant differences in the scores of minority and majority language students, and the groups obtained age-appropriate results for both German reading and writing. Similar findings for bilingual (IM) programs have been reported in previous studies conducted in Germany (e.g., Steinlen & Gerdes, 2015; Steinlen & Piske, 2013, 2015, 2018; Steinlen, 2016, 2017, 2018) and elsewhere with older learners (e.g., Lindholm-Leary & Hernández, 2011; Marian et al., 2013; Thomas & Collier, 2012). Three possible reasons for the encouraging results obtained for minority language children in this this study may be noted. First, the German lessons were often conducted by the same teachers who also taught the content subjects in English, and thus may have used the same methodological and didactic principles and strategies in German and in English lessons. This may have given minority language children more opportunities to acquire and improve their German (Burmeister & Pasternak, 2004), because the content had been presented in a contextualized manner, i.e., with a strong link between content 106 6. Discussion and language. Second, owing to factors such as low socio-economic status and mismatch between their L1 and the language of instruction (German), minority language students in Germany are often considered at-risk students. In the present study, minority group status and language background were confounded as majority language children exclusively spoke German at home and minority language children used German in addition to their family language. Although both groups differed with respect to their parents’ educa‐ tional background (Fachhochschulreife vs. Hochschulreife), this difference is a rather small one within the German educational system. Independent of their migration status and/ or their language background, the parents of the present study, therefore, may be described as middle-class, highly educated and concerned about their children’s educational welfare; the children may therefore not be considered at-risk children, at least with respect to their SES. Third, according to the Interdependence Hypothesis (e.g., Cummins, 1991), positive transfer effects from the minority language students’ family language (i.e., their L1) to the FL would also be possible. i.e., their high proficiency in their L1 would positively affect their L2 (German) skills and there would not be a mismatch between the minority language children’s L1 and German as the main language of instruction. However, it is difficult to compare minority language students’ L1 and FL skills, mainly because standardized reading and writing tests for minority language students’ family language are lacking, particularly for elementary schoolers (e.g., Brehmer & Mehlhorn, 2018). Fourth graders’ reading skills in their family language could be assessed with PIRLS tests (e.g., Mullis et al., 2017), which are available in 48 languages. However, it should be noted that minority language children generally do not receive explicit instruction in their family language (e.g., Brehmer & Mehlhorn, 2018). However, in Baden-Württemberg (and in the Hügelschule), Turkish is offered as “Konsulatsunterricht” once a week, which takes place outside the regular lessons but usually in the school itself, with teachers usually recruited by the Turkish government, which is also responsible for the curriculum. It is not clear whether any other family languages are taught outside of school. It is very likely, however, that minority language children’s listening and speaking skills in their family language are much superior to their reading and writing skills, because they have often not been exposed to written language in their L1 (as noted by the children themselves in informal interviews, see also Brehmer & Mehlhorn, 2018; Carreira, Jensen & Kagan, 2009). However, further studies are necessary to assess the influence of the children’s family language in more detail, in particular regarding the minority language students’ literacy proficiency. 107 6.2 Language background The analyses of this study yielded different results for the EFL and Bili-70 programs, where majority language students outperformed their minority language peers in the German reading and writing test. Turning first to the German reading test, it is important to note, though, that the minority language groups, independent of the FL program, still performed age-appropriately. Thus, these group differences due to language background do not play a decisive role, because the minority language children’s reading scores still correspond to the norm values. In the regression analyses for the English reading test, the effect of language background also disappeared when cognitive abilities were added as factors (Table 4). This result contests findings of other large-scale studies conducted in regular elementary school programs in Germany, where minority language students were usually not only outperformed by their majority language peers but also did not obtain age-appropriate results in German reading tests (e.g., Bos & Pietsch, 2006; Chudaske, 2012; Dollmann, 2010; Haag et al., 2017; Hussmann et al., 2017: and see e.g., Hesse et al., 2008; Haag et al., 2016; OECD, 2019 for older learners in grade 9). The results for the German writing test HSP of the EFL and Bili-70 group showed that the majority language groups scored within norm values (i.e., age-appropriately). However, the minority language students in both groups performed below norm and, on average, spelt only 60 % of the German words correctly. Similar results have been reported for eighth grade students in a middle immersion program in Canada (Hart, Lapkin & Swain, 1988) and in many mainstream elementary school programs in Germany (e.g., Bos & Pietsch, 2006: Chudaske, 2012; Dollmann, 2010; Haag et al., 2016; Hussmann et al. 2017; May, 2006). An informal look some background variables revealed that this result may be attributable to a substantial number of new children, who either had a refugee status (EFL program) or who had recently moved to Germany due to parents having new jobs (Bili-70). These students had only recently been accepted and had not attended this particular program since grade 1, as had the other students. A future study should, therefore, employ a matched subgroup technique to control for such effects. As regression analyses indicated, though, language background did not significantly predict the outcomes of the German writing test, when cognitive abilities were added as a factor (Table 4, and see chapter 6.4 for a discussion on the effects of nonverbal intelligence). Nevertheless, it seems that writing in the majority language (here German) poses a greater challenge to minority language students than reading does. This may be due to the fact that writing is inherently more complex than reading, as it requires, among other things, more time, effort and preparation (e.g., De La Paz & McCutchen, 2016). Therefore, the HSP seems to be a better test to distinguish German literacy skills 108 6. Discussion of minority and majority language students than reading tests such as ELFE, at least when examining those with high proficiency in the majority language German. As a conclusion, the variable language background does not seem to exert too great an influence on elementary schoolers’ proficiency in reading in the majority language German, but does seem to affect writing to a larger degree, in particular when minority language students are new to a program. Apart from that, other factors such as the children’s cognitive skills and their social background apparently play a more important role. These factors need to be accounted for in any study on minority language students’ proficiency in any language. 6.2.2 English In this study, significant differences between majority and minority language students were only noted for the English reading test (but not for the writing test) in the mainstream EFL program, with majority language students outper‐ forming their minority language peers. In other studies on the mainstream FL context, language background did not affect FL reading skills (e.g., Haenni-Hoti, 2008; Kessler & Paulick, 2010; Wilden & Porsch, 2015; see also Hesse et al., 2008; Rauch et al., 2010 for older learners). However, our students’ cognitive and social background had not been controlled for in this first within-group comparison (Table 1). As subsequent investigations showed, any discrepancy between minority and majority language children in the EFL program for English reading disappeared in the regression analyses: for both English reading and writing proficiency, language background did not appear as a significant predictor, in any of the five models, even when it was entered as the sole predictor (Table 4). Thus, the children’s language background does not seem to affect foreign language skills in reading and writing, regardless of the teaching program. For the three bilingual programs Bili-70, Bili-50, and Bili-20, the results of the present study were consistent. No significant differences between majority and minority language children were found regarding their scores in the English reading and writing test (Table 1). A similar result has been reported in other studies, for example from bilingual school programs in Germany (e.g., Steinlen & Gerdes, 2015; Steinlen & Piske 2013, 2014, 2015, 2016, 2018; Steinlen, 2016, 2017, 2018a, b) or Canada (Bild & Swain 1989; Genesee & Jared, 2008; Swain & Lapkin, 1991; Swain et al., 1990), although these were conducted either with a smaller sample size and/ or with older learners. A variety of reasons has been 109 6.2 Language background proposed to account for this result, and these relate to the minority language children’s L1 and L2. Starting with the children’s L1, Swain and her colleagues suggested that minority language children’s L3 literacy skills may be affected by the typological proximity between the family and the target language (i.e., the children’s L1 and L3, e.g., Swain & Lapkin, 1991; Swain et al., 1990; and see Hermas, 2010, 2014b; Jin, 2009; Na Ranong & Leung, 2009 in Hopp, 2018, who proposed that the L1 constitutes the default language from which any kind of (grammatical) transfer originates). For example, Bild & Swain (1989) found that learners with a Romance background performed better in a French writing test than their peers with non-Romance backgrounds. It could, therefore, be argued that our learners with a Germanic background (e.g., Danish, Dutch) would obtain better results in the English literacy tests than learners with a non-Germanic language background (e.g., Arabic, Russian, and Turkish). However, in the present sample, the number of children with a Germanic background (except for German) is very small, and the results would be confounded with L2 German (as another Germanic language) as the ambient language. Nevertheless, such a typological comparison of children’s family languages in relation to their L3 performance in English reading and writing tasks is worth pursuing in a future study, including a larger number of students with a Germanic background. Swain and her colleagues also pointed out that students’ literacy experiences at home may play an important role, i.e., the better they know how to read and write in their family language, the better they perform in L3 literacy tests (e.g., Swain & Lapkin, 1991; Swain et al., 1990). As mentioned previously, our students have usually acquired their L1 literacy skills either in school in the family’s home country, in heritage language lessons in Germany and/ or at home. Although we did collect data on minority language students’ L1 literacy background using a student questionnaire, these data have not yet been analyzed; these will hopefully shed some light on the question as to how L1 and L3 literacy skills may be related to each other. Testing minority language students’ L1 would be even more ideal, but it is difficult to provide standardized literacy tests for the minority language students’ L1 a) when over thirty different languages are spoken in one school (e.g. Hügelschule), and b) when such tests are not available for many of these languages (e.g., Albanian, Arabic, Kurdish, and Romanian, to name just a few), at least not for elementary schoolers (e.g., Brehmer & Mehlhorn, 2018). Third, many minority language children in our sample had not received instruction in their family language, which became apparent when they were asked to complete a C-test in their family language and were unable 110 6. Discussion to do so. To test the relationship between minority language students’ L1 and their L3 English, therefore, is a topic that needs to be addressed in future studies. Many previous studies, however, reported both L1 and L2 transfer on L3 literacy skills. For example, Griva & Chostelidou (2013) found that success in L3 writing depended on competence in L1 and L2, and that metacognitive strategies (e.g., revising a text) acquired in L1 and L2 were transferred to the L3. Cenoz & Gorter (2011) examined students’ texts in Basque, Spanish and English for content, structure, vocabulary, grammar, and orthography, and found these aspects to be cross-linguistically interconnected. Thus, students with a high score in orthography or vocabulary in one of the languages also obtained good results in the other languages and vice versa. In contrast, a few studies noted a stronger influence of the L2 than of the L1 on the L3. In their study of L3 writing by adult students with different L1s, L2 German and L3 English, Angelovska & Hahn (2012) identified only a few cases of L1 transfer but found more evidence of L2 syntactic and lexical elements being transferred to L3 English. Similar results have also been reported for L3 English grammar skills of Turkish third graders, where children relied on their L2 German and not on their L1 Turkish (e.g., Hopp, 2018). Likewise, in this study it also appears reasonable to assume that L2 effects are stronger than L1 effects. For example, the phonological and orthographic (i.e., typological) similarities between many German and English words (the minority language children’s L2 and L3) may have positively affected the outcomes of the English reading and writing tests, with German being closer to English than to the children’s family languages (e.g., Arabic, Kiswaheli, Kurdish, Persian, Serbian, Croatian, Spanish, or Turkish). Ringbom (2001) pointed out that for lexis, L3 learners at an early stage will frequently make use of L2 words in their L3 production if the L2 and L3 are related and have a number of common cognates. In our English tests, a number of words are very similar in German and English, which may have facilitated English reading comprehension of both majority and minority children alike (e.g., hand, apple, bananas, brown, ball, or t-shirt). Indeed, tasks including familiar words were solved more successfully than other tasks, as an informal look at the data revealed. The same applies to the writing task at A1 level, in which the children had to label nine English pictures with the English word provided in a box. Independent of the language background, the children had the least number of problems with words that were similar to German (e.g., toilet, bed). A more detailed analysis, however, is necessary in order to confirm these impressions. Apart from typological similarities, Williams & Hammarberg (2001) also identified L2 proficiency, L2 status, and recency of L2 use as influential for 111 6.2 Language background the activation of a language in L3 production (see also Hammarberg, 2018). In terms of proficiency, our minority language students in all programs performed age-appropriately in the German reading and writing tests, with test results corresponding to norm values, except for the minority language students in the EFL and Bili-70 programs, who scored below norm in the German writing test (Table 1, see also section 6.2.1). Some studies indeed indicated a relationship between L2 German and L3 English. In her study on English listening skills, Elsner (2007) compared fourth graders with German and Turkish backgrounds and found Turkish students to perform lower than their German peers. Her findings led Elsner (2007: 245-246) to argue that the linguistic skills many of her Turkish children showed in the majority language German did not seem to be sufficient to enable these children to achieve good results in their L3 English. Turning to the present study, it is likely that the insignificant differences between majority and minority language students in both English tests may be due to the minority language students’ high proficiency in German reading and writing. In a future study, however, the relationship between L2 and L3 writing skills should be examined in more detail. For example, in two FL programs of our study (i.e., EFL and Bili-70), the minority language children achieved rather low scores in the German writing test but their scores in the L3 English writing test did not differ from their majority language peers. It is possible that these children may have been exposed to English outside school more intensively than was indicated in the questionnaires. It is also possible that the different formats of the German and English writing test may play a role, as the German writing test is concerned with spelling only, but the English test includes vocabulary and grammar as well. Weaker spellers, therefore, would have had more possibilities to receive higher scores on the English writing test than on the German writing test. In terms of L2 recency (i.e., the extent to which the learner has used the language recently), our minority language students largely used the L2 German at school, i.e., with their peers, with many teachers and other school staff (unless they used English). Because many languages are represented in the four programs, there are hardly any groups including many students speaking the same language. In other words, the children speak their family language only to a very limited extent in school, and the L2 German is, therefore, constantly activated during this time. Finally, the importance of L2 status, defined as “a desire to suppress L1 as being ‘non-foreign’ and to rely rather on an orientation towards a prior L2 as a strategy to approach the L3” (Hammarberg, 2001: 36-37), has been reported in many studies on L3 vocabulary and L3 syntax, mainly with older learners 112 6. Discussion (e.g., Angelovska & Hahn, 2012; Bardel & Falk, 2007; Cenoz, 2001; De Angelis, 2005, 2007; Hopp, 2018). With regard to the present study, the minority language students had learned to rely on their L2 German because it had been the main language of communication outside the children’s home since preschool (at the latest). In addition, the minority language children usually learnt the alphabet in German first (i.e., at school) and may associate any literacy activities more with German than with their family language. More studies are, however, necessary to determine how background languages are suppressed in L3 literacy tasks, how a young learner can consciously decide on which strategy to use and on the question of which components of the L2 (apart from L2 vocabulary, L2 grammar and L2 writing system) may be transferred to L3 reading and writing and under which conditions. In sum, when following Hammarberg’s hypothesis (2018), namely that the language that reaches the highest total score on the scales mentioned above is the one which the learner will activate most often, then German would be the language on top, outscoring the children’s family languages. It is thus assumed that the minority language children’s L2, namely German, constitutes the source of transfer in L3 English reading and writing. A few examples from the children’s free text production may serve to illustrate this point. Minority language children produced the same kind of errors as their majority language peers, which are clearly based on German. For example, a girl with a Turkish background and some Turkish language lessons spelt nouns (*<Judo>, *<Football>, *<Mother>) with capital letters (as in German), although this spelling does not correspond to Turkish (Schroeder & Şimşek, 2014). Another boy used German syntax in a sentence like *<I go yesterday morning to school>), which is not reflected in his L1 Farsi where the verb is in final position. As noted in Steinlen (2018b), minority language students’ L2 text production regarding transfer patterns for spelling, grammar and vocabulary has to be explored in more detail in future, large-scale studies. The question still remains whether transfer from German reflects effects of typology or proficiency. Especially in the context of early L3 acquisition, models of cross-linguistic influence will need to be expanded to include the effects of relative proficiency and degrees of use of prior languages, particularly with respect to L3 reading and writing. Various reasons may account for these positive findings. First, minority language students may benefit from the fact that English is new to almost all students and are more likely to “sit in the same boat” in the FL classroom, compared to other subjects where students’ performance depends more on their competences in German (e.g., Steinlen & Piske, 2016a; Piske, 2017, 2018). Second, unlike in many other studies (e.g., Chudaske, 2012; Hopp, Kieseier, Vogelbacher 113 6.2 Language background & Thoma, 2018), the parents of the majority and minority language students did not differ significantly from one another in terms of their SES, so that there was no interaction between linguistic and social economic background, which is likely to have positively influenced the results of this study. Third, minority language students often grow up with more than one language and may thus develop a high level of language awareness at an early age, which could help them to acquire additional languages. These family languages could also be used as a resource in FL teaching, for example in tasks that activate metalinguistic awareness (e.g., Hopp et al., 2017). As shown in the Hopp et al. (2017) study, when asked to comment on different grammatical structures in German and English sentences, minority language children sometimes also referred to their family language to account for structural differences, i.e., they were able to exploit an additional resource, which, so far, hardly receives any attention in FL teaching (see also Angelovska & Hahn, 2014, for similar results for adults). Finally, of particular importance for the results presented here is the quality of teaching. The teachers in all four FL programs had a very high command of English, i.e., all students, independent of their language background, were exposed to high-quality input. Minority language students in particular may have benefitted from the clarity and vividness of how the content in the FL was presented, i.e., the strong interaction between FL and content (so-called contextualization), in addition to other teaching strategies such as scaffolding techniques, negotiation of meaning (e.g., paraphrasing, clarification requests or confirmation checks to facilitate the understanding of FL input) and multisensory learning, which should make it easier for students, regardless of their linguistic background, to develop a new language and specialist content relatively quickly (e.g., Burmeister, 2006; Piske, 2015, 2017, 2018; Steinlen & Piske, 2013, 2018; Steinlen, 2016, 2017, 2018a, b, and see Kersten, 2019 for a review). In his literature review, Somers (2017: 515) concluded that minority language students in bilingual (CLIL) programs may not only develop age-appropriate proficiency in the majority language. In addition, they also gain opportunities “(a) to acquire advanced levels of functional proficiency in an additional language without endangering their academic success; (b) to take advantage of precisely the kind of (scaffolded, interactive, motivating) pedagogy that is intended to facilitate access to both content and language; (c) to develop positive self-image and motivation; and (d) to acquire important capital for achieving success economically and socially.” Furthermore, Somers argues that policy makers may limit minority language students’ opportunities and risk a number of negative consequences (e.g., loss of self-esteem, lost opportunity to learn an 114 6. Discussion additional language) if they withhold them from bilingual programs. Already in 1976, Genesee concluded: “There is nothing in the data to suggest that (French) immersion would not be suitable for third language children” (p. 591). Our data clearly support this statement for bilingual programs with varying FL intensity in Germany. 6.3 Gender Girls outperform boys in any language test - this has been the result of many studies on German and FL English proficiency. Will the results of the present study agree with this finding? 6.3.1 German Regression analyses (see Table 4) did not indicate gender to be a significant predictor for German reading. In previous studies with subsamples of the Bili-50 group, girls and boys also showed similar performance in the German reading comprehension test (Steinlen & Piske, 2016a); the same finding was reported on the standardized German reading fluency test Würzburger Leise Leseprobe-R (Steinlen, 2016). Many studies in Germany and elsewhere, though, reported gender-specific effects for reading, pointing to advantages for girls in regular and bilingual elementary and secondary school programs (e.g., Baumert et al., 2017; Böhme et al., 2016; DESI-Konsortium, 2008; Hussmann et al., 2017; Schipolowsky et al., 2017, see also Mullis et al., 2017; OECD, 2019 for the international context). It is not clear how to account for these differing findings. It is possible that different test formats may have affected the results; or that variables such as the children’s cognitive and social background may have overridden effects of gender, which some studies have suggested (e.g., Howard et al., 2014). Many studies also pointed to the influence of reading habits, reading motivation and self‐perceptions of competence while reading (e.g., McElvany et al., 2017). In the present study, these aspects have been included in a student questionnaire, but the data have not yet been analyzed. It is expected that girls and boys will not differ regarding the relationship between German reading skills and their reading habits, reading motivation and self‐perceptions of competence while reading. In contrast, gender keeps its role as a significant predictor for German writing even after all other factors (i.e., SPM, maternal education, teaching 115 6.3 Gender 14 ANOVA’s yielded significant differences for the HSP between boys and girls for the EFL (F (1, 107) = 9.125, p<0.05) and the Bili-20 program (F(1, 101) = 8.458, p<0.05) but not for the Bili-50 (F(1, 121) = 0.526, p>0.05) and the Bili-70 program (F(1, 74) = 0.668, p>0.05). program) were added (Table 4). A closer look at the data revealed significant gender-differences in the EFL and Bili-20 program but not in the Bili-50 or in the Bili-70 program. 14 It seems that gender does not affect German writing performance across all FL programs but that its effect is limited to the two programs, which are, incidentally, the least intensive of the four programs. Similar results were reported from the regular elementary and secondary school context, with girls outperforming boys in German writing tests (e.g., Böhme et al., 2016; Drucks, 2011; Hartig & Jude, 2008; Kennedy et al., 2012; May, 2006; Richter & Brügelmann, 1994). So far, it is not clear why effects of gender on German writing skills were restricted to less intensive FL programs and not obtained for the more intensive bilingual ones. In her study with a subsample of the Bili-50 group, Steinlen (2018b) found the same non-significant effect and argued that parental support may have been a decisive factor. As an informal look at the parent question‐ naires suggested, the parents (with a middle-class and academic background) supported their children (girls and boys alike) on a regular basis in terms of homework monitoring and joint reading activities. Additional studies are needed to examine the relationship between socio-economic status, parental support, familial literacy activities and outcomes in German writing tests in more detail. In future student questionnaires, for example, questions on parental support would not only include instrumental support (e.g. “My family goes over my homework with me”) but also affective support (e.g. “My family helps me to feel better when I am worried about school-related tasks”). As some studies noted, outcomes in language (writing) tests are indeed influenced by parental support, but this effect may be differentiated according to gender, with girls being more sensitive to the affective dimension of parental support and boys to the instrumental one (e.g., Pires, Candeias, Grácio, Galindo & Melo, 2017). To diminish gender effects in writing, teachers may engage students more frequently in writing activities which focus more on the students’ interests in general. For example, Horst (w/ o year) pointed out that creative writing activities may be particularly useful to increase the motivation to write texts. Older elementary school students, for example, could be presented with a picture (e.g., by Paul Klee) as a prompt. The task would then be designed in such a way that the students are free to choose the format of the text (e.g., argumentation, narration, informational text) and the genre (e.g., diary entry, comic, fairy 116 6. Discussion tale or a combination). Such tasks would consider gender-typical differences (e.g., girls often write longer texts than boys and prefer narratives) without cementing them, and without neglecting the diversity within the gender groups, in order to support optimal learning. As for spelling, May (1994) noted that girls seemed to be better at following spelling rules, whereas boys rather remember morphemeand word-specific peculiarities (e.g., length marks). In addition, he found that particular words were better spelt by boys than by girls and vice versa. For example, girls spelt words like <Tierärztin> (veterinarian, female) or <Sekretärin> (secretary, female) better than boys, and boys outperformed girls in words like <Schiedsrichter> (referee, male) and <Computer>. The present data have not yet been analyzed along these lines. May (1994) pointed out that teachers should more frequently use such words in spelling tasks which are of interest and relevance to the children in their particular class. However, more research is needed in order to examine gender-specific teaching strategies for spelling in more detail. In conclusion, the effects of gender on German reading and writing are limited to just two programs for just one test format, namely writing (aka spelling). Reading is not affected by gender differences at all, and the more intensive the program, the fewer differences between girls and boys are noted. Thus, it may be speculated that the teachers of German already employ gender-sensitive teaching strategies for reading. Informal interviews indicated that teachers try to draw boys’ attention to fictional genres and continuous texts, because boys often prefer non-fiction and discontinuous texts (e.g., Merga, 2017). Teachers also often make sure to ask boys more frequently to read aloud texts in class and to offer them appropriate and diverse reading materials to cover a wide range of topics to encourage them to read for pleasure more often (see e.g., VBW, 2009). Some teachers also mentioned that they sometimes critically discuss the school books with their class, as many tasks represent certain family forms, attribute certain activities to women and men (or girls and boys), or establish optical characteristics for genders (see e.g., Jochim, 2014). Finally, it has often been pointed out that teachers in elementary school are predominately female (90 % were women in 2008, VBW, 2009), and that this lack of male role models may negatively affect boys’ performance in elementary school. However, such a correlation was not found when comparing the PIRLS-2008 reading data internationally (e.g., VBW, 2009); nor does this seem to apply to the teachers of the present study, who were mostly female, a reality which also did not seem to negatively affect boys’ reading skills in German. In his review on gender differences in L1 acquisition, Dale (1976) concluded that there seemed to be few differences or no differences at all between boys 117 6.3 Gender and girls during their preschool and early school years. However, “around the age of ten or eleven years, girls establish a definite pattern of superior verbal performance, which continues through the high-school and college years” (Dale, 1976: 311), which leads to “consistent, though modest differences” in tests evaluating spelling, punctuation, verbal creativity or comprehension of complex written texts and logical verbal relations. In Dale’s view, extralinguistic factors, such as society or personality, are likely to play a part in girls’ apparent superiority in writing skills in adolescence, which should also be examined in more detail for bilingual classrooms in elementary school. 6.3.2 English In all four programs, the children’s gender did not significantly predict the outcomes of the English reading and writing tests (Table 4). This finding is surprising because it has repeatedly been reported that girls perform better than boys, at least regarding regular FL programs at the elementary school level (e.g., BIG Kreis, 2015; Burstall, 1975; Duursma et al, 2007; Lynn & Wilson, 1993; May, 2006; but see Edelenbos & Vinjé, 2000; Steinlen 2018a). In contrast, the results for the three bilingual programs (Bili-20, Bili-50, and Bili-70) correspond well with findings from other bilingual programs in different countries which did not find any gender-specific effects (but with older secondary school students, e.g., Canga Alsonso, 2015; Dallinger, 2015; Roquet et al., 2016; Thomas & Collier, 2002; Zumhasch, 2010; Zydatiss, 2007 or with a smaller sample from the Bili-50 program, e.g., Steinlen, 2016, 2018a, b; Steinlen & Piske, 2016a, 2018). This result strongly indicates a more comprehensive view on successful FL learning, which does not only take variables such as ‘language proficiency’ and ‘gender’ into consideration but also neurological, cognitive, affective, social and educational factors, including interactional effects (e.g., López Rúa, 2006; Uhl, 2016). Factors such as learning strategies, motivation and social interaction in the classroom have often been mentioned to account for the differences in girls’ and boys’ FL performance (e.g., Böttger, 2016; 2017; López Rúa, 2006; Kissau & Turnbull, 2006; Schmenk, 2002). However, in the present study, girls and boys did not differ significantly regarding their scores in the English reading and writing tests, independent of the program. This means that for these programs, there is no “fundamental need” to reconsider and restructure FL lessons in a gender-specific way, as called for by Böttger (2014: 148). The teachers of the present study have apparently ensured that boys and girls participated equally well in FL interactions. They may also have encouraged boys to use learning strategies such as cognitive, metacognitive, and social strategies to 118 6. Discussion a greater extent, as boys have been found to use them less often than girls (e.g., Lopez-Rúa, 2006; Sunderland, 2010). Finally, the teachers may have used creative, experimental, kinesthetic, and performative activities in the classroom in order to acquire literacy skills in the FL that have engaged girls and boys alike (e.g., games, role-plays, projects, videos, e.g., Böttger, 2016, 2017). As gender-related differences have been found for the mainstream FL context, it seems as if the EFL program of the present study implemented various aspects of gender-sensitive FL teaching (i.e., social and educational factors as well as the children’s individual background) in such a way that it was beneficial for both girls’ and boys’ FL literacy development. For bilingual programs, Lasagabaster (2008) pointed out that gender-specific motivational differences may be more easily compensated for in certain con‐ texts. For example, boys may indirectly be more encouraged to use the new language by the instrumental necessity of being interested in the topics of a content subject (i.e., the topic water in science lessons). The topics in bilingual lessons may often be more authentic and relevant to girls and boys alike than the topics of regular FL textbooks, which often appeal to girls rather than boys (e.g., Roquet et al., 2016). Any future study focusing on gender issues in FL learning should system‐ atically take into consideration teaching practices (i.e., the teachers, their methods, the materials in the classroom, e.g., using video observations) and students’ individual predisposition (i.e., cognitive, social and affective factors, using standardized tests and questionnaires), thus allowing for a more holistic approach to determine possible gender effects. 6.3.3 Interaction between gender and language background Although the interaction between language background and gender was not the focus of the present study, the data allow for some analyses (Table 4). The results, though, differ for the two German tests: for German reading, only language background was a significant predictor, while gender was not; and both lost their predictive significance once cognitive background (i.e., nonverbal intelligence) was entered as a variable. For German writing skills, both language background and gender were significant predictors when entered as the only factors. However, when cognitive background was entered as a variable, language background lost its predictive power (although gender did not). For both German tests, the two factors together only accounted for 3 % and 4 % of the variance, respectively. In other words, even a combination of language background and gender does not seem to explain well the outcomes 119 6.3 Gender of the German reading and writing tests. A similar result has been noted for the bilingual preschool context (Steinlen et al., 2019), but not in other studies (for older learners), which found male immigrant students at age 15 to be particularly disadvantaged in German reading tests (e.g., Segeritz et al., 2010) as well as in any other academic test in school for ninth graders (e.g., Gottburgsen & Gross, 2012). The results are more conclusive for the English tests: regression analyses revealed language background and gender as not being significant predictors for English reading and writing, even when they were both entered as the only factors. This result parallels findings from the bilingual preschool context (e.g., Steinlen et al., 2019, using receptive grammar and vocabulary tests), and a previous study on three FL elementary school programs with a smaller sample (e.g., Steinlen, 2018a). Thus, regardless of the intensity of the FL input, the differences between multilingual girls and multilingual boys, between monolingual and multilingual girls or between monolingual and multilingual boys have been found to be non-significant in English tests These findings may be attributed to many reasons, relating for example to minority language boys’ cognitive and social background (e.g., in terms of parental school-related activities). It is also likely that the educational setting played an important role, in the sense that these children may not have felt marginalized in the FL classroom, felt sufficiently supported by their teachers and their peers and received tasks and topics that were of interest and of relevance (Segeritz et al., 2010) in order for them to perform appropriately. However, more research is needed in order to determine the effects of teaching strategies which are carried out in a genderand ethnicity-sensitive manner in more detail, and whether such effects (i.e., gender and language background not being significantly predicting FL literacy skills) will last beyond the elementary school years. 6.4 Control variables In the following, the role of the control variables used in the present study is briefly discussed, i.e., nonverbal intelligence as a proxy for cognitive skills (chapter 6.4.1) and maternal education as a proxy for SES (chapter 6.4.2). 120 6. Discussion 6.4.1 Cognitive background It has often been claimed that students’ cognitive abilities affect their academic performance (see e.g., Kersten, 2019, for a review). For the bilingual (immersion) context in Germany and elsewhere, it has also been shown that students in such a program show superior cognitive abilities, operationalized as nonverbal intelligence, as compared to their peers in mainstream programs (e.g., Baumert et al., 2017; Deventer et al., i. pr.; Gebauer et al., 2012; Genesee, 1976; Swain & Lapkin, 1982; Zaunbauer & Möller, 2007, 2010; Zaunbauer et al., 2005). In the present study, significant differences of the nonverbal intelligence test SPM were also noted (Table 2), with students in the Bili-70 group outperforming those in the Bili-50 group, who, in turn performed better than the students in the Bili-20 group, with the EFL group scoring lowest (Table 1). Several reasons have been suggested to account for such a finding. First, it may be possible that the different groups had already started their respective programs with different cognitive abilities. Zaunbauer & Möller (2010) found such differences already for students attending either mainstream or bilingual (immersion) programs at the end of grade 1. However, in a subsample of this project (which also included students of the present study), Yadollahi et al. (2020) did not notice any significant differences between first graders either attending the EFL or the Bili-50 program. It appears that at least students from this particular school in the city of Tübingen start their respective program with their cognitive abilities on par (at least with regard to nonverbal intelligence). This finding also replicates results from the St. Lambert experiment in Canada, where differences in Raven tests were not found for Year 1 students in IM and regular classes (e.g., Lambert & MacNamara, 1969; Lambert & Tucker, 1972; Genesee, 1987). Second, it may also be possible that the students’ cognitive abilities are affected by their SES, i.e., that students with high SES also obtain better scores on cognitive tests due to the fact that cognitively more demanding activities are more frequently carried out in families with high SES in contrast to families with low SES (see e.g., Kersten, 2019 for a review). In the present study, maternal education was used as a proxy for SES, and indeed, significant differences between the programs were noted: the more intensive the FL program, the higher the mothers’ educational background (Table 2). Third, some studies reported positive long-term cognitive effects of intensive FL programs. These relate to cognitive skills such as metalinguistic awareness (Bialystok, Peets & Moreno, 2014) and executive functions (e.g., auditory selective attention, divided attention and mental flexibility, Nicolay & Poncelet, 2013); but apparently not to nonverbal intelligence: Zaunbauer & Möller (2010) 121 6.4 Control variables did not find any differences in the rate of development from grade 1 to grade 2, comparing students in bilingual (immersion) and regular EFL programs with the Coloured Progressive Matrices (Raven, 1976). Future studies should, therefore, take into consideration students’ longitudinal development of a range of cognitive skills from grade 1 to 4, including nonverbal intelligence but also executive function, metalinguistic awareness, working memory, short-term phonological memory and verbal intelligence. Such a study, integrating various cognitive skills, would reveal a clearer picture of whether more intensive FL programs indeed positively affect long-term cognitive development. In the present study, though, nonverbal intelligence (using the SPM) has mainly been used as a covariate to control for the outcomes of the language tests (Table 3) because the four groups differed in their cognitive abilities (Table 2), following the approach of many other studies (e.g., Baumert et al., 2017; Bialystok et al., 2014; Edele et al., 2018; Gebauer et al., 2012, 2013; Genesee & Jared, 2008; Genesee, 1987; Jared et al., 2011; Reeder et al., 1999; Swain & Lapkin, 1982; Thomas & Collier, 2002; Zaunbauer & Möller, 2007, 2010; Zaunbauer et al., 2005). The results of the regression analyses (Table 4) revealed nonverbal intelligence (SPM) to be a significant predictor for all tests even when the other variables were entered into the models. The only exception is the English writing test PSAK-W, where nonverbal intelligence loses its role as a significant predictor when school program is added as a factor. Thus, nonverbal intelligence in the present study predicts German reading and writing outcomes as well as test results for English reading in grade 4, independent of the program. This result parallels findings of other studies on the relationship between nonverbal intelligence and literacy skills in the L1 (e.g., Gebauer et al., 2012; Morvay, 2015; Naglieri, 2001; Stanovich et al., 1984; Zaunbauer & Möller, 2010) and in the FL (e.g., Genesee, 1976; Morvay, 2015; but see MacCoubrey, Wade-Woolley, Klinger, & Kirby, 2004). As noted earlier (e.g., chapter 2.6), the relationship between nonverbal intelligence and literacy skills in the L1 and L2 is not well understood yet, as the limited number of studies clearly shows. Reading and writing skills are affected to a much greater extent by cognitive variables like working memory, short-term phonological memory, phonological awareness, and verbal intelligence (e.g., Berninger et al., 1996; Grabe, 2009; Leki et al., 2008), which have not been employed in the present study. However, these variables should be included in any future studies on reading and writing to obtain a more holistic view of how the development of cognitive and literacy skills interact in different FL programs during the elementary school years and beyond. 122 6. Discussion 6.4.2 Social background A number of studies have pointed out that bilingual (immersion) programs often attract parents with a higher SES and educational background (e.g., Apsel, 2012; Baumert et al., 2012; Genesee, 1987 Möller et al., 2017; Zaunbauer & Möller, 2007; Zydatiss, 2007). This is also true for the samples of the present study: the more intensive the FL program, the higher the maternal educational background (Table 1), with many of the between-group differences being significant (Table 2). For the Bili-70 program, this result does not come as a surprise, because it is offered by a private school where the parents pay tuition fees (Steinlen & Piske, 2018c). Although some children were exempted from the fee due to economic reasons, this school is not likely to attract a great number of parents who would have to apply for a scholarship for their children to cover the fees. One would envision the clientele of the Bili-70 to be particularly “elitist” (see e.g., Klemm, Hoffmann, Maaz & Stanat, 2018) but, indeed, in terms of SES and educational background, it does not differ significantly from the Bili-50 program (Table 2) which, in turn, is offered by a public school (i.e., without tuition fees), mainly geared towards children in the neighborhood (see also Steinlen, 2018a; Steinlen & Piske, 2018a; Tamm, 2010). In both programs, the mothers’ educational background on average corresponds to a university entrance certificate (“Hochschulreife”, see Table 1). As noted above, the mothers’ educational background was, on average, lower in the Bili-20 and EFL programs, and corresponded to an entrance certificate for special universities in Germany (“Fachhochschulreife”, see Table 1). However, the parents of the present sample had a higher educational background than parents in other schools in Germany. In 2008, only 39 % of all elementary school parents held a university entrance certificate (“Fachor Hochschulreife”) and 29 % a secondary school certificate (“Realschulabschluss”) (Nold, 2010: 143). Independent of the program, the parents of the present sample may thus be described as highly educated, with an upper middle SES, and interested in their children’s educational well-being (see Steinlen, 2018a, b; Steinlen & Piske, 2018a). Because the maternal educational background differed significantly between the four groups, it was expected that this variable would also affect the children’s scores in the German and English tests. In order to control for such effects, maternal educational background had been entered as a covariate (Table 3), following other studies (e.g., Baumert et al., 2017; Bialystok et al., 2014; Gebauer et al., 2012; Reeder et al., 1999; Rolff et al., 2008; Swain & Lapkin, 1982; Thomas & Collier, 2002; Wilden & Porsch, 2015; Zaunbauer & Möller, 2007, 2010; 123 6.4 Control variables Zaunbauer et al., 2005). In regression analyses (Table 4), maternal education lost its role as significant predictor for the outcomes of the English tests as soon as the FL program was entered as a factor. Similarly, other studies from Germany and elsewhere also reported that students’ SES would not affect their FL proficiency, particularly for those from bilingual (immersion) programs (e.g., Baumert et al., 2017; Genesee, 1987, 2004; Genesee & Fortune, 2014; Holobow et al., 1985; Lindholm-Leary, 2001; Lindholm-Leary & Borsato, 2006; Thomas & Collier, 2017; Zaunbauer et al., 2012). Furthermore, maternal education did not significantly predict the results of the two German tests, contradicting studies which reported a significant influence of SES (which included parents’ educational background) on German reading and writing skills for elementary school children (e.g., Haag et al., 2017; Hussmann et al., 2017; Krüger et al., 2010; Mullis et al., 2017; May, 2006; Zöller & Roos, 2009). It is possible that such effects are not found in the present sample because the parents already had a very high level of education. In sum, SES affected the composition of the samples in the individual FL programs, although the parents’ social background may generally be character‐ ized as upper middle class, highly educated and interested in their children’s academic progress. Nevertheless, this factor did not provide significant predic‐ tions for the German and English reading and writing tests, particularly when the FL program was added as an additional factor. 6.5 Summary, limitations and future studies The present study dealt with German and English reading and writing skills of students in grade 4 in four elementary school programs with different FL intensity, with a focus on children’ language background, their gender and their social and cognitive background. The main results of this study in particular, and this project in general, are briefly summarized in chapter 6.5.1. Chapter 6.5.2 outlines some additional limitations associated with the project, and further ideas for future research are presented in chapter 6.5.3. 6.5.1 Summary of results To summarize, the foci of this study were on the following aspects, which had not (or only partly) been addressed in previous studies: instead of comparing only two programs (e.g., regular vs. immersion programs), this study compared four FL programs with different intensity. Many other studies only focused on 124 6. Discussion high-intensity programs (e.g., partial vs. total immersion programs), whereas this study also included a bilingual program with lower FL intensity (i.e., a Bili-20 program). Furthermore, writing skills (particularly in the FL) had rarely been addressed in prior studies. Finally, not only one language was examined (for example the students’ FL) in the present study but both their German and English reading and writing skills, which provides a more holistic view of what students in these four different programs can achieve after four years of elementary school with respect to their literacy skills in the school language and in the FL. Some findings of the present study did not support results of previous studies. For example, in this study, gender did not affect the results of any language test in any group, probably due to gender-sensitive teaching styles (see also Steinlen, 2016, 2018a). In many other studies, though, girls regularly outperformed boys in literacy tests (e.g., Baumert et al., 2017; Böhme et al., 2016; Hussmann et al., 2017; Mullis et al., 2017). Furthermore, the minority language students in this EFL program performed as well as their majority language peers in the German reading test, probably due to their social / cognitive background. In contrast minority language children have been found to perform worse than their majority language peers in many other large-scale studies on German reading skills (e.g., Bos & Pietsch, 2006; Chudaske, 2012; Dollmann, 2010; Haag et al., 2017; Hussmann et al., 2017). Most results of this study, however, corresponded well with previous research and extended their scope. First, intensity matters, i.e., the higher the FL input, the better the results on FL tests (see also e.g., Couve et al., 2016; Steinlen, 2016; Steinlen & Piske, 2018). Second, in all bilingual programs, the students generally achieved age-appropriate results in the German reading and writing tests, despite some / many subjects being taught in English (see also e.g., Baumert et al., 2017; Genesee, 1987; Steinlen, 2016, 2018a, b; Swain & Lapkin, 1982; Turnbull, Lapkin & Hart, 2001; Zaunbauer & Möller, 2006, 2007, 2010). Third, students’ social and cognitive background was found to significantly differ across the FL programs. Therefore, these variables were used as covariates, particularly regarding preselectional effects for students in the bilingual programs (see also e.g., Baumert et al., 2017; Bialystok et al., 2014; Gebauer et al., 2012; Reeder et al., 1999; Swain & Lapkin, 1982; Thomas & Collier, 2002; Zaunbauer & Möller, 2007, 2010; Zaunbauer et al., 2005). Fourth, in high intensity FL programs (i.e., Bili-50 and Bili-70), fourth graders reached level A2/ B1 for English reading and level A2 for English writing. Level B1 is not expected of students in mainstream secondary school programs until the end of grade 10 (KMK, 2013). Fifth, the minority language students in the EFL program obtained scores below average 125 6.5 Summary, limitations and future studies 15 German translation: „Als wie wohlhabend schätzen Sie Ihre Familie im Vergleich zu anderen Familien ein? “ Five categories were provided: „Überhaupt nicht wohlhabend / eher nicht wohlhabend / durchschnittlich / eher wohlhabend / wohlhabend“. (In com‐ in the German writing test (see also e.g., Chudaske, 2012; Haag et al., 2017; May 2006). And, most importantly, these minority language students in the four programs performed as well as their majority language peers in all English tests and in the German reading test (see also e.g., Steinlen & Piske, 2013; Steinlen, 2016, 2018a, b). Thus, this study corroborates results of many previous studies and, and the same time, provides additional findings on minority and majority language children’s literacy development in programs with different FL intensity. 6.5.2 Limitations As already indicated, this study shows a number of shortcomings and some of them have already been addressed in previous sections. Aspects related to test materials and statistical analyses will be dealt with first. The parent questionnaire of the present study was rather short and only included four pages. In the course of the analyses, it was noted that minority language parents had, for example, not been asked to provide information on whether their children were immigrants of the first, second or third generation. In many studies it has been reported that minority language students of the first generation performed better in language tests than comparable peers of later generations (e.g., Kemper, 2015). Furthermore, the parents filled out the questionnaire only once, most of them at the end of grade 1. However, the use of the family language and German may have changed over the elementary school years, with German becoming more dominant over time, particularly among siblings (e.g., Brehmer & Mehlhorn, 2018). Finally, the parents were not asked to state how well their child understood, spoke, read and wrote the family language/ s, which would have given an indication of its state of development, which in turn may affect the development of German as the L2 (see e.g., Cummin’s Interdependence Hypothesis, 1981) and in turn the development of English as the L3 (e.g., Rauch et al., 2010; Hammarberg, 2018). With regard to the family’s economic status, the parents were asked to estimate their wealth in comparison to other families. This question (as well as the one on the parents’ educational background) was often not answered at all because many parents felt the question to be too personal, as informal interviews revealed. Furthermore, the parents were provided with five responses 15 . It is well known 126 6. Discussion parison to other families, as how wealthy would you categorize your family? Not wealthy at all / rather not wealthy / average / rather wealthy / wealthy.) that when in doubt, people tend to choose the middle (e.g., Hansen, Hurwitz, Marks & Mauldin, 1951). In addition, it turned out that the phrasing of this particular question was rather ambiguous: some parents indicated that they found the phrase “in comparison to other parents” very subjective, as this may depend on the social context of a family. In future studies, the family’s socio-economic status should be assessed differently, for example by compiling a hierarchical score for socio-economic status of the parents on the basis of parents’ highest level of education, household income and educational profiles of professions and professional groups, following e.g., Ganzeboom, de Graaf & Treimann (1992). It is, however, not clear whether this would have increased the response rate. Several questions from the parents’ questionnaire have not been evaluated yet. For one, the parents provided information about the family languages. Kessler & Paulick (2010), for example, reported that students with a Russian background performed better on English language tests than students with a Turkish background. Although in this project, the individual languages are represented by only a small number of students, this could be examined in more detail in a future study with a greater sample. Furthermore, any information on family activities (e.g., telling stories, singing songs, watching TV, using the internet) have not been analyzed yet, nor have data on parental attitudes towards and their knowledge of English as a FL; nor minority language parents’ reading habits (in the family language, as compared to German). Finally, many children had attended a bilingual preschool before they entered a bilingual elementary school. This could have ramifications on the outcomes of the FL tests, although, as Möller (2015) pointed out, such effects are stronger in grade 1 and seem to disappear in grade 4, at least with respect to her study on cohesion in English oral narrations in a bilingual German-English school. As far as test materials are concerned, one shortcoming relates to the use of the PSAK, which is a non-standardized test (in contrast to the standardized German tests). It was originally designed for non-English students in Ireland in transition classes in order to move them to mainstream classes once they have reached level B1 in English speaking, listening, reading and writing, according to the Common European Framework of Reference. In terms of reading, the PSAK is easy to assess, and the responses leave hardly any room for doubt. However, the PSAK writing subtest is often rather subjective, at least regarding the task on level B1 (free text production), where the texts are to be evaluated in terms 127 6.5 Summary, limitations and future studies of grammar, vocabulary, handwriting, spelling and punctuation. According to the guidelines, one point is given when the student “can use a small amount of vocabulary”, and three points when s/ he “can use a lot of vocabulary additional to basic items” (Little et al., 2003: 14). However, how can “a small amount of vocabulary” or “a lot of vocabulary” be quantified, and which vocabulary do the so-called “basic items” include? An informal comparison of different raters’ evaluations did not show large discrepancies; nevertheless, interrater reliabilities should be reported in future studies. Furthermore, the test formats in the two languages do not match, which is more severe for writing than for reading. Both the ELFE and the PSAK-R roughly assess reading at the word level, sentence level and text level, although in slightly different forms. However, the formats of the German and English writing tests are not equivalent because the HSP focuses on orthography whereas the PSAK-W is not only concerned with spelling and punctuation but also with grammar and vocabulary. It is not clear how this mismatch affected the results, or whether such effects are more prone in correlation analyses or when using writing skills in one language as predictors for writing outcomes in another language. Unfortunately, writing tests displaying the same format in German and English were not available, but future cross-language comparisons should take care to use test formats that examine similar skills in writing. As regards cognitive skills, the present study employed tests on nonverbal intelligence (i.e., Colored and Standard Progressive Matrices, Raven, 1976 a, b), but in this project children were also assessed with a test of attention / concentration (D2-R, Brickenkamp et al., 2010). Results of these tests were reported in earlier studies on the Bili-50 program (Steinlen & Piske, 2013, 2015, 2016b), where minority and majority language students achieved age-appropri‐ ate results, without any significant group differences. However, the effect of attention / concentration on literacy skills is not well explored, and a future study may take a closer look at children with attention problems and how they would score on such tests. Many studies also found a positive effect of immersion programs on cognitive skills, particularly on children’s executive functions (e.g., Bialystok et al., 2014; Lazaruk, 2007; Lee, 1996; Nicolay & Poncelet, 2013; Nicolay & Poncelet, 2015; see Kersten, 2019 for a review). Future studies will have to explore whether such effects also hold for the present sample, for example by using the Simon or the Flanker task (see e.g., Poarch, 2018 for a review of these two tests in multilingual learners), with students in high intensity FL programs probably outperforming their peers in less intensive FL programs. A number of statistical analyses were carried out in the present study, which included, among others, t-tests, ANOVAS, correlation analyses and regression 128 6. Discussion analyses. However, not all students participated in all tests, due to sickness or other school-related activities, and not all parents completed the questionnaire (i.e., item non-responses). Therefore, some data (i.e., data values, not variables) in the present sample is missing. There are three alternatives to deal with missing data. First, the data set is taken as such, including missing data. The disadvantage is that the results may not be representative (e.g., Field, 2013). This approach was taken in the present study. Second, depending on the topic, a matched subgroup technique can be used, where groups may be matched for social, cognitive and language background, and only complete data sets are used (e.g., Flege, 2009). The disadvantage of such an approach is that only a limited number of subjects (with complete data sets) can be used. Third, missing data can be imputed, i.e., missing data is replaced with substituted values; and these values are usually predicted based on already existing values. This approach is worth taking in future studies as it may reduce bias and increase the representativeness of the results (e.g., Field, 2013). Finally, Structural Equation Modelling (SEM, e.g., Blunch, 2012) may be employed in future studies to analyze causal structures (e.g., between the results of the language tests and the individual learners’ background variables) in more detail. SEM are often used because of their ability to impute relationships between unobserved constructs (latent variables, e.g., intelligence, or, in our case, effectivity of the FL program) from observable variables (e.g., test scores). In addition to testing the overall theory, SEM therefore allows the researcher to diagnose which observed variables are good indicators of the latent variables. However, large sample sizes are required to provide sufficient statistical power and precise estimates using SEM, which is therefore an approach that will be taken once more data has been collected. 6.5.3 Future studies It has already been noted that not much is known about the development of minority language students’ family language. Assuming that there is an interaction between the languages being learned (e.g., Hammarberg, 2018, Hufeisen, 2018), the acquisition of one language should in some way affect the development of the other language/ s. Thus, even if a child has not explicitly been taught to read and write in her/ his family language, knowing how to read and write in another language (in this case German) may also positively affect literacy skills in the family language (provided both languages share the same writing system). Thus, students with a Turkish background, for example, should not have any problems reading and writing in Turkish, once they know 129 6.5 Summary, limitations and future studies how to do it in German, because Turkish has a fairly transparent orthography. In our project, data on some family languages (e.g., Albanian, French, Italian, Polish, Russian, Spanish, Turkish) were collected using a C-test and some free writing, as standardized reading and writing tests in most family languages are not available for elementary schoolers. A first evaluation of these data for 46 Turkish students in grades 3 and 4 (Reinhardt, 2018) suggested that the Turkish children indeed had severe problems with the C-test, and they expressed their frustration already during the test. Many of them had been attending Turkish lessons (“Konsulatsunterricht”, see chapters 2.4 and 6.2), if only for a year or so, and all of the students reported using Turkish at home with their family members. The results of the C-test indicated that they depended on language use at home: the more Turkish was used at home, the higher the scores of the C-test. However, only half of the children were able to complete the four texts at all, and many errors were noted regarding spelling, some of these attributable to German, e.g., in cases where <ı> was dotted (Reinhardt, 2018). However, more detailed analyses are warranted for in this respect. First, additional analyses should include more information about the children’s reading skills in Turkish on the word, sentence and text level. Second, not much is known about the development of Turkish writing skills in Turkish “Konsulatsunterricht” courses and how this may affect the children’s development of German writing skills. Thus, proper reading and writing tests should be used instead of C-tests. Third, the data from the C-tests can also be used to examine the relationship of L1 Turkish, L2 German and L3 English literacy skills. In the present study, for example, no significant differences were found between majority and minority language students in terms of their German reading test and their English test. But this was a quantitative analysis and does not automatically mean that these two groups also provided the same types of errors. More analyses of the individual language tests are necessary to study possible transfer effects in more detail. Finally, as noted above, the C-tests are also available for other family languages (e.g., Croatian, Italian, Polish, Russian, Spanish), but these data have not been analyzed yet. For Russian, for example, it would be possible to examine whether the Cyrillic script may affect children’s writing in an alphabetic script such as German and English. In the present study, only the data of fourth grade students were introduced. In Steinlen (2016, 2017, 2018b), the data of third grade students were included. But the data of this project also allow for a longitudinal approach, where the same students can be followed through their elementary school years, i.e., from grade 1 to 4, as they were tested at the end of every school year (i.e., a cohort panel study). The advantage of such an approach is that developmental patterns 130 6. Discussion can be determined, which allow for more concise results because changes can be more accurately observed. One important drawback of a longitudinal study is sample attrition. For example, some of our students were no longer able to participate because they had moved away, which cut down useable data (e.g., Caruana, Roman, Hernández-Sánchez & Solli, 2015). Currently, longitudinal data are available from approximately 100 students in the EFL and Bili-50 programs who attended the programs from grade 1 to 4. Such a longitudinal approach would ideally not only include variables such as gender, cognitive skills, language background, and prior exposure to the FL (to name just a few), but it would also assess effects of other factors such as reading/ writing strategies, motivation, vocabulary and grammar knowledge in both English and German, which have been found to affect reading and writing skills in any language (see chapter 2.1). Finally, the presentation of the data was based on average values of all subjects. This approach, though, did not consider learners who overor under‐ performed in the literacy tasks. For example, around 10-15 % of the students in the average classroom in Germany are nowadays classified as having dyslexia (e.g., Bundesverband Legasthenie und Dyskalkulie, 2016), which is defined as a delayed acquisition of reading and/ or writing skills which deviate markedly from those that are generally observed in children, with sight, hearing, IQ, as well as the student’s emotional and family background being normal (e.g., Nijakowska, 2010). Often such children are not even registered for a bilingual program or drop out later because teachers and parents feel that the exposure to an additional language would “overburden” such children (e.g., Apsel, 2010, Genesee & Fortune, 2014, see also section 2.2). Indeed, dyslexic students’ development of FL reading and writing skills is not well understood yet (e.g., Gerlach, 2019), and literature, particularly on bilingual programs, is scarce and usually geared towards older learners (e.g., De Bree & Unsworth, 2014; Genesee & Fortune, 2014; Kruk & Reynolds, 2012). An informal analysis of these data indicate that dyslexic students in the Bili-50 not only outperformed comparable dyslexic students in the EFL program regarding English receptive vocabulary and grammar, but also in tests on English reading and writing. However, more in-depth analyses are necessary to validate these observations, which will also examine whether such results may extend to minority language students as well (Steinlen & Piske, i. pr.). None of the reading and writing skills that the students in this study have obtained in either German or English would have been possible without the literacy activities that were conducted in the classroom. The following chapter, 131 6.5 Summary, limitations and future studies therefore, is devoted to best practice methods which seem to positively affect elementary schoolers’ foreign language reading and writing skills. 132 6. Discussion 7. Recommendations for teaching reading and writing in the FL classroom As noted in chapter 2.2, FL reading and writing used to play only a minor role in the elementary school curricula of the Federal States of Germany until recently, which is why FL reading and particularly writing activities are, more often than not, still conducted to only a limited degree, especially in regular EFL programs. It has, however, been shown that such activities strongly affect students’ competence in FL reading and writing (e.g., Bassetti & Cook, 2011; Frisch, 2013; Lesaux & Geva, 2006; NEPS, 2019; Oxley & de Cat, 2019; Peregoy & Boyle, 2013). The purpose of this chapter is to present recommendations for FL literacy activities in the elementary school classrooms based on research results, lesson observations, and informal interviews carried out with teachers of the three schools involved in the present project. Many of these recommendations and suggestions are applicable both to regular EFL programs and bilingual programs. However, some recommendations only relate to bilingual programs, when the focus is exclusively on fostering subject content. Moreover, some recommendations in this chapter may also be employed by teachers in regular school programs in Germany who teach subject content (e.g., science and mathematics) to minority language students (in German). As Piske (2018: 23 f.) pointed out, many of these minority language children acquire German as their L2 in similar conditions as students in bilingual programs: in both contexts, subject content is presented in a language that does not correspond to the students’ family language (e.g., English for majority language students in bilingual programs and German for minority language students in regular programs in Germany). Therefore, subject content must always be presented in such a way that language learning is fostered at the same time (see Demming, 2018 for a discussion of similarities and differences between CLIL and “sprachsensibler Fachunterricht”). In this chapter, the recommendations for teaching FL reading and writing to elementary school students include ideas for the literacy-rich classroom, different types of scaffolding to facilitate students’ reading and writing output, and awareness-raising activities, which should be embedded in authentic and relevant contexts (chapters 7.1-7.6). In order to improve their reading and writing skills, young learners also need to know how FL words are spelt. Teaching spelling, however, is a neglected area in early FL classrooms, which is why chapter 7.7 is exclusively devoted to invented spelling, spelling activities, the role of spelling errors and teaching English spelling rules. In addition, recommendations on how to give feedback on student writing are included in chapter 7.8, because many teachers pointed out that they felt more uncomfort‐ able providing feedback for FL writing than for FL reading. A short conclusion on teaching literacy is provided at the end of this chapter (chapter 7.9). In the following, some of the recommendations for FL literacy activities aimed at elementary school children are illustrated with photos taken at the three schools involved in this project. The recommendations are also accompanied by teachers’ statements drawn from informal interviews, and by informal lesson observations, which have taken place since the beginning of the current study in 2012. A last word on terminology: as pointed out in chapter 2.2, bilingual teaching in Germany is often subsumed under the umbrella term CLIL (Content and Language Integrated Learning), ranging from bilingual modules to partial IM programs (see Figure 1). This is why the term CLIL is used in this chapter (following the European tradition), which, as noted above, may also include subject content in German provided for minority language students in elemen‐ tary schools. 7.1 Starting with FL literacy activities as early as possible The schools of the current study introduce literacy activities in English as early as possible (i.e., from grade 1 onwards), because in their view the advantages outweigh the disadvantages. Very briefly, these merits include the following (see also chapter 2.2.1): it may help students in their FL literacy development when simultaneously confronted with the differences between a comparatively transparent writing system such as German and a more intransparent writing system such as English (see also e.g., Frisch, 2013; Piske, 2010; Rymarczyk, 2010, 2016; Schmid-Schönbein, 2008; Zaunbauer & Möller, 2007). Furthermore, writing may have a positive effect on the pronunciation of a FL (e.g., Bassetti & Cook, 2011). It may also be easier for some children to recognize FL word and sentence boundaries; and for many children, writing helps with remembering words and sentences (e.g., Schmid-Schönbein, 2008). Finally, teachers have a wider range of activities to choose from in the FL classroom (e.g., Hancock & McDonald, 2000). The teachers of the present study all mentioned that children will start to write in the FL whether they are instructed to do so or not. Even though many schools start as early as possible with FL literacy activities, the question often arises as to how young students should be introduced to the 134 7. Recommendations for teaching reading and writing in the FL classroom FL writing system and how these FL literacy skills should be fostered during the elementary school years. Unfortunately, a systematic approach is still missing, and in most FL classrooms, the students are introduced to the English writing system only in passing (e.g., Burwitz-Melzer, 2010; Frisch, 2013). In one of the schools of this project (i.e., the Bili-70 program), however, the teachers systematically introduced English sounds and sound connections and their grapheme correspondences using the Phonics program “Letters and Sounds” (www.letters-and-sounds.com/ ), starting in the second half of the first year. At this point the children had already learnt part of the alphabet in German. The focus of Phonics is on the acquisition of letter-sound correspondences (i.e., it emphasizes the alphabetic principle as a bottom-up approach), in contrast to the Whole Word Method, where the reader constructs a personal meaning for a text based on using their prior knowledge to interpret the meaning of what they are reading (see Frisch, 2013, for a detailed review). The introduction of Phonics in this particular school was probably due to the fact that many teachers were native English speakers and had already been acquainted with this program before they started teaching in Germany. The Phonics approach, as well as a strong focus on the use of English books and materials with different levels of difficulty, themes and genres to train reading comprehension, were likely to contribute greatly to the positive results in the English reading and writing tests of the Bili-70 program (see chapter 5). As Phonics is not an undisputed approach to learn reading and writing in English as the L1 or a FL, Frisch (2013) compared two groups of German EFL students in grade 2 whose English teachers had either used the Phonics or the Whole Word approach. Her results indicated that the students in the Phonics group pronounced English words more accurately to the target in reading-aloud tasks and were better able to comprehend English texts than the students in the Whole Word Method group. It seemed that the development of the FL reading skills of weak learners in particular improved when they were made aware of some peculiarities of the FL spelling system. Unfortunately, the usefulness of these two approaches has not been systematically evaluated in the various bilingual school contexts in Germany yet (but see Sendlmeier & Oertel, 2015, for a comparison of phonetic, syllabic, and holistic approaches to introduce German orthography in elementary schools). According to Hempel et al. (2018: 43 f.), many curricula of the Federal States of Germany point to a combination of analytical and holistic approaches to FL literacy. Generally, the holistic approach is preferred for reading, but many curricula also recommend a contrastive (i.e., analytical) approach regarding differences in phoneme-grapheme correspond‐ ences. The curriculum of Berlin is the only one which requires teachers to use 135 7.1 Starting with FL literacy activities as early as possible the analytical approach to writing, in that first graders are taught typical letter combinations in English as well as simple rules of word formation to facilitate the FL writing process. 7.2 Creating a literacy-rich environment Peregoy & Boyle (2013: 196) among many others recommend turning the class‐ room into a literacy-rich environment right from the beginning to emphasize the importance and functionality of print. Such an environment may also serve as a means to build the basic skills necessary for literacy development by demonstrating to students the function and use of FL print in an intentional, purposeful, and intensive way. Given the support of this environment, students are better prepared to work on their FL writing skills as well as on other FL literacy skills including phonemic awareness, phonics, reading fluency, vocabulary, and reading comprehension (e.g., Burmeister & Piske, 2011; Peregoy & Boyle, 2013). Creating this environment involves selecting materials that facilitate language and literacy opportunities, reflecting on classroom design, and intentional instruction and facilitation by teachers. Different ways may be employed to create a literacy-rich environment. For example, the objects in the classroom may be labeled with words so that students constantly connect written language with the things they represent (e.g., Cameron, 2001). Such labels may also be employed outside the classroom (Figure 5). Figure 5: FL labels used outside the classroom (Hügelschule, EFL and Bili-50) Moreover, it is common in many classrooms, and especially in CLIL class‐ rooms, to hang posters with English words and sentences on the walls (Cameron, 2001; Peregoy & Boyle, 2013). Such posters may include, for example, sight word vocabulary (i.e., high-frequency words that regularly recur in texts, e.g., <the, a, of, and, that> etc., Figure 6). These words are 136 7. Recommendations for teaching reading and writing in the FL classroom difficult to guess from the context and are equally difficult for teachers to sound out or to illustrate. It has been claimed that the 100 most frequently used words represent 50 % of all written text (e.g., Kress & Fry, 2015). Automatic recognition of these words will help children comprehend and create text more fluently (Shin, 2017). Students may also keep their own word banks (i.e., a list of words which relates to the core content and academic material the student is learning) or dictionaries with these words along with other words they choose to include (see also Kierepka, 2020). Figure 6: A list of fifty sight words for the English classroom (adapted from Peregoy & Boyle 2013: 205) Often, wall charts are employed which include themed words and phrases. Figure 7 illustrates a so-called “weekly word wall” which was employed in the Bili-50 program, dealing with the topic “nature and life”, more specifically contrasting body parts of animals and humans in grade 4. As the term “weekly word wall” suggests, the terms on this chart change on a weekly basis, and the teachers and students work together to determine which words should go on the word wall. These words provide reference support for learners during reading and writing activities. Specifically, they may help students see patterns and relationship in words, thus building phonics and spelling skills. 137 7.2 Creating a literacy-rich environment Figure 7: Weekly word wall (Bili-50, grade 4) The second example is from grade 4 (Figure 8). The students in Baden-Württem‐ berg have to pass a test to obtain a bicycle license, which is an obligatory part of the curriculum for the subject MeNuK (Mensch, Natur und Kultur, Ministerium Baden-Württemberg für Kultus, Jugend und Sport, 2016b). Together with the students, the teacher labeled all the parts of a bicycle in order to facilitate FL discussion in class and to support students with their FL written assignments. Figure 8: Parts of a bicycle (Bili-50, grade 4) Books constitute the most important resource in a literacy-rich environment. An in-class library is ideally organized in such a way that children can easily find the books and put them away after use (Peregoy & Boyle, 2013), and it is even more ideal if books can be borrowed over the weekend (see also Cameron, 138 7. Recommendations for teaching reading and writing in the FL classroom 2001: 157). Such a library should, according to Peregoy & Boyle (2013: 196 f.), not only contain normal textbooks, but also books created by the children, parents or teachers themselves. The books should cover various genres (e.g., poems, short stories, whole books, songbooks and rhyming books) as well as a variety of topics, for example books related to theme studies, alphabet books, author-of-the-month books, big books, children’s journals, collections of songs or poems children have learned at school, cookbooks, craft books, dictionaries and encyclopedias, easy-to-read books for emergent learners, fables and folktales, holiday books, informational books (non-fiction), joke and riddle books, photo album books with pictures labeled, picture storybooks, poetry books, realistic literature, recipe books, sign language books, traditional literature (e.g., nursery rhymes and fairy tales) and even wordless books (see Figure 9). Bookshelves are often organized with labels by subject, author, title, or size (tall, wide, and small). In each case, alphabetical order may provide the sequence. Some teachers put books in color-coded bins according to topics, authors, or level of reading difficulty. For example, single-book display racks may showcase a special theme book, the author of the month, or a current story or poem. In sum, all good literacy programs build on children’s fascination with books (Peregoy & Boyle, 2013). Figure 9: Library corner (grade 3, Bili-20) courtesy of Susanne Wittmann. 7.3 Scaffolding for FL reading and writing activities Scaffolding is a key concept in education, not just in CLIL; but in such a context, it is of even greater importance. If learners successfully access content through a vehicular language, then that content must be carefully scaffolded. According to Van der Pol, Volman & Beishuizen (2010), the scaffolding metaphor was adopted 139 7.3 Scaffolding for FL reading and writing activities by Wood, Bruner, & Ross (1976) to explain the role that adults can play in joint problem-solving activities with children. The concept is based on Vygotsky’s notion of the “Zone of Proximal Development”, and the emphasis is on providing assistance to enable learners to reach beyond what they are able to achieve alone (see e.g., Van der Pol et al., 2010 for a critical discussion of the concept of scaffolding and its effectiveness for L1 teaching). Learners initially need support and guidance until they are able to perform the activity themselves unaided. Thus, to be of benefit, the scaffolding is always temporary, so that responsibility for the learning process is gradually transferred to the learner, with the aim of promoting autonomous learning (e.g., Kersten, 2019). In the literature, a variety of scaffolding techniques are listed which, however, differ in categorization and terminology. In the following, the focus will be on procedural, verbal and instructional scaffolding (following the categorization by e.g., Echevarrìa, Vogt & Short, 2010), including content scaffolding as well (e.g., Ioannou-Georgiou, 2011; Massler & Ioannou-Georgiou, 2010), which mainly relates to the CLIL classroom. 7.3.1 Verbal scaffolding Verbal scaffolding may be inputor output-related. In general, it “refers to ways and methods to make […] input comprehensible for the students. It also denotes methods which enable students to participate in the lesson actively and meaningfully, despite any limitations imposed on them by their current linguistic competence” (Massler & Ioannou-Georgiou, 2010: 61). For speech, input-oriented verbal scaffolding relates to the use of level-appropriate speech (e.g., adjusting the rate of speech, pauses between phrases, repeating and paraphrasing, purposefully using synonyms and antonyms), animate language (e.g., miming, gestures, facial expressions, intonation of speech) and redundancy (e.g., repetition of key words or phrases, simple paraphrasing and synonyms already known by the students, Massler & Ioannou-Georgiou, 2010: 63-66). The aim of output-oriented verbal scaffolding is to help students both in input comprehension (e.g., reading) and output production (e.g., writing). Peregoy & Boyle (2013: 275) use the term “literacy scaffolds” to refer to such temporary frameworks that allow beginning and intermediate FL learners to construct or comprehend a written message. These instructional strategies may help students read or write whole, meaningful texts at a level somewhat beyond what they could do on their own. In general, literacy scaffolds include elements that are predictable as a result of repetition of language patterns or routines. For example, teachers may offer key vocabulary, sentence starters or pre-formulated 140 7. Recommendations for teaching reading and writing in the FL classroom sentence chunks on the board (e.g., Massler & Ioannou-Georgiou, 2010: 68, see also Figure 10). Literacy scaffolding may also relate to supportive error correction (see chapter 7.8) and allowing students to code-switch (e.g., to enable beginners to write in all languages that they know, e.g., Massler & Ioannou-Georgiou, 2010; Ioannou-Georgiou, 2011; Peregoy & Boyle, 2013). Other types of literacy scaffolds include journals and clustering (as prewriting activities), which provide support for beginning-level writers, preferably includ‐ ing peer feedback. Finally, Peregoy & Boyle (2013: 275) pointed out that the writing process itself may be viewed as a powerful scaffold because it breaks a complex process into smaller sub-processes. Figure 10: Worksheet as an example for output-oriented verbal scaffolding (Bili-50, grade 3). 7.3.2 Content scaffolding Particularly for the CLIL classroom, content scaffolding refers to techniques which can assist and support students’ understanding of and engagement with the FL content (see Massler & Ioannou-Georgiou, 2010: 62 and 70 ff, see also 141 7.3 Scaffolding for FL reading and writing activities 16 Many instructional methologies that Snow pointed out in the 1990s (e.g., Snow, 1990) for IM teaching are nowadays commonplace in any FL classroom, for example the use of body language and objects, comprehensible FL input, indirect error correction, a variety of teaching methods and types of activites, or the use of clarification and comprehension checks (e.g. Fleckenstein et al., 2020; Legutke et al., 2009). Echevarrìa et al., 2010; Ioannou-Georgiou, 2011; Snow, 1990 16 ). On the one hand, content scaffolding may be employed to help learners to understand content concepts. For example, teachers may refer to previous knowledge and experiences in order to link the known to the unknown and thus provide a scheme of reference for new material to increase students’ reading compre‐ hension. Teachers may also define, display and review content and language objectives, for example, by using the KWL chart (Olgle, 1986), where K refers to what the students Know, W to what the students Want to know, and L what the students have Learned. The students complete the K and W sections at the beginning and the L section at the end of the lesson (see Figure 11). Figure 11: KWL chart (adapted from Gondová, 2015: 159). K: What do the students Know about the topic? W: What do they Want to know? L: What have they Learned? Moreover, CLIL teachers foster active discovery of concepts (e.g., they provide a variety of activities such as hands-on materials or manipulatives (e.g., literacy activities which also include observing, deciding, experimenting, cutting, gluing, painting, selecting, drawing) for students to explore new content knowledge and stimulate learning. Finally, teachers review key vocabulary (see also chap‐ ter 7.3.1) and key content concepts during lessons, keeping in mind that CLIL learners need more time and opportunities to practice and revise content (e.g., Burmeister, 2006; Ioannou-Georgiou, 2011; Massler & Ioannou-Georgiou, 2010; Snow, 1990). Such opportunities can be provided through permanent or temporary displays of key words and concepts (see chapter 7.2). 142 7. Recommendations for teaching reading and writing in the FL classroom In addition, content scaffolding may also be used to explain tasks (Massler & Ioannou-Georgiou, 2010: 72 ff, see also CLIL4Children, 2018; Snow, 1990). Most importantly, CLIL teachers give clear instructions. Using these, as well as predictability, in instructional routines (e.g., lesson openings or closings, assignments or homework) helps provide CLIL learners with more support and structure so they can more easily anticipate meaning when they do not understand the FL. Furthermore, CLIL teachers also provide a concrete model of the process, task, or assignment. For example, giving complex explanations or instructions for doing an experiment may often be beyond the scope of students‘ FL competence. For beginning learners, therefore, teachers often rely on gestures, real objects and modelling to give clear instructions. In Figure 12, the setup of an experiment is illustrated by pictures including labels so that children can follow each step when they do an experiment on their own without having to rely on oral instructions in the FL. Figure 12: How to do an experiment (Bili-50, grade 4). 143 7.3 Scaffolding for FL reading and writing activities Finally, it is also essential to check students’ understanding of task instructions, as students often struggle to evaluate which pieces of information are important to remember. Such comprehension checks may include asking a student to explain the instructions to the rest of the class (sometimes perhaps in the L1), or by asking a pair or group of students to carry out a model task for the rest to see how it works. 7.3.3 Learning process scaffolding and procedural scaffolding Learning process and procedural scaffolding may be characterized as general strategies employed in the classroom, independent of the language or subject. Learning process scaffolding refers to techniques which assist teachers in supporting students’ working processes as well as their learning processes. They can involve strategies to improve learning or to develop reflection skills which promote learner autonomy (e.g., Massler & Ioannou-Georgiou, 2010: 67-68). Within the American context, learning process scaffolding is also called instructional scaffolding and refers to any tools which support learning (e.g., Echevarrìa et al., 2010). In the FL or CLIL classroom, teachers may use graphic organizers, e.g., timelines, flow charts, semantic maps etc. to provide students with structures in which they can write down the information they interpret from a picture. Mnemonics strategies include short poems, tongue twisters or special words which help students remember key concepts through associations (a classic mnemonic is, for example, My Very Excellent Mother Just Sent Us Nine Pies, which is a useful mnemonic for CLIL science lessons for students to remember the planets, in order of mean distance from the sun). In addition, reading techniques such as scanning (reading to find specific information) and skimming (reading to get the main idea) are important strategies so that students are able to focus on the information they are looking for, without being distracted by unknown words, for example when they fill in a chart on a specific topic. Finally, content-specific working strategies are also a learning process scaffold, for example, when students in a science lesson carry out a survey and present the results of this survey in a chart (e.g., CLIL4Children, 2018: 47). According to Echevarrìa et al. (2010), procedural scaffolding is also one of the key concepts in education and relates to grouping techniques and activity structures. Grouping techniques may include anything from individual to pair to group activities. Cooperative learning activities are nowadays an integral part of teaching. Such activities may, for example, include role plays, while others may relate to dictoglosses, jigsaws, or simulations (e.g., Echevarrìa et al. 2010). Teachers may also partner students for FL reading and content activities with more experienced 144 7. Recommendations for teaching reading and writing in the FL classroom readers who can assist those with less experience. In a think-pair-share (TPS) activity, the students first individually responds to a CLIL or EFL teacher’s question by jotting down key ideas, key words or completing a graphic organizer (think); the students then tell a partner what they have written down and both will write some more (pair). In the share-stage, the pairs meet with other pairs, compare notes, do some more writing and finally present their results to the rest of the class (e.g., Met, 2002). It is often noted that TPS activities are particularly beneficial in producing output when the questions are targeted at students’ higher order thinking skills (e.g., “How many sums can you think of to which the answer is 64? ”) instead of aiming at their lower order thinking skills (e.g., “What is 16 times 4? ”, e.g., Leung, 2015: 127). Finally, another common group activity relates to process FL writing, which focuses on the process of creating writing and on learning the stages of the writing process (i.e., pre-writing, drafting, revising, editing, and publishing), rather than on the end product (Peregoy & Boyle, 2013: 254). Many of these stages involve peer activities. For example in the prewriting stage, students may brainstorm, read and discuss ideas for writing in groups. At a later stage, peer editing groups may correct spelling, grammar, punctuation, and mechanics. Finally, students’ writing may be shared in many formats (e.g., papers placed on bulletin boards or shared in school book fairs) and not only with their classmates and other students but also with parents (e.g., Peregoy & Boyle, 2013: 259). Personal topics are preferred for process writing (such as “I remember”) because personal memories work well with young learners and may also enhance personal relationships among students as they share their experiences. In sum, even young FL students will grow both in writing and overall English language development when the exposure to books and print is combined with opportunities to write often; when strategies are offered to solve problems in writing; and when writing is shared and published (Peregoy & Boyle, 2013: 263). 7.4 Regular FL literacy activities According to Peregoy & Boyl (2013), daily classroom routines can enhance children’s awareness of the forms and functions of print and may highlight how literacy serves everyday purposes. At the same time, attention should be drawn to the actual processes of reading and writing as teachers go about daily activi‐ ties. One classroom routine is the “morning message”, in which teachers preview the day’s activities for young students (Peregoy & Boyl, 2013). Many teachers start class by writing the day’s activities on the board (see Figure 13). As they say the words, they model the organizational/ mnemonic function of writing 145 7.4 Regular FL literacy activities and the form (i.e., the left-to-right, letter-by-letter sequence corresponding to the spoken words), which is particularly beneficial for emergent readers. Figure 13: Activities of the day (Bili-70, grade 1). Alternatively or additionally, in many CLIL classrooms in Germany, the day is started with the weather routine (Burmeister, 2006, see also Cameron, 2001: 143, see Figure 14). If more or less the same routine happens for each student in the class, the written text becomes part of the event and will be learned through participation in the event. Figure 14: Weather routine (Bili-70, grade 2). 146 7. Recommendations for teaching reading and writing in the FL classroom Another routine that lends itself to a functional literacy learning opportunity includes classroom rules and procedures that teacher and students establish together at the beginning of the year (Peregoy & Boyle, 2013). As decisions are made, the rules and procedures for duties (such as table cleaning, floor sweeping, pet care, and any other routine chores) are written down on a large chart (see Figure 15). In this way, the students see the spoken words written on the board. The duties are often illustrated with a broom, an animal, a table, or other appropriate pictures to support student understanding of the chart. By highlighting print uses during day-to-day routines, the teacher makes FL literacy natural and unintimidating so that children begin to read without even knowing it. This is often also accompanied by phrases commonly used in the classroom, such as utterances for praise, blame, correction, etc., which may also be displayed on a wall in the classroom. Figure 15: Classroom duties (Bili-50, grade 2) 147 7.4 Regular FL literacy activities Many authors suggest that teachers should read aloud to their students on a regular basis (e.g., Cameron, 2001: 141; Peregoy & Boyle, 2013) because the stu‐ dents are then involved in the pleasure function of print, the reading is modeled for them and they develop general knowledge and literary notions about story plots and characters. Several techniques are used by teachers to facilitate FL comprehension because following a story line places heavy cognitive-linguistic demands on young listeners in terms of attention, comprehension, and memory. For example, the teacher may stop at certain places in the book to discuss a picture as it relates to the story or to review the plot. The teacher may also focus on comprehension by asking prediction questions. If the book is short and simple, repeated readings will assist in comprehension. Most importantly, the reading-aloud moments should be a special time when students feel comfortable to simply sit and enjoy listening to stories in the FL (Peregoy & Boyle, 2013). Beyond these conventions of print, reading aloud familiarizes children with the language of written English (Cameron, 2001: 141), e.g., formulaic openings (Once upon a time …), closings (… and so they all lived happily ever after), and the patterns of text types (e.g., stories and information texts). Affectively, reading aloud can motivate children to want to read themselves. Sometimes students are also asked to read aloud to the class (Cameron, 2001: 142). On the one hand, this can help learning, but on the other hand, the child reading aloud to the whole class may not speak loud enough for all to hear and, if s/ he stumbles over words, the other children will lose the meaning and probably also the motivation to listen. Reading aloud tasks may be particularly painful for weak readers. Therefore, paired reading, where children take turns reading to each other in pairs, may be more helpful. It is nevertheless important that children regularly read aloud individually to their teacher, since it is only by listening carefully to how children are making sense of written FL words that teachers can understand their progress in learning to read in the target language. Finally, the “Daily 5” (Boushey & Moser, 2014; Burmeister, Muller & Nistl, 2018) is a relatively new concept for structuring literacy time with the aim that students develop lifelong habits of reading, writing, and working independently. The idea is that students select from five authentic reading and writing choices and work independently toward personalized goals, while the teacher meets individual needs through whole-group and small-group instruction, as well as one-on-one conferring. The Daily 5 contains five activities in which students read to themselves, work on some writing, read to someone, listen to reading, or work on words. In Germany, the usefulness of this framework for the CLIL classroom has unfortunately not been examined yet, which is also due to the 148 7. Recommendations for teaching reading and writing in the FL classroom fact that in its original version, three 7to 10-minute rounds are allocated each day, which is not feasible even within the time frame of CLIL lessons at the elementary school level. Informal interviews with FL teachers in Germany, however, indicated that they found their students’ reading and writing skills to increase when only completing one task for five minutes a day over an extended period of time. 7.5 Awareness-raising literacy activities Frisch (2013) emphasized the benefits of awareness-raising techniques that focus students’ attention on a specific linguistic phenomenon. In such activities, the students receive sample material on a particular FL linguistic problem and are then asked to describe what they notice. This is a direct teaching method in which learners independently discover regularities in the FL (i.e., an inductive approach, see e.g., Ellis, 2008: 17). Already in grade 2, students can be taught how some letters are pronounced in word-initial position. It may facilitate their reading performance knowing that <c> is usually pronounced as [k] as in <cat, cow, crocodile>, <sh> as [ʃ] as in <shirt, show, shine> or <j> as [dʒ] as in <jump, jeans, jam>. Such knowledge may be particularly useful for students when reading aloud in class (e.g., Kierepka, 2020; see also Ur, 1996). Another example of an awareness-raising activity may focus on the pho‐ neme-grapheme relation of the letter <a> and the sound [æ] in English (see Frisch, 2010 on the procedure in the EFL classroom). The learners are presented with words on flash cards (e.g., <cat>, <map>, < pack>), illustrating the <a> - [æ] grapheme-phoneme correspondence. The learners then verbalize the rule for this relationship in English. In a next step, the learners describe to what extent this English phoneme-grapheme relation differs from the German one (in which the letter <a> is pronounced as [a], as in <Pack>, <Matte>, or <Kappe>, e.g., Ossner, 2008). In doing so, the learners provide a cognitive abstraction to understand the linguistic phenomenon. Finally, the learners are encouraged to summarize their findings in the form of a rule (i.e., the short-vowel rule which states that a vowel usually has a short sound when one-syllable words have a vowel in the middle), for example on a learning poster where students collect English words in which the letter <a> is articulated as [æ]. The teachers in the current study also like to draw students’ attention to words in German and English that share similar spelling, meaning, and pronunciation (i.e., cognates). Such words include pairs like <arm - Arm> or <finger - Finger>, which are important words for the elementary school 149 7.5 Awareness-raising literacy activities level. Cognate awareness benefits learners by providing a bridge of familiarity between one language and the next, but it can grow in the level of sophistication and difficulty. Cognate awareness may also, however, lead to negative transfer (Biebricher 2012: 108), for example, when words are spelt identically but pronounced differently (e.g., <Tiger> and <tiger>, Frisch, 2013: 203). This is one of the reasons why Rymarcyk (2008) pointed out - as did many other authors before her - that the pronunciation of an English word should be consolidated before its written form is introduced, at least for the elementary school context. However, because experience showed that children start to write and read in any language as soon as they are able to, FL teachers should not wait too long to introduce the written form of words in order to avoid fossilization errors. In the field of teaching methodology, and closely related to learning and working techniques, awareness-raising activities may also relate to learning processes in more open forms of teaching, such as free work, learning stations or project learning. In the Bili-70 program, learning stations are frequently employed in CLIL lessons on math and science. For example, students in grade 2 dealt with multiplication and division as part of the requirements of the curriculum for the subject math (Berlin Senatverwaltung, 2016). In one of the mathematics lessons, four learning stations were set up: students (six per group) played a board game on the first table, which comprised several tasks on multiplication and division. On the second table, the students individually worked on more complex addition tasks. On the third table, students did pair work, matching several questions and answers on multiplication on a worksheet and providing explanations. Finally, the students at the last table solved division tasks using hands-on material and formulated ideas as to how to deal with division problems with odd numbers (e.g., 9 / 2 = ? ). At these learning stations, the students described mathematical phenomena (table 3), formulated possible rules (table 4), revised and extended their previous knowledge of another topic (table 2), received a multisensory experience while learning content (table 4) and combined the joy of playing a board game with content tasks (table 1). In addition, the children almost exclusively used English with each other during this lesson but also worked on their FL reading skills as all tasks were presented on worksheets in English. 7.6 Relevant and authentic literacy activities As the previous examples showed, students in CLIL classrooms learn both content and language. They also train their reading skills in authentic and 150 7. Recommendations for teaching reading and writing in the FL classroom relevant contexts. This is one of the reasons why Rymarczyk (2008, 2016) and many other authors have argued that any literacy activities should be embedded in genuine reading and writing opportunities. These may include students predicting the meaning of individual words and phrases in context to increase understanding and to free new capacities for new and difficult combinations of letters. For example, fourth graders in the Bili-50 program were required to present a poster on bats in a group of 3-4 students (Figure 16) as part of the curricular requirements for the subject combination Mensch - Natur - Kultur (Ministerium Baden-Württemberg für Kultus, Jugend und Sport, 2004, 2016b). In this project, which lasted for approximately four weeks, each group had to decide on a particular type of bat, collect information, prepare the poster and present a five-minute talk. Consequently, this task trained the traditional five skills, i.e., reading (for background informationand revising the poster text), writing (e.g., providing information on the poster and taking notes for the talk), listening (e.g., to YouTube clips on bats and to their peers in class), speaking (i.e., practicing the talk, interacting in group discussions in class), and mediation (as some groups found background information on this topic in German, which they translated into English.). Figure 16: Posters on bats (Bili-50, grade 4). In this context, the students are also provided with information as to how they should present their topic to class, as depicted in Figure 17: 151 7.6 Relevant and authentic literacy activities Figure 17: Golden rules for presentation (Bili-50, grade 4) According to the teachers, the students were very motivated to present their findings, and such tasks also increased students’ motivation to produce FL texts. It is one of the advantages of the CLIL classroom that the students are continuously exposed to authentic and relevant FL texts which cover a wide range of topics and genres that may in turn increase students’ motivation to engage in FL literacy activities. Of course, similar projects may also be carried out in the regular FL classroom, although in a less elaborate way, including, for example, traditional topics such as the students’ family or body parts. 7.7 Some notes on FL spelling Spelling is a rather neglected area in FL learning, particularly for the elementary school classroom; little research has been conducted on the question of which FL spelling activities may actually be beneficial for young learners. This section is devoted to selected FL spelling activities which have been reported to posi‐ tively affect students’ FL spelling skills. Before these activities are introduced, however, it is worth taking a closer look at young students’ spelling errors in the FL. 7.7.1 Invented spelling As noted above, it is often the case that young learners start to write in the FL simply because they are able to do so in German. However, their English spellings do not always correspond to conventional English orthography. Such errors have often been referred to as “invented spelling” (e.g., Rymarcyk & Musall, 2008; Rymarczyk, 2010, 2016), based on Read’s (1975) observation that L1 native English preschoolers commonly “invented” spellings for words in their 152 7. Recommendations for teaching reading and writing in the FL classroom daily vocabulary by rearranging letters to fit their perception of the rules of the English language. Such misspellings constitute the focus of Steinlen & Piske’s study (2020, see chapter 6.1.2). They examined English spelling skills of 117 fourth graders attending the Bili-50 program. The students produced a short free text on the topic “Yesterday”, for which five sentences were required. On average, the children spelled 84 % of the words in their texts correctly, i.e. five spelling errors in 31 words. The fourth graders may thus be characterized as transitional spellers (see Peregoy & Boyle, 2013: 210). Two examples are provided in Figure 18. As Figure 18 illustrates, the children had already begun to develop their own spelling rules for English words, which, according to Peregoy & Boyle (2013: 184) “represent their logical, developmental hypotheses about how to spell” (see also Parlindungan, 2018). On average, of the five spelling errors which occurred in children’s English text production, three errors related to substutition, i.e., two errors on upper and lower case (e.g., *<Homework> instead of <homework> and *<i> instead of <I>) and one on vowel or consonsant substitutions (such as *<startet> instead of <started>). Additionally, one error was due to omission (e.g., *<yumy> instead of <yummy>), and one word contained multiple errors 153 7.7 Some notes on FL spelling (e.g., *<Frend, scholl> instead of <friend, school>). The examples are all taken from the two texts in Figure 18. Following James (1998), Steinlen & Piske (2020) attributed misspellings in the English texts to either intralingual (developmental) or interlingual reasons. In their sample, interlingual errors (i.e., transfer from German) were slightly more frequent than intralingual ones (which may also occur in texts of beginning native English writers). Misspellings due to negative transfer from the first to the target language have been noted in many studies conducted in Germany (e.g., Luelsdorff, 1991; Rymarcyk & Musall, 2008; Rymarczyk, 2010 for EFL programs, Burmeister, 2010, for bilingual programs) and elsewhere (e.g., Howard, Green & Arteagoitia, 2012; Swain, 1975, and see Figueredo, 2006 for a review). In Steinlen & Piske’s study (2020), interlingual errors included, for exam‐ ple, insertions (*<sche> instead of <she>), omissions (*<wen> instead of <when>), and transposition (*<tabel> instead of <table>). These were, how‐ ever, far less frequent than substitutions of individual sounds and graphemes: of these, errors of upper and lower case made up the largest part (61 %). Such errors mainly concerned the word <I>, which was written in lower case, as in German (see also Figure 18). Furthermore, English nouns from the areas of sports activities, school and family were often written in upper case instead of lower case (e.g. *<Football, Mother, School>). A look at the free text productions from grade 3 indicated that errors regarding upper and lower case were less frequent (with phonetic spelling being more frequent in grade 3 than in grade 4). In informal interviews, the teachers of the present study recalled that the students in grade 4 had revised the topic upper and lower case letters in German-as-a-subject lessons and had apparently transferred the German rules to English. The CLIL teachers therefore decided to explicitly revise the same topic for English as well, which resulted in students immediately improving their capitalization skills. Another group of substitution errors reported by Steinlen & Piske (2020) based on transfer from German to English in student texts regarding pho‐ neme-grapheme rules, e.g., when <y> was replaced by *<j> (*<jesterday> instead of <yesterday>), or <c> by <k> (*<kome> instead of <come>). Spellings such as *<thet> (instead of <that>, see Figure 18) indicate that English auditory input had been treated as German and the corresponding German phoneme-grapheme rules had been applied, because the English short vowels / e, æ/ are notoriously difficult for Germans to distinguish (see also Luelsdorff, 1991). Yet another group of substitutions was due to the final devoicing rule in German, where, for example, verbs in the simple past whose 154 7. Recommendations for teaching reading and writing in the FL classroom <-ed> ending is pronounced as a voiceless sound (e.g. *<startet> instead of <started>, see Figure 18). The anecdotally-reported use of German umlauts (e.g., Burmeister 2010) was only found in one word (*<börstäg> instead of <birthday>) in Steinlen & Piske’s corpus. In conclusion, transfer errors are very frequent in FL writings, independent of the intensity of the FL program (see e.g., BIG-Kreis, 2015; Rymarcyk & Musall, 2008; Rymarczyk, 2010, 2016 for EFL programs or Burmeister, 2010, for bilingual programs). However, as Figueredo (2006) pointed out, transfer errors in spelling may change as a function of students’ first-language proficiency level, FL proficiency, and their age of learning. It is far from clear, though, what these changes imply for FL text production in elementary schools, because most studies involve older learners (see Figueredo’s review, 2006). Thus, more research is warranted in order to determine how the quality and the quanity of interlingual FL spelling errors change over time. Apart from negative transfer, Steinlen & Piske (2020) also noted that almost half of the misspellings in the children’s English texts were intralingual er‐ rors, namely developmental errors (see e.g., James, 1998). There are many authors who reported children’s idiosyncratic writing to support their literacy acquisition process, particularly for L1 German or L1 English (e.g., Gentry, 2006: xiii; Scheerer-Neumann, 1998: 56; Steinig & Ramers, 2020). In Steinlen & Piske’s (2020) sample, such errors included omisions of letters in double consonant groups (e.g., *<tenis> instead of <tennis>, which is also found in German students’ German texts), omissions of silent letters (e.g., *<climing> instead of <climbing>), omissions of final <e> (e.g., *<something> instead of <something>), homophones (e.g., *<there> instead of <their>) and, most prominently, omissions of vowel graphemes in vowel groups (e.g., *<Tusday, brekfast> instead of <Tuesday, breakfast>). In these examples, the students apparently tried to adjust spelling and pronunciation or based their spelling on analogies from spellings of other words (e.g., *<breakfeast>). If transfer from the first language constitutes an importance source for mis‐ spellings in FL texts, how did this affect the production of FL texts by minority language students’ for whom English is the L3? Interestingly, the minority language children in Steinlen & Piske’ study (2020) committed the same quantity and the same quality of spelling errors as their majority language peers; that is, there were no significant differences in terms of the number of misspellings or in terms of the quality of interlingual and intralingual errors between the two groups. This result indicates that minority language students also relied on German (which was their L2). Based on Williams & Hammarberg (1998), Steinlen & Piske (2020) argued that this finding was 155 7.7 Some notes on FL spelling probabaly due to the fact that German was the language which the minority language students were (orthographically) more proficient, compared to their family language. Moreover, German was the language which they had used most recently (namely in the school context) and which, in most cases, is typologically more similar to English than their family language (e.g., Arabic, Greek or Turkish). Thus, it is not the minority language students’ first language which constitutes the source of transfer regarding spelling errors in English, but rather their L2 German, paralleling findings for L3 English reading and writing (e.g., Steinlen, 2016, 2017 2018b; Steinlen & Piske, 2016a, 2018a; Steinlen 2018 a, b) and L3 English receptive grammar and vocabulary (Steinlen & Piske, 2013, 2015, 2016b). This result strongly indicates that minority language students can achieve the same level of FL spelling competence as comparable majority language students. In conclusion, the use of invented spelling as such, according to Rymarcyk & Musall (2008: 71), is “neither extraordinary nor alarming” and also extends to FL writing. However, unlike the name implies, “invented spellings” are not arbitrary, as this literature review indicated. Regarding FL spelling, the learners not only followed their perception and knowledge of the rules of the target language (like their native English peers) but, when in doubt, they also resorted to the language that they were orthographically most familiar with. Additional research is needed in order to examine the role of “invented spelling” and its consequences for the FL classroom in a more systematic way, especially because the quantity and quality of FL intraand interlingual spelling errors are likely to change as a function of students’ age, their cognitive maturation and their language proficiency in the L1, L2 and L3 (if applicable). Spelling instruction has been found to be very effective, not only for native speaker students (e.g., Graham & Santangelo, 2014) but also for FL students (e.g., Pérez-Cañado, 2006). This is why some recommendations on FL spelling activities are provided in the next sections. 7.7.2 Spelling activities There are only a limited number of papers dealing with teaching activities for spelling / writing in the FL elementary school classroom, and most of them, as Parlindungan (2018) pointed out, relate to older, more advanced FL learners. Some of the studies cited here date back to the 1990s (e.g., Homstad & Thorson, 1996; James & Klein, 1994) but these spelling activities still receive attention in publications and recommendations for FL teaching today (see Parlindungan, 2018). 156 7. Recommendations for teaching reading and writing in the FL classroom It is very common that such recommendations, geared towards beginning writers in the FL classroom (e.g., Burmeister & Piske, 2011; Cameron, 2010; Cook, 1999; Department of Education, Newfoundland and Labrador, 2010; Parlindungan, 2018; Russak & Kahn-Horwitz, 2015; Shin, 2017; Wilson, 1989), focus on explicit teaching of sound-symbol correspondence (i.e., phonics), spelling rules (see below), and a variety of spelling strategies (e.g., rhyming, patterning, phonemic segmenting, and analysis), which should be carried out as frequently as possible, particularly for FL learners in the CLIL classroom. Prediction, confirmation and integration also play an important role in spelling strategies. For example, in analogy-based word activities, students are made aware of the fact that when they know the <-at> rime (as in <cat>), then they also know how to read and spell <hat, mat, sat>, and <pat>. Additional spelling activities relate to phonological awareness tasks, which involve rhyming, clapping syllables, and deleting initial and final sounds of FL words. In word sorts, students sort words for one commonality, such as rhyme or alliteration, or pick a word that does not belong. Similarly, word works involve writing words in lists, manipulating letters to complete the words, or cloze texts. Young learners also like to engage in spelling-based games, such as Bingo, homophones (“Teekesselchen”), or crosswords. Finally, more advanced FL students are also trained to analyze new vocabulary in terms of verb and noun endings, and they are also frequently asked to edit their peers’ texts. In Cameron’s view (2010: 153 f.), any focused teaching about written language forms should comprise five steps, i.e., to start from a meaningful context, to focus the students’ attention on the key feature, to give input (e.g., examples or a rule to provide varied practice), and finally, to give students plenty of opportunities to apply their new knowledge and skills in different, meaningful contexts. Cameron’s activity revolved around the rime <-ai-> and was aimed at 9-10 year-old students with a Malaysian background. The students first listened to a particular poem on a postman and saw a picture. Next, the written version was used, with the <-ai-> rime highlighted in red at the end of the lines (e.g., <snail, tail, rail>). Then the words containing <-ai> were taken from the chant and presented in isolation on the board. This was followed by a quick game-like activity in which the students placed various initial consonants next to the rime. Finally, the students’ task was to compose a new chant, using other (known) words like <sail, pail, jail> along with some of the original words. Such short, intense activities may, according to Cameron, also be used to fill gaps in children’s knowledge and skills, or to highlight areas in which the FL literacy works differently from the L1. 157 7.7 Some notes on FL spelling Some additional activities relate to spelling strategies. As James & Klein (1994) pointed out, these include the phonemic route to spelling (where the words are spelt as they are pronounced), the visual route to spelling (i.e., a visual memory strategy, where the word is retrieved as a whole, using a visual imagery) and the use of metalinguistic knowledge (which includes significant regularities and conventional irregularities, i.e., spelling rules. Good spellers not only rely on the most frequently used phonemic route but utilize metalinguistic knowledge and visual imagery of the words as well. Thus, these spelling strategies should also be trained explicitly in the FL classroom. First, the phonemic route to spelling may be trained with pronunciation tasks, because the better the FL learners are able to produce FL words, the better they are able to spell them. In order to improve the visual route to spelling, teachers may encourage students to frequently copy FL words and phrases (despite the fact that such activities are often considered to be “mindlessly mechanical”, James & Klein, 1994: 44) and to ask students to read a lot so that they become familiar with the correct visual imagery of the words, which strengthens the visual lexical entry in the memory and thus facilitates retrieval. Finally, the use of metalinguistic knowledge includes contrastive teaching when words are cognate but not identical, in order to avoid “false friends” associations. According to James & Klein (1994) the problem with metalinguistic knowledge is, however, that rules are often associated with certain words but are not sufficiently automatized, and the only solution seems to be constant practice (i.e., applying rule knowledge) which eventually leads to an automatization of rules. However, studies are still lacking which examine the effectiveness of teaching these three spelling strategies to young FL learners, particularly for those attending a bilingual program. 7.7.3 The role of spelling errors in the FL classroom The teachers of the present study also pointed out that spelling errors need to be dealt with differently in different activities. In their view, students’ written productions should not contain any errors when these texts are used as references, e.g., when displayed on walls (e.g., posters) or taken home. In contrast, errors are allowed when they occur in children’s written journals or learning diaries, which are left in school because the focus of these activities is not on spelling but on producing free texts on a certain topic (Homstad & Thorson, 1996). Although this kind of writing is not corrected, teachers may want to respond to these texts every now and then, particularly in terms of content (Cameron, 2001). More advanced FL writers (e.g., in grade 4) may 158 7. Recommendations for teaching reading and writing in the FL classroom then focus on correct spellings in the final editing stage of process-writing assignments (e.g., Homstad & Thorson, 1996). Spelling errors are, however, a good indicator of literacy skills. Many authors favor a qualitative approach to spelling errors, which, in their view, should be analyzed according to their phonetic, morphological, and ortho‐ graphic nature instead of only scoring words as correct or incorrect. Such a dichotomous paradigm for marking spelling (i.e., right / wrong) may overlook opportunities to indicate to students how their application of L1 knowledge is inappropriate in contrast to FL English norms (Fashola, Drum, Mayer & Kang, 1996). Such an approach may also reinforce orthographic principles and emphasize word-specific spellings (e.g., Bebout, 1985; Burmeister & Piske, 2011; Fashola et al., 1996). In the CLIL classroom, spelling only plays a minor role in formal assessment contexts. In most cases, spelling errors (in German or English) do not count, with the exception of words which are already given in the FL and which only need to be copied in the answer (e.g., Massler, 2011). In the EFL classroom, FL spelling is only assessed at the secondary school level, with a focus on fluency before accuracy (e.g., KMK, 2013a: 5). 7.7.4 Teaching spelling rules Ur (1996) pointed out that some of the spelling peculiarities of English are easy to implement in the FL classroom. Contrary to general belief, the large majority of words in English (about 84 %, according to Ur, 1996: 163) are spelt either phonetically or according to regular rules which are fairly straightforward to teach and practice. These include, for example, the digraphs <th>, <ch>,< sh>, <wh> (in high-frequency words such as <the, chin, ship, what>) and the less common <ph> (e.g., <physics>). In addition, final <e> causes a previous vowel to be pronounced like its alphabetic name (e.g., <late, these, time, hope, tune>). The letter <c> usually represents the sound [k] (<call, cool>) but is regularly pronounced [s] before <i, e, y> (<ceiling, cypress, cycle>, see Figure 19). In addition, the letter <g> usually represents the sound [g] (<give, go>) but is sometimes (not always) pronounced [dʒ] before <e, i> (<George, gin>), and finally the letter <q> in English is always followed by <u> (e.g., <queen, quaint>). Two examples of spelling rules which are displayed on posters in the CLIL classroom are provided in Figure 19. 159 7.7 Some notes on FL spelling Figure 19: Spelling of word-initial <k> vs. <c> and word-final <k> vs. <ck> (Bili-70, grade 3) In addition to teaching these rules, it is quite useful to give students advice on some generalizations that apply to a variety of words (Ur, 1996: 163-4). For example, the letter <z> is rare (e.g., <zoo, zombie>) and so is the letter <j>, which is pronounced [dʒ] in word-initial position (e.g., <Japan, Jim, joke>). However, [dʒ] is often spelt with <g> or <dg/ dge> (<George, dodge>). The letter <k> is relatively rare in word-initial position (e.g., <king>), and the sound [k] is usually represented by <c> (e.g., <crocodile, cow, come>), except at the end of short words, where it is likely to be <ck> (e.g., <kick>, see Figure 19 for the distinction between <k> and <ck> in wordfinal position). Moreover, a double consonant usually causes the previous vowel to be pronounced short (not like its alphabetic name), in words like <apple, filling> (as compared to <paper, filing>). However, in elementary schools in Germany, implicit approaches are preferred: this means that students acquire spelling rules by reading, for example, when they realise that in English a word can never start with a doubled consonant. Using such an approach, it is either expected that the learners abstract a rule which they subsequently apply unconsciously to new materials, or that learners’ performance is based on the frequency of the appearance of certain grapheme combinations in words to which they have been exposed (e.g., Kemper, Verhoeven & Bosman, 2012). It is not clear, though, to what extent young learners in classrooms with differing FL intensity actually benefit from explicit spelling rules or from an implicit approach to spelling; from which age onwards spelling rules should be introduced; or which measures should be taken so that this knowledge may be stored permanently in students’ long-term memory. 160 7. Recommendations for teaching reading and writing in the FL classroom In summary, spelling is important for successfully acquiring writing and reading skills in any language. As students learn the various spelling patterns of English, they can apply this knowledge to word identification during reading and writing. Attention to students’ spelling and thoughtful instruction yield benefits to students’ literacy development. However, there is still a lack of studies examining effective spelling instructions for young learners in CLIL and EFL programs in Germany and elsewhere. 7.8 Teacher feedback on FL students’ writing Teachers play an important role in the FL writing process because young students need encouragement, time and appropriate feedback to produce any kind of FL text (e.g., Burmeister, 2010; Burmeister & Piske, 2011; Cameron, 2001; Grabe & Kaplan, 1996; Peregoy & Boyle, 2013; Rymarczyk, 2010, 2016; Ur, 1996). According to Hancock & McDonald (2000), at beginning levels of FL writing development, the purpose for responding and feedback is to support student efforts. Such an example for teacher feedback on a student’s test on weekdays (and months, not provided), conducted for the subject science, is depicted in Figure 20. 161 7.8 Teacher feedback on FL students’ writing Figure 20: Example of a graded test in the subject science (first page, Bili-20, grade 2, courtesy of Silke Greubel) In other words, teachers encourage students to continue, to feel good about their writing, and to carry the activity through to completion. Teachers also respond positively to the ideas in the learner’s work, not just the language. Grabe & Kaplan (1996: 275) also recommend teachers to build a helpful spirit and give directions for positive criticism (e.g., find the funniest sentence, find two good words, or find something good about the opening). Thus, teachers place 162 7. Recommendations for teaching reading and writing in the FL classroom value on student writing regardless of the degree of accuracy and the stage of development. Young learners should also be provided with the time and resources needed to write about their own ideas, because FL writing is a time-consuming process that requires extended periods to allow for creative expression (see also Hancock & McDonald, 2000; Rubin & Galván Carlan, 2005). Resources include the classroom as a literacy-rich environment (e.g., wall papers with sight words and sentence patterns), dictionaries (monolingual and bilingual), and individual resources such as word banks and mind maps (e.g., Burmeister & Piske, 2011; Grabe & Kaplan, 1996: 274, see chapter 7.2). In order to build young FL writers’ confidence, teachers give very specific, well-scripted writing tasks and break any FL writing process into smaller steps (e.g., pre-writing, drafting, revising, editing, and publishing, e.g., Homstad & Thorson, 1996; Peregoy & Boyle, 2013). Burmeister & Piske (2011) suggested so-called “writing conferences” which focus on functional writing and on the process of writing. The goal is to have the children work individually or in small groups on written products that can be published, for example, in the school library or on the school’s homepage. The children should go through the individual phases of text production as independently as possible: from the writing preparation phase, in which ideas are generated, material is collected (prewriting) and a writing plan is drawn up (drafting); through the phases of text revision (revising); to error-free completion and publication (editing and publishing; cf. Peregoy & Boyle 2013). In the first years of learning, the teacher has an important role to play, especially in the phases of text revision. She helps the students to improve the respective draft versions, which the children write in pencil until they are error-free and thus ready for publication. This phase has diagnostic potential, as the teacher can use the errors to determine developmental stages and respond to them individually - without simply painting over the errors. It is important that the children experience the corrections as benevolent, constructive feedback. Under no circumstances should criticism lead to them losing the joy of creative writing for fear of making mistakes. In addition, mini-lessons have been found useful in which writing strategies are made conscious, or errors that frequently occur in class can be discussed. Running through such “writing conferences” takes a comparatively long time, but can contribute to a constructive and individualized handling of errors. In terms of appropriate feedback, teachers mark students’ texts positively for good and ambitious use of language. FL learners need to be told what they are doing right as well as where their own particular problems are. Apart from spelling (see above), teachers will tell students to find any sentence that is not 163 7.8 Teacher feedback on FL students’ writing clear or to think of something that might be added; they will also point out strong or weak openings, the need for more descriptive words, or the over-repetition of vocabulary (Grabe & Kaplan, 1996: 275). Any feedback on writing should, ideally, be dialogical in nature so that students are not only given the opportunity to learn, comprehend, process and negotiate feedback but also to react to it (Louis, 2017). This can be done, for one, by asking students to identify specific aspects on which they would like to receive feedback before they hand in their assignments. Peregoy & Boyle (2013: 212) advise teachers to document their students’ emergent literacy skills by keeping a portfolio that includes both reading and writing information for each student. For writing, samples from students’ journals and dictated stories may be used; for reading, teachers may include a list of favorite stories. In such a portfolio, teachers may also keep a checklist of their students’ knowledge of letters/ sounds, sight words and any other word recognition strategies they have been taught. Additional checklists may be used to document the development of students’ FL reading and writing skills in a more holistic way. These are illustrated in Table 5. According to Peregoy & Boyle (2013: 213), Levels 1-6 represent behaviors, Level 7 transitional writing, and Levels 8-11 a developmental progression of more mature writing. Similarly for reading, Levels 1-5 represent emergent reading behaviors, Levels 6-8 transitional reading, and Levels 9-11 developmental progressions of more mature reading. Scale of Writing Development Scale of Reading Devel‐ opment Level 11 Child uses a variety of strategies for revision and editing. Child uses a variety of literary techniques to build suspense, create humor. etc. Child reads fluently from a variety of books and other materials. Level 10 Child willingly revises and edits. Child writes creatively and imaginatively. Child writes original poetry. Child writes clearly. The message makes sense. Child uses commas, quotation marks, and apostrophes. Child seeks out new sources of written information. Child voluntarily shares in‐ formation with other chil‐ dren Level 9 Writing includes details, dialogue, a sense of humor, or other emotions. Spelling becomes more conventional. Child willingly revises. Child uses context clues, sentence structure, struc‐ tural analysis, and phonic analysis to read new pas‐ sages in a functional, effec‐ tive and strategic manner. Level 8 Child writes a story with a beginning, mid‐ dle, and end. Child reads unfamiliar sto‐ ries haltingly, with little adult assistance. 164 7. Recommendations for teaching reading and writing in the FL classroom Child uses different forms for several differ‐ ent purposes (narrative, expository, persuasive). Revisions include adding to the story or piece. Child uses basic punctuation purposefully and consistently. Level 7 Child writes the start of a story. Child writes several short sentences. Child uses both phonics and sight strategies to spell words. Child rewrites a familiar story or follows the pattern of a known story or poem. Child reads familiar stories fluently. Level 6 Child invents spellings. Story is a single factual statement. Message is understandable (decipherable). Child reads short texts word by word. Level 5 Child labels or makes statement about draw‐ ings. Letters have some connection to sounds. Child writes lists. Child separates words with space or marker. Child memorizes text and can pretend to read a story. Level 4 Letters do not match sounds, but child can explain written message. Child writes strings of letters. Child participates in read‐ ing by supplying rhyming words and some predictable text. Level 3 Child copies words s/ he sees around the room. Child writes letters and mock letters in a line across the page. Child writes in left·to-right sequence, top to bottom of page. Child talks about each pic‐ ture (attends to pictures, less so to the story itself). Level 2 Child copies words s/ he sees around the room. Child writes mock letters, but these may not be in any conventional sequence. Child pretends to write. Child watches pictures as adult reads a story. Level 1 Child attempts to write in scribbles or draws patterns. Child listens to a story but does not look at pages. Table 5: Scale of Writing Development and Scale of Reading Development (adapted from Peregoy & Boyle, 2013: 213). According to Peregoy & Boyle (2013), these reading and writing development descriptors may be used to document and evaluate individual student progress. They should not be considered as lockstep sequences because students develop 165 7.8 Teacher feedback on FL students’ writing in individual ways: some levels may be skipped or intermixed. These develop‐ mental checklists are used in such a way that teachers put a check beside each statement that describes the child’s behaviors exhibited during a particular observation period. By collecting and marking these observations over time, the teacher will be able to portray each student’s progress, communicate their progress to parents and others, and adjust her instruction accordingly. Students may also work on a portfolio individually (e.g., Babocká, 2015: 182 f). In summative terms, a portfolio is product-oriented and judgmental because it comprises the results of a learner’s work, such as written work, projects, test results, essays, drawings, notes etc. Self-assessment and peer-assessment sheets, teacher comments and parent comments are included in a portfolio, too. At the same time, it can be considered a formative assessment, because this collection of samples of a learner’s work is purposeful, systematic and ongoing. Thus, a learner portfolio offers “a more complete picture of a child’s [learner’s] work and development than any other assessment technique” (Ioannou-Georgiou & Pavlou, 2003: 9). The European Languages Portfolio (Council of Europe, 2001, 2018) is the most well-known one. It consists of a language passport (where students fill out personal details), a language biography (where personal goals and competences are logged) and a language dossier (where all good language work is logged both in and out of school). However, this type of portfolio only assesses language and not content. As the comparative self-assessment sheet in Figure 21 illustrates, such a portfolio can also be adapted to the CLIL classroom, where it relates to both language and content. Figure 21: Example of a comparative self-assessment sheet for CLIL classes (adapted from Babocká, 2015: 183). 166 7. Recommendations for teaching reading and writing in the FL classroom Massler, Stotz & Queisser (2014) designed a model called CLILA (Content and Language Integrated Learning Assessment) to strive for differentiated assessment in the elementary school CLIL classroom, which takes into consideration three important dimensions of CLIL, namely the topics in the domain of subject teaching, the skills and competences in the various CLIL subjects (subject knowledge), and communicative competences in the FL. In this model, FL skills are measured by using the Common European Framework of Reference (Council of Europe, 2001), whose levels are fine-grained with an additional instrument called Lingualevel (Lenz & Studer, 2007). Based on the topics of the subject science and the competences that students should attain in science, Massler et al. (2014: 143) identified descriptors for key activities in this subject (which are part of any CLIL classroom and at the same time represent desirable learning outcomes). They then linked these descriptors (e.g., distinguishing, observing, describing, measuring, estimating, matching, ordering, comparing, displaying, designing, orienting oneself, defining, analyzing, interpreting, explaining and carrying out) to more fine-grained levels as provided by Lingualevel (i.e., A1.1, A1.2, A2.1, A2.2, etc.). The CLILA model thus consists of a framework of dimensions and descriptors, which serves as a basis to generate suitable assessment scenarios, each of which is made up of several tasks (some of them may be found in Massler et al., 2014: 144 ff.). According to Massler et al. (2014: 147), the CLILA model may help CLIL teachers to understand what good CLIL teaching and assessment practices consist of, because it exemplifies how language and content-related competences can be described right from the beginning of the learning process. Furthermore, the sets of task prototypes for the diagnostic, formative and summative assessment in CLIL teaching are examples of how content and language learning can be integrated meaningfully. Many additional examples on assessing language and content for the elementary school CLIL context can be found in Massler, Stotz, Rehm & Rimmele (2016). 7.9 Conclusion on recommendations In sum, systematic instruction and plenty of practice (e.g., awareness-raising activities, The Daily 5, explicit teaching of sound-symbol correspondence and spelling strategies, and benevolent, constructive teacher feedback) can very positively affect young students’ FL spelling and writing skills, because the level of students’ spelling, reading and writing skills depends on the quantity and quality of the FL input. Furthermore, if children leave elementary school and are able to read and write texts in the FL and use a good range of reading and writing 167 7.9 Conclusion on recommendations strategies, they will have a solid foundation for future literacy development. At the same time, it is very important for them to feel positive about reading and writing in the FL, to understand why literacy is useful and to enjoy tackling a text in the foreign language, confident that they will be able to get something from it. A wide range of material motivates young students to try reading in the foreign language and to enjoy the process, without having to understand every word. Finally, the world of cyberspace also relies on literacy skills (particularly in English), and it offers a wide range of possibilities for literacy skills development, for students of any age and teachers alike (see e.g., Whyte & Schmid, 2018, on recommendations for the effective integration of new technologies in the elementary EFL classroom). 168 7. Recommendations for teaching reading and writing in the FL classroom 8. Conclusion This book presented the data of four elementary school programs which offer either a regular mainstream (EFL) program or a bilingual program. In such a bilingual program English was used as the language of instruction in either only one subject (i.e., Bili-20) or in several subjects (i.e., Bili-50 and Bili-70). The focus of this study was on the development of reading and writing skills in English and German, with special reference to the influence of child-internal and child-external factors such as the children’s language skills in German and English, their social environment (particularly their parents’ socio-economic and educational background), the children’s cognitive skills (using nonverbal intelligence as a proxy) and their institutional environment regarding the school program (i.e., the intensity of the FL program). Of special interest were students with a minority language background (sometimes also called “students with a migration background”). These students have often been claimed to constitute an “at-risk” group due to the mismatch of family language and school language, and because a lower socio-economic status is often associated with a migrant background, which may result in academic underachievement in school. Partic‐ ularly regarding this group, the purpose of this book was to disentangle the effects of language background, cognitive background, and social background, and to describe the impact of these factors on the development of German and English reading and writing skills in various elementary school programs in more detail. To summarize, the focus of this book was on the following aspects, which had not been examined in previous studies: • This study compared four FL programs with different intensity (instead of comparing only two programs, e.g., regular vs. immersion programs). • This comparison also included a bilingual program with lower FL inten‐ sity (i.e., a Bili-20 program) instead of focusing only on high-intensity programs (e.g., partial vs. total immersion programs). • The students were examined for their German and English reading and writing skills (rather than for one language only), providing a more holistic view on what students in these four different programs can achieve after four years of elementary school with respect to their literacy skills in the school language and in the FL. • Writing skills (particularly in the FL) had only rarely been examined in prior studies. Most results of this study corresponded well to previous research and extended the scope thereof: • Intensity matters, i.e., the higher the FL input, the better the results on FL tests (see also e.g., Couve et al., 2016; Steinlen, 2016; Steinlen & Piske, 2018a). • In all bilingual programs, the students generally achieved age-appropriate results on the German reading and writing tests despite some/ many subjects being taught in English (see also e.g., Baumert et al., 2017; Genesee, 1987; Steinlen, 2016, 2018a, b; Swain & Lapkin, 1982; Turnbull, Lapkin & Hart, 2001; Zaunbauer & Möller, 2006, 2007, 2010). • Students’ social and cognitive background significantly differed across the FL programs. Therefore, these variables were used as covariates, particularly regarding preselectional effects for students in the bilingual programs (see also e.g., Baumert et al., 2017; Bialystok et al., 2014; Gebauer et al., 2012; Reeder et al., 1999; Swain & Lapkin, 1982; Thomas & Collier, 2002; Zaunbauer & Möller, 2007, 2010; Zaunbauer et al., 2005). • In high intensity FL programs (i.e., Bili-50 and Bili-70), fourth graders reached level upper A2/ B1 for English reading and level A2 for English writing. Level B1 is expected of students in mainstream secondary school programs at the end of grade 10 (KMK, 2013b). • In low intensity FL programs (i.e., Bili-20 and EFL), level A1 and lower level A2 were obtained for English writing and reading, respectively, at the end of grade 4. Especially for FL reading, this level surpasses level A1 which is envisioned by the KMK (2013a) for regular EFL programs at the end of the elementary school years in Germany. • The minority language students in the four programs performed as well as their majority language peers on all English tests and on the German reading test (see also e.g., Steinlen & Piske, 2013; Steinlen, 2016, 2018a, b). • The minority language students in the EFL program obtained scores below average on the German writing test (see also e.g., Chudaske, 2012; Haag et al., 2017; May, 2006) but performed age-appropriately on the German reading tests. However, some findings did not support results of previous studies: • The minority language students in the EFL program performed as well as their majority language peers on the German reading test, probably due to their social/ cognitive background (but see e.g., Bos & Pietsch, 2006; 170 8. Conclusion Chudaske, 2012; Dollmann, 2010; Haag et al., 2017; Hussmann et al., 2017 for different results). • In contrast to many other studies (e.g., Baumert et al., 2017; Böhme et al., 2016; Hussmann et al., 2017; Mullis et al., 2017), gender did not affect the results of any language test in any group in the present study, probably due to gender-sensitive teaching styles (see also Steinlen, 2016, 2018a). As the focus of this book was on minority language students, a few summa‐ rizing words on their achievements are in order. The results of this study clearly indicate that minority language children fared as well as their majority language peers on the English reading and writing test and also on the German reading test, regardless of the FL intensity of the program. Thus, a multilingual background per se does not automatically lead to poorer school performance. Inversely, any students - i.e., minority or majority language students - can actually turn into at-risk students if they receive little parental support, grow up under difficult socio-economic circumstances, or receive their school instruction in less supportive learning environments. Accordingly, any examination of children’s language development in FL programs calls for a more holistic approach that takes into account as many variables as possible (such as the children’s language skills in all their languages, their family and social background, their cognitive skills, their personality and their institutional environment). The results of this study confirmed that such a holistic approach is indeed feasible to determine possible confounding effects of learners’ linguistic, cognitive, and social background. This book also included recommendations regarding FL literacy activities. Of particular interest were awareness-raising activities, The Daily 5, explicit teach‐ ing of sound-symbol correspondence, spelling strategies, and teacher feedback to writing. These recommendations may be adapted to any target language, relating to second or foreign language learning in regular, CLIL or immersion classrooms. However, additional research is still necessary to determine the influence of these FL activities and their interactions on improvements of students’ FL spelling, reading and writing skills and to disentangle possible effects of high-quality teaching on the one hand and structural (e.g. quantitative) effects of bilingual programs on the other hand. The aim of any research is not only to provide answers but also to generate new questions. It is my hope that this book will inspire new studies examining the linguistic development of heterogeneous learner groups of any age in differ‐ ent school programs, including research-based recommendations for teaching literacy skills to different learner groups in more detail. 171 8. Conclusion 9. References Abbott, Robert; Berninger, Virginia W. & Fayol, Michel (2010). Longitudinal relationships of levels of language in writing and between writing and reading in grades 1 to 7. Journal of Educational Psychology, 102, 281-298. Aktionsrat Bildung (2009). Geschlechterdifferenzen im Bildungssystem. Jahresgutachten für 2009. Wiesbaden: Verlag für Sozialwissenschaften. Angelovska, Tanja & Hahn, Angela (2012). Written L3 (English): Transfer phenomena of L2 (German) lexical and syntactic properties. In: Gabrys-Barker, Danuta (Ed.). Cross-Linguistic Influences in Multilingual Language Acquisition, Second Language Learning and Teaching. Berlin: Springer, 23-40. Angelovska, Tanja & Hahn, Angela (2014). Raising language awareness for learning and teaching L3 grammar. In: Benati, Alessandro; Arche, Maria & Laval, Cecile (Eds.). The Grammar Dimension in Instructed Second Language Learning. London: Bloomsbury Academic, 185-207. Apsel, Carsten (2012). Coping with CLIL: Dropouts from CLIL streams in Germany. International CLIL Research Journal, 1, 47-56. Archibald, Alasdair Neil & Jeffery, Gaynor (2000). Second language acquisition and writing: A multi-disciplinary approach. Learning and Instruction, 1, 1-11. Arnett, Katy & Mady, Callie (Eds., 2013). Minority Populations in Canadian Second Language Education. Bristol: Multilingual Matters. Avvisati, Francesco (2020). The measure of socio-economic status in PISA: A review and some suggested improvement. Large-Scale Assessment in Education, 8/ 8. doi.org/ 10.1186/ s40536-020-00086-x. Babocká, Mária (2015). Assessment in CLIL classes. In: Pokrivčáková, Silvia (Ed.). CLIL in Foreign Language Education: E-Textbook for Foreign Language Teachers. Nitra: Constantine the Philosopher University, 176-188. Bae, Jungok (2007). Development of English skills need not suffer as a result of immersion: Grades 1 and 2 writing assessment in a Korean/ English two-way immersion program. Language Learning, 57, 299-332. Baker, Colin (2001). Foundations of Bilingual Education and Bilingualism. Third Edition. Clevedon: Multilingual Matters. Bardel, Camilla & Falk, Ylva (2007). The role of the second language in third language acquisition: The case of Germanic syntax. Second Language Research, 23/ 4, 459-484. Bardel, Camilla & Falk, Ylva (2012). The L2 status factor and the declarative/ procedural distinction. In: Amaro, Jennifer; Flynn, Suzanne & Rothman, Jason (Eds.). Third Language Acquisition in Adulthood. Amsterdam: John Benjamins, 61-78. Barik, Henry C. & Swain, Merril (1976). A longitudinal study of bilingual and cognitive development. International Journal of Psychology, 11/ 4, 251-263. Bassetti, Benedetta & Cook, Vivian (2011). Language and cognition: The second language user. In: Cook, Vivian & Bassetti, Benedetta (Eds.). Language and Bilingual Cognition. Oxford: Pychology Press, 143-190. Baumert, Jürgen & Schümer, Gundel (2001). Familiäre Lebensverhältnisse, Bildungsbe‐ teiligung und Kompetenzerwerb. In: Baumert, Jürgen; Klieme, Eckhard; Neubrand, Michael; Prenzel, Manfred; Schiefele, Ulrich; Schneider, Wolfgang; Stanat, Petra; Tillmann, Klaus-Jürgen & Weiß, Manfred (Eds.). PISA 2000. Basiskompetenzen von Schülerinnen und Schülern im internationalen Vergleich. Opladen: Leske + Budrich, 323-407. Baumert, Jürgen; Köller, Olaf & Lehmann, Rainer (2012). Leseverständnis im Englischen und Deutschen und Mathematikleistungen bilingual unterrichteter Schülerinnen und Schüler am Ende der Grundschulzeit. Ergebnisse eines Zwei-Wege-Immersionsprogramms. Unterrichtswissenschaft, 40/ 4, 290-315. Baumert, Jürgen; Hohenstein Friederike; Fleckenstein, Johanna; Preusler, Sandra; Pau‐ lick, Isabell & Möller, Jens (2017). Die schulischen Leistungen an der SESB - 4. Jahrgangsstufe. In: Möller, Jens; Hohenstein, Friederike; Fleckenstein, Johanna; Köller, Olaf & Baumert, Jürgen (Eds.). Erfolgreich integrieren - die Staatliche Europa-Schule Berlin. Münster: Waxmann, 95-188. Bear, Donald R. & Templeton, Shane (1998). Explorations in developmental spelling: Foundations for learning and teaching phonics, spelling and vocabulary. The Reading Teacher, 52, 222-243. Bebout, Linda (1985). An error analysis of misspellings made by learners of English as a first and as a second language. Journal of Psycholinguistic Research, 14/ 6, 569-593. Berninger, Virginia W. & Chanquoy, Lucile (2012). What writing is and how it changes across early and middle childhood development: A multidisciplinary perspective. In: Grigorenko, Elena; Mambrino, Elisa & Preiss, David (Eds.). Handbook of Writing: A Mosaic of Perspectives and Views. New York, NY: Psychology Press, 65-84. Berninger, Virginia W.; Cartwright, Ana C.; Yates, Cheryl M.; Swanson, H. Lee & Abbot, Robert D. (1996). Developmental skills related to writing and reading acquisition in the intermediate grades. Shared and unique functional systems. Reading & Writing, 6/ 2, 161-196. Berninger, Virginia W. & Swanson, H. Lee (1994). Modifying Hayes and Flower’s model of skilled writing to explain beginning and developing writing. In: Butterfield, Earl C. & Carlson, Jerry S. (Eds.). Advances in Cognition and Educational Practice: Vol. 2. Children’s Writing: Towards a Process Theory of the Development of Skilled Writing. New York: JAI Press, 57-81. 174 9. References Bérubé, Daniel & Marinova-Todd, Stefka H. (2012). The development of language and reading skills in the second and third languages of multilingual children in French Immersion. International Journal of Multilingualism, 9/ 3, 272-293. Bialystok, Ellen; Peets, Kathleen F. & Moreno, Sylvain (2014). Producing bilinguals through immersion education: Development of metalinguistic awareness. Applied Psycholinguistics, 35/ 1, 177-191. Biebricher, Christine (2012). Literacy and teaching reading in English as a foreign language. Anglistik - International Journal of English Studies, 23, 107-117. Biedinger, Nicole (2009). Der Einfluss von elterlichen Investitionen auf die Entwicklung deutscher und türkischer Kinder. Berliner Journal für Soziologie, 19/ 2, 268-294. BIG-Kreis (Eds., 2015). Der Lernstand im Englischunterricht am Ende von Klasse 4 - Ergebnisse der BIG-Studie. München: Domino Verlag. Bild, Eva R. & Swain, Merril (1989). Minority language students in a French immersion program: Their French proficiency. Journal of Multilingual and Multicultural Develop‐ ment, 10/ 3, 255-274. Björklund, Siv & Mård-Miettinen, Karita (2011). Integrating multiple languages in immersion: Swedish immersion in Finland. In: Tedick, Diane; Christian, Donna & Fortune; Tara W. (Eds.). Immersion Education. Practices, Policies, Possibilities. Bristol: Multilingual Matters, 13-35. Bleakley, Hoyt & Chin, Aimee (2004). Language skills and earnings: Evidence from childhood immigrants. Review of Economics and Statistics, 86, 481-496. Bleyhl, Werner (2000). Empfehlungen zur Verwendung des Schriftlichen im Fremd‐ sprachenerwerb der Grundschule. In: Bleyhl, Werner (Ed.). Fremdsprachen in der Grundschule. Grundlagen und Praxisbeispiele. Hannover: Schroedel, 84-91. Bleyhl, Werner (2007). Schrift im fremdsprachlichen Anfangsunterricht - ein zweisch‐ neidiges Schwert. Take Off! Zeitschrift für frühes Englischlernen, 1, 47. Blom, Elma; Boerma, Tessel; Bosma, Evelyn; Cornips, Leonie & Everaert, Emma (2017). Cognitive advantages of bilingual children in different sociolinguistic contexts. Frontiers in Psychology, 8, 552. doi: 10.3389/ fpsyg.2017.00552. Blunch, Niels (2012). Introduction to Structural Equation Modelling Using SPSS and AMOS. Second Edition. Los Angeles, CA: Sage. BMBF (Bundesministerium für Bildung und Forschung, 2007). Förderung von Lesekom‐ petenz - Expertise. Bildungsforschung Band 17. Bonn, Berlin: BMBF. Böhme, Katrin; Sebald, Susanne; Weirich, Sebastian & Stanat, Petra (2016). Geschlechts‐ bezogene Disparitäten. In: Stanat, Petra; Böhme, Katrin; Schipolowski, Stefan & Haag, Nicole (Eds.). IQB-Bildungstrend 2015. Sprachliche Kompetenzen am Ende der 9. Jahrgangsstufe im zweiten Ländervergleich. Münster: Waxmann, 377-408. 175 9. References Bonnet, Andreas; Breidbach, Stefan & Hallet, W. (2003). Fremdsprachlich handeln im Sachfach: Bilinguale Lernkontexte. In: Bach, Gerhard & Timm, Johannes-Peter (Eds.). Englischunterricht. Grundlagen und Methoden einer handlungsorientierten Unterrichtspraxis. Frankfurt am Main: Lang, 172-196. Börner, Otfried; Böttger, Heiner; Kierepka, Adelheit & Lohmann, Christa (2017). Fremd‐ sprachenunterricht in der Primarschule - Potenziale für zukünftige Standards. Fremd‐ sprachen Lehren und Lernen, 46/ 2, 85-103. Bos, Wilfred & Pietsch, Marcus (2006). KESS 4. Kompetenzen und Einstellungen von Schülerinnen und Schülern Jahrgangsstufe 4. Hamburg: Bergmann & Sohn. Bos, Wilfred; Tarelli, Irmela; Bremerich-Vos, Albert & Schwippert, Knut (Eds., 2012). IGLU 2011: Lesekompetenzen von Grundschulkindern in Deutschland im internationalen Vergleich. Münster: Waxmann. Böttger, Heiner (2014). Genderdifferenzierung im Fremdsprachenunterricht: Geschlechtsspezifische funktionale Unterschiede in der neuronalen Organisation und didaktische Konsequenzen. In: Böttger, Heiner & Gien, Gabriele (Eds.). The Multilin‐ gual Brain: Zum neurodidaktischen Umgang mit Mehrsprachigkeit. Konferenzband 2014. Eichstätt: Eichstaett Academic Press, 133-154. Böttger, Heiner (2016). Neurodidaktik des frühen Sprachenlernens. Wo die Sprache zuhause ist. Stuttgart: utb. Böttger, Heiner (2017). About boyz ‘n’ girlz: Consider the differences. Grundschule Englisch, 58, 31-33. Böttger, Heiner & Müller, Tanja (2020). Schulversuch Lernen in zwei Sprachen. Bilinguale Grundschule Englisch Bayern. Abschlussbericht der wissenschaftlichen Evaluation. https: / / bildungspakt-bayern.de/ wp-content/ uploads/ 2020/ 07/ Abschlussbericht.pdf. Boudon, Raymond (1974). Education, Opportunity, and Social Inequality: Changing Pros‐ pects in Western Society. New York: Wiley. Boushey, Gail & Moser, Joan (2014). The Daily 5: Fostering Literacy in the Elementary Grades. Portsmouth, NH: Stenhouse. Bredenbröcker, Martina; Brune, Jasmin & Elsner, Daniela (2015). Sally - Allgemeine Ausgabe. Englisch ab Klasse 3 - Pupil’s Book. München: Oldenbourg. Brehmer, Bernhard & Mehlhorn, Grit (2018). Herkunftssprachen. Tübingen: Narr. Bulon, Amélie; Hendrik, Isa; Meunier, Fanny & Van Goethem, Kristel (2017). Using global complexity measures to assess second language proficiency: Comparing CLIL and non-CLIL learners of English and Dutch in French-speaking Belgium.Travaux du CBL, 11/ 1, 1-25. Bundesverband Legasthenie (BVL, 2018). Ratgeber zum Thema Legasthenie - Erkennen und Verstehen. Informationsbroschüre. Twelfth Edition. https: / / www.bvl-legasthenie.de. 176 9. References Burmeister, Petra (2006). Immersion und Sprachunterricht im Vergleich. In: Pienemann, Manfred: Keßler, Jörg- U. & Roos, Eckhard (Eds.). Englischerwerb in der Grundschule. Ein Studien- und Arbeitsbuch. Paderborn: Schöningh, 197-216. Burmeister, Petra (2010). *Did you now that 15 difrent Fish arts in the Kiel Canal live? On foreign language writing in partial immersion primary school classrooms. In: Diehr, Bärbel & Rymarczyk, Jutta (Eds.). Researching Literacy in a Foreign Language among Primary School Learners. Forschung zum Schrifterwerb in der Fremdsprache bei Grundschülern. Frankfurt am Main: Peter Lang, 131-145. Burmeister, Petra (2013). Immersion. In: Hallet, Wolfgang & Königs, Frank (Eds.). Handbuch Bilingualer Unterricht/ Content and Language Integrated Learning. Seelze: Klett Kallmeyer, 160-167. Burmeister, Petra & Daniel, Angelika (2002). How effective is late partial immersion? Some findings from a secondary school program in Germany. In: Burmeister, Petra; Piske, Thorsten & Rohde, Andreas (Eds.). An Integrated View of Language Development. Papers in Honor of Henning Wode. Trier: Wissenschaftlicher Verlag Trier, 499‐515. Burmeister, Petra & Massler, Ute (2010). Introduction. In: Burmeister, Petra & Massler, Ute (Eds.). CLIL und Immersion: Fremdsprachlicher Sachfachunterricht in der Grundschule. Braunschweig: Westermann, 7-9. Burmeister, Petra; Muller, Jennifer & Nistl, Lea (2018). The Daily 5: Lese-und Schreib‐ kompetenzen fördern. Grundschulmagazin Englisch, 2/ 2018, 31-34. Burmeister, Petra & Pasternak, Ruth (2004). Früh und intensiv: Englische Immersion in der Grundschule am Beispiel der Claus-Rixen-Grundschule in Altenholz. fmf-Landes‐ verband Schleswig-Holstein. Mitteilungsblatt August 2004. www.fmks-online.de. Burmeister, Petra & Piske, Thorsten (2008). Schriftlichkeit im fremdsprachlichen Sa‐ chunterricht an der Grundschule. In: Böttger, Heiner (Ed.). Fortschritte im frühen Fremdsprachenlernen. Ausgewählte Tagungsbeiträge Nürnberg 2007. München: Dom‐ ino, 183-193. Burmeister, Petra & Piske, Thorsten (2011). Der Umgang mit Rechtschreibfehlern im Englischunterricht der Grundschule. Take Off! Zeitschrift für frühes Englischlernen, 2/ 2011, 48-49. Burwitz-Melzer, Eva (2010). Lesen und Schreiben im Englischunterricht der Primarschule und des Übergangs (Klasse 3 bis 6). In: Appel, Joachim; Doff, Sabine; Rymarczyk, Jutta & Thaler, Engelbert (Eds.). Foreign Language Teaching: History, Theory, Methods. Berlin: Langenscheidt, 101-122. Burstall, Clare (1975). Factors affecting foreign-language learning: a consideration of some relevant research findings. Language Teaching and Linguistics Abstracts, 8, 105-125. Cameron, Lynne (2001). Teaching Languages to Young Learners. Cambridge: Cambridge University Press. 177 9. References Canga-Alonso, Andrés (2015). Receptive vocabulary of CLIL and non-CLIL primary and secondary school learners. Complutense Journal of English Studies, 23, 59-77. Carreira, Maria; Jensen, Linda & Kagan, Olga (2009). The Heritage Language Learner Survey: Report on the Preliminary Results. www.international.ucla.edu/ media/ files/ pa per.pdf. Carrell, Patricia L. & Grabe, William (2002). Reading. In: Schmitt, Norbert (Ed.). An Introduction to Applied Linguistics. London: Arnold, 233-250. Caruana, Edward J.; Roman, Marius; Hernández-Sánchez, Jules & Solli, Piergiorgio (2015). Longitudinal studies. Journal of Thoracic Disease, 7/ 11, E537-40. Cashion, Marie & Eagan, Ruth (1990). Spontaneous reading and writing in English by students in total French immersion: Summary of final report. English Quarterly, 22/ 1-2, 30-44. Cenoz, Jasone; Genesee, Fred & Gorter, Durk (2014). Critical analysis of CLIL: Taking stock and looking forward. Applied Linguistics, 35/ 3, 243-262. Cenoz, Jasone & Gorter, Durk (2011). Focus on multilingualism: A study of trilingual writing. The Modern Language Journal, 95/ 3, Special Issue: Toward a Multilingual Approach in the Study of Multilingualism in School Contexts, 356-369. Cenoz, Jasone (2000). Research on multilingual acquisition. In: Cenoz, Jasone & Jessner, Ulrike (Eds.). English in Europe. Clevedon: Multilingual Matters, 39-53. Cenoz, Jasone (2003). The role of typology in the organization of the multilingual lexicon. In: Cenoz, Jasone; Hufeisen, Britta & Jessner, Ulrike (Eds.). The Multilingual Lexicon. Dordrecht: Kluwer, 103-116. Cenoz, Jasone (2013). The influence of bilingualism on third language acquisition: Focus on multilingualism. Language Teaching, 46/ 1, 71-86. Cheng, Liying; Li, M. Miao; Kirby, John R.; Quiang, Haiyan & Wade-Woolley, Lesley (2010). English language immersion and students’ academic achievement in English, Chinese and mathematics. Evaluation & Research in Education 23/ 3, 151-169. Chudaske, Jana (2012). Sprache, Migration und schulfachliche Leistung. Einfluss spra‐ chlicher Kompetenz auf Lese-, Rechtschreib- und Mathematikleistungen. Wiesbaden: Verlag für Sozialwissenschaften. CLIL4Children (2018). Teacher’s Guide on CLIL Methodology in Primary Schools. Volume 1. www.clil4children.eu/ wp-content/ uploads/ 2018/ 06/ Guide_Addressed_to_ Teachers_Vol01.pdf. Collier, Virginia & Thomas, Wayne (2017). Validating the power of bilingual schooling: Thirty-two years of large-scale, longitudinal research. Annual Review of Applied Lin‐ guistics, 37, 203-217. Cook, Vivian (1999). Teaching L2 Spelling. Unpublished Paper. www.viviancook.uk/ Writ ings/ Papers/ TeachingSpelling.htm. 178 9. References Cook, Vivian & Bassetti, Benedetta (2005). Second Language Writing Systems. Clevedon: Multilingual Matters. Council of Europe (2001). Common European Framework of Reference for Languages: Learning, Teaching and Assessment. Strasbourg: Language Policy Unit. Council of Europe (2018). Common European Framework of Reference for Languages: Learning, Teaching and Assessment. Companion with New Descriptors. Strasbourg: Language Policy Unit https: / / rm.coe.int/ cefr-companion-volume-with-new-descripto rs-2018/ 1680787989. Couve de Murville, Stephanie; Kersten, Kristin; Maier, Esther; Ponto, Katharina & Weitz, Martina (2016). Rezeptiver L2 Wortschatz in der Grundschule. In: Steinlen, Anja K. & Piske, Thorsten (Eds.). Wortschatzlernen in bilingualen Schulen und Kindertagesstätten. Frankfurt am Main: Peter Lang, 85-121. Coyle, Do; Hood, Philip & Marsh, David (2010). CLIL: Content and Language Integrated Learning. Cambridge: Cambridge University Press. Cummins, James (1979). Linguistic interdependence and the educational development of bilingual children. Review of Educational Research, 49/ 2, 222-251. Cummins, Jim (1991). Interdependence of firstand second-language proficiency in bilingual children. In: Bialystok, Ellen (Eds.). Language Processing in Bilingual Children. Cambridge, New York: Cambridge University Press, 70-89. Dale, Philip S. (1976). Language Development: Structure and Function. New York: Holt, Rinehart and Winston. Dallinger, Sara (2015). Die Wirksamkeit bilingualen Sachfachunterrichts: Selektionseffekte, Leistungsentwicklung und die Rolle der Sprachen im deutsch-englischen Geschichtsun‐ terricht. Unveröffentlichte Dissertation. Ludwigsburg, Germany: Pädagogische Hoch‐ schule Ludwigsburg. De Angelis, Gessica (2005). The acquisition of languages beyond the L2: Psycholinguistic perspectives. Rassegna Italiana di Linguistica Applicata, 2-3, 397-409. De Angelis, Gessica (2007). Third or Additional Language Acquisition. Clevedon: Multi‐ lingual Matters. De Bree, Elise & Unsworth, Sharon (2014). Dutch and English literacy and language outcomes of dyslexic students in regular and bilingual secondary education. Dutch Journal of Applied Linguistics, 3/ 1, 62-81. De La Paz, Susan & McCutchen. Susan (2016). Learning to write. In: Mayer, Richard E. & Alexander, Patricia A. (Eds.). Handbook of Research on Learning and Instruction. New York: Routledge, 32-56. Decke-Cornill, Helene (2007). The issue of gender and interaction in the L2 classroom. In: Decke-Cornill, Helene & Volkmann, Laurenz (Eds.). Gender Studies and Foreign Language Teaching. Tübingen: Narr, 77-90. 179 9. References Demmig, Silvia (2018). CLIL, sprachsensibler Fachunterricht, sprachbewusster Unter‐ richt, Bildungssprache, DaZ, Language Awareness? Welche Konzepte brauchen Lehrkräfte in der Primarstufe in Österreich? Zeitschrift für Interkulturellen Fremd‐ sprachenunterricht, 23/ 1, 111-120 Department for Education and Skills, UK (2007). Letters and Sounds: Principles and Practice of High Quality Phonics. Primary National Strat‐ egy. https: / / assets.publishing.service.gov.uk/ government/ uploads/ system/ uploads/ at tachment_data/ file/ 190599/ Letters_and_Sounds_-_DFES-00281-2007.pdf. Department of Education, Newfoundland and Labrador Language Programs (2010). Developing Spelling Skills in the French Immersion Classroom - Primary and Ele‐ mentary: A Support Document for Teachers. www.ed.gov.nl.ca/ edu/ k12/ french/ immer sion/ autres_documents/ developing_spelling_skills_in_the_french_immersion_classr oom_2010.pdf. DESI-Konsortium (Eds., 2008). Unterricht und Kompetenzerwerb in Deutsch und Englisch: Ergebnisse der DESI-Studie (Deutsch-Englisch Schülerleistungen International). Beltz: Weinheim. Deutscher Bundestag (2017). Türkischer Konsulatsunterricht. www.bundestag.de/ blob/ 53 6698/ 414a949206f558633c8d8909d6b10abd/ wd-8-038-17-pdf-data.pdf. Deventer, Jennifer; Machts, Nils; Gebauer, Sandra Kristina & Möller, Jens (under review). Immersion education and school achievement: A three-level meta-analysis. Kiel: CAU, Manuscript. Diehr, Bärbel & Rymarczyk, Jutta (Eds., 2010). Researching Literacy in a Foreign Language among Primary School Learners. Forschung zum Schrifterwerb in der Fremdsprache bei Grundschülern. Frankfurt am Main: Peter Lang. Ditton, Hartmut; Krüsken, Jan & Schauenberg, Magdalena (2005). Bildungsungleichheit - der Beitrag von Familie und Schule. Zeitschrift für Erziehungswissenschaft, 8/ 2, 285-304. DJI (Deutsches Jugendinstitut, 2020). DJI-Kinder- und Jugendmigrationsreport 2020. Datenanalyse zur Situation junger Menschen in Deutschland. Bielefeld: wbv Media. Dollmann, Jörg (2010). Türkischstämmige Kinder am ersten Bildungsübergang. Wiesbaden: Verlag für Sozialwissenschaften. Doyé, Peter (2008). Sprechen. Zuhören. Schreiben? Lesen? Gehört die Schrift in den Fremdsprachenunterricht der Grundschule? Grundschule, 40/ 3, 53. Drucks, Stefan (2011). Die Hamburger Schreibprobe als Instrument innerhalb der Reprä‐ sentativbefragung des Forschungsprojekts HaBil. Bericht der Fakultät für Bildungswis‐ senschaften. Universität Duisburg-Essen. www.uni-due.de/ imperia/ md/ content/ biwi/ bauer/ hsp_bericht_homepage_290711.pdf20.7.2009_.pdf. Duarte, Joana (2011). Migrants’ educational success through innovation: The case of the Hamburg bilingual schools. International Review of Education, 57, 631-649. Dubowy, Minja; Ebert, Susanne; von Maurice, Jutta & Weinert, Sabine (2008). Sprach- 180 9. References lich-kognitive Kompetenzen beim Eintritt in den Kindergarten: Ein Vergleich von Kindern mit und ohne Migrationshintergrund. Zeitschrift für Entwicklungspsychologie und Pädagogische Psychologie, 40/ 1, 124-134. Duursma, Elisabeth; Romero-Contreras, Silvia; Szuber, Anna; Proctor, Patrick & Snow, Catherine E. (2007). The role of home literacy and language environment on bilinguals’ English and Spanish vocabulary development. Applied Psycholinguistics, 28/ 1, 171-190. Dyson, Ann & Freedman, Sarah (1991). Handbook of Research on Teaching in the Language Arts. Second Edition. New York: MacMillan. Echevarrìa, Jana; Vogt, Mary Ellen & Short, Deborah (2010). Making Content Comprehen‐ sible for Elementary English Learners. The SIOP Model. Second Edition. Boston, MA: Pearson. Eckert, Hartwig & Barry, William (2002). The Phonetics and Phonology of English Pronun‐ ciation. Trier: Wissenschaftlicher Verlag Trier. Edele, Aileen; Kempert, Sebastian & Schotte, Kristin (2018). Does competent bilingualism entail advantages for the third language learning of immigrant students? Learning and Instruction, 58, 232-244. Edelenbos, Peter & Vinjé, Maria P. (2000). The assessment of a foreign language at the end of primary (elementary) education. Language Testing, 17, 144-162. Ehri, Linnea (1995). Phases of development in learning to read words. Journal of Research in Reading, 18/ 2, 116-125. Ehri, Linnea (2003). Systematic phonics instruction: Findings of the National Reading Panel. Paper presented at the Invitational Seminar organized by the Standards and Effectiveness Unit, Department for Education and Skills, British Government (London, England, March 17, 2003). www.standards.dfes.gov.uk/ pdf/ literacy/ lehri_phonics.pdf. Ellis, Rod (2008). The Study of Second Language Acquisition. Second Edition. Oxford: Oxford University Press. Elsner, Daniela & Keßler, Jörg Uwe (Eds., 2013). Bilingual Education in Primary School. Aspects of Immersion, CLIL, and Bilingual Modules. Tübingen: Narr. Elsner, Daniela & Lohe, Viviane (2016). Introduction to teaching gender in the EFL classroom. In: Elsner, Daniela & Lohe, Viviane (Eds.). Gender and Language Learning. Research and Practice. Tübingen: Narr, 9-18. Elsner, Daniela (2007). Hörverstehen im Englischunterricht der Grundschule. Ein Leistungs‐ vergleich zwischen Kindern mit Deutsch als Muttersprache und Deutsch als Zweitsprache. Frankfurt am Main: Peter Lang. Engel, Gaby; Groot-Wilken, Bernd & Thürmann, Eike (Eds., 2009). Englisch in der Primarstufe - Chancen und Herausforderungen. Evaluation und Erfahrungen aus der Praxis. Berlin: Cornelsen. Eurydice (2006). Content and Language Integrated Learning at School in Europe. Eurydice European Unit. www.indire.it/ lucabas/ lkmw_file/ eurydice/ CLIL_EN.pdf. 181 9. References Eurydice (w/ o year). National Education Systems: Germany: Teaching and Learning in Primary Education. https: / / eacea.ec.europa.eu/ national-policies/ eurydice/ content/ tea ching-and-learning-primary-education-18_en. Fashola, Olatokunbo S.; Drum, Priscilla A.; Mayer, Richard E. & Kang, Sang-Jin (1996). A cognitive theory of orthographic transitioning: Predictable errors in how Spanish-speaking children spell English words. American Educational Research Journal, 33, 825-843. Field, Andy (2013). Discovering Statistics Using IBM SPSS Statistics. Fourth Edition. Los Angeles, CA: Sage. Figueredo, Lauren (2006). Using the known to chart the unknown: A review of first-lan‐ guage influence on the development of English-as-a-second-language spelling skill. Reading and Writing, 10, 873-905. Fischer, Uta (2019). Stolpersteine beim Aufbau eines bilingualen Bildungsganges. In: Rohde, Andreas & Steinlen, Anja (Eds.). Sprachenvielfalt als Ressource begreifen. Mehr‐ sprachigkeit in bilingualen Kindertagesstätten und Schulen. Band II. Berlin: Dohrmann, 148-154. Fleckenstein, Johanna; Möller, Jens; Hohenstein, Friederike; Radmann, Susann; Becker, Michael & Baumert, Jürgen (2017). Die schulischen Leistungen an der SESB - 9. Jahrgangsstufe und 15-Jährige. In: Möller, Jens; Hohenstein, Friederike; Fleckenstein, Johanna; Köller, Olaf & Baumert, Jürgen (Eds.). Erfolgreich integrieren - die Staatliche Europa-Schule Berlin. Münster: Waxmann, 89-153. Fleckenstein, Johanna; Möller, Jens & Baumert, Jürgen (2020). Ist jünger immer besser? Frühes Fremdsprachenlernen in der Grundschule. Zeitschrift für Pädagogische Psycho‐ logie, 2020, 1-16. Flege, James E. (2009). Give input a chance! In: Piske, Thorsten & Young-Scholten, Martha (Eds.). Input Matters in SLA. Bristol: Multilingual Matters, 175-190. Flynn, Suzanne, Foley, Claire & Vinnitskaya, Inna (2004). The cumulative-enhancement model for language acquisition: Comparing adult’ and children’s patterns of develop‐ ment in first, second and third language acquisition of relative clauses. International Journal of Multilingualism, 1/ 1, 3-16. FMKS, Verein für frühe Mehrsprachigkeit in Kindertagesstätten und Schulen (2014). Ranking: bilinguale Kitas und Grundschulen im Bundesvergleich. www.fmks.eu. Fortune, Tara. (2004). Scaffolding Techniques in Content-Based Instructed Classrooms. www.carla.umn.edu/ cobalt/ modules/ strategies/ ust.html. Frisch, Stefanie (2010). Bewusstmachende Verfahren beim Umgang mit dem Schriftbild im Englischunterricht der Grundschule: erste Ergebnisse der LiPs Studie. In: Diehr, Bärbel & Rymarczyk, Jutta (Eds.). Researching Literacy in a Foreign Language among Primary School Learners. Forschung zum Schrifterwerb in der Fremdsprache bei Grund‐ schülern. Frankfurt am Main: Peter Lang, 107-130. 182 9. References Frisch, Stefanie (2013). Lesen im Englischunterricht der Grundschule. Eine Vergleichsstudie zur Wirksamkeit zweier Lehrverfahren. Tübingen: Narr. Frith, Uta (1985). Beneath the surface of developmental dyslexia. In: Patterson, Karalyn E; Marshall, John C. & Coltheart, Max (Eds.). Surface Dyslexia: Neuropsychological and Cognitive Studies of Phonological Reading. Hillsdale, NJ: Laurence Erlbaum, 301-330. Frost, Ram (2005). Orthographic systems and skilled word recognition in reading. In: Snowling, Margaret & Hulme, Charles (Eds.). The Science of Reading: A Handbook. Oxford: Blackwell, 272-295. Fuchs, Stefanie (2015). Geschlechtsunterschiede bei motivationalen Faktoren im Kontext des Englischunterrichts der Sekundarstufe I. Zeitschrift für Fremdsprachenforschung, 25, 175-205. Gamoran, Adam (2008). Creaming. In: Darity, William A. (Ed.). International Encyclopedia of the Social Sciences: Vol. 2. Cohabitation - Ethics in Experimentation. Second Edition. Chicago, IL: Gale, 160-171. www.encyclopedia.com/ doc/ 1G2-3045300477.html. Gamsjäger, Erich & Sauer, Joachim (1996). Determinanten der Grundschulleistung und ihr prognostischer Wert für den Sekundarschulerfolg. Psychologie in Erziehung und Unterricht, 43, 182-204. Ganzeboom, Harry B. G.; de Graaf, Paul M. & Treiman, Donald J. (1992). A Standard International Socio-Economic Index of Occupational Status. Social Science Research, 21/ 1, 1-56. Gardner, Robert C., & Lambert, Wallace E. (1972). Attitudes and Motivation in Second Language Learning. Rowley, MA: Newbury House Publishers. Gardner, Robert C.; Masgoret, Anne-Marie & Tremblay, Paul F. (1999). Home background characteristics and second language learning. Journal of Language and Social Psychol‐ ogy, 18/ 4, 419-437. Gebauer, Sandra Kristina; Zaunbauer, Anna C.M. & Möller, Jens (2012). Erstsprachliche Leistungsentwicklung im Immersionsunterricht: Vorteile trotz Unterrichts in einer Fremdsprache? Zeitschrift für Pädagogische Psychologie, 3, 183-196. Gebauer, Sandra Kristina; Zaunbauer, Anna C.M. & Möller, Jens (2013). Cross-language transfer in English immersion programmes in Germany: Reading comprehension and reading fluency. Contemporary Educational Psychology, 38, 64-74. Genesee, Fred H. (1976). The role of intelligence in second language learning. Language Learning, 26, 267-280. Genesee, Fred H. (1978). A longitudinal evaluation of an early immersion program. Canadian Journal of Education, 3, 31-50. Genesee, Fred H. (1987). Learning Through Two Languages: Studies of Immersion and Bilingual Education. Cambridge. MA: Newbury House. 183 9. References Genesee, Fred H. (2004). What do we know about bilingual education for majority language students? In: Bhatia, Tay K. & Ritchie, William C. (Eds.). Handbook of Bilingualism and Multiculturalism. Malden, MA: Blackwell, 547-561. Genesee, Fred H. (2007). French Immersion and at-risk students: A review of research evidence. Canadian Modern Language Review, 63/ 5, 655-688. Genesee, Fred H. & Fortune, Tara W. (2014). Bilingual education and at-risk students. Journal of Immersion and Content-Based Language Education, 2/ 2, 196-209. Genesee, Fred H. & Jared, Debra (2008). Literacy development in early French immersion programs. Canadian Psychology, 49/ 2, 140-147. Genesee, Fred H., Holobow, Naomi, Lambert, Wallace E. & Chartrand, Louise (1989). Three elementary school alternatives for learning through a second language. Modern Language Journal, 73, 250-263. Gentry, J. Richard (2000). A retrospective on invented spelling and a look forward. The Reading Teacher, 54/ 3, 318-332. Gentry, J. Richard (2006). Breaking the Code: The New Science of Beginning Reading and Writing. Portsmouth, NH: Heinemann. Gerlach, David (2019). Lese-Rechtschreib-Schwierigkeiten (LRS) im Fremdsprachenunter‐ richt. Tübingen: Narr Francke Attempto. Geva, Esther & Clifton, Susan (1994). The development of first and second language reading skills in early French immersion. The Canadian Modern Language Review, 50, 646-667. Gondová, Danica (2015). Selecting, adapting and creating CLIL materials. In: Pokrivčá‐ ková, Silvia (Ed.). CLIL in Foreign Language Education: E-Textbook for Foreign Language Teachers. Nitra: Constantine the Philosopher University, 153-163. www.klis.pf.ukf.sk / dokumenty/ CLIL/ e-textbook%20CLIL%20in%20FLE%20-%20final.pdf. Gonzalez, Virginia (1999). Language and Cognitive Development in Second Language Learning. Boston, MA: Allyn and Bacon. Gottburgsen, Anja & Gross, Christiane (2012). Welchen Beitrag leistet Intersektionalität zur Klärung von Kompetenzunterschieden bei Jugendlichen? In: Becker, Rolf & Solga, Heike (Eds.). Soziologische Bildungsforschung. Wiesbaden: Springer, 86-110. Grabe, William & Kaplan, Robert B. (1996). Theory and Practice of Writing. An Applied Linguistic Perspective. Harlow: Pearson. Grabe, William (2005). The role of grammar in reading comprehension. In: Frodesen, Jan & Holten, Christine (Eds.). The Power of Context in Language Teaching and Learning. Boston: Heinle & Heinle, 268-282. Grabe, Willliam (2009). Reading in a Second Language: Moving from Theory to Practice. Cambridge: Cambridge University Press. Grabe, Willliam (2010). Fluency in reading-thirty-five years later. Reading in a Foreign Language, 22/ 1, 71-83. 184 9. References Grabe, Willliam (2014). Key issues in L2 reading development. In: Deng, Yudong & Seow, Richard (Eds.). Alternative Pedagogies in the English Language and Communication Classroom: Fourth CELC Symposium Proceeding. Singapore: National University of Singapore, Center for English Language Communication, 8-18. www.nus.edu.sg/ celc / research/ books/ 4th%20Symposium%20proceedings/ 2).%20William%20Grabe.pdf. Graham, Steve; Morphy, Paul; Harris, Karen; Fink-Chorzempa, Barbara; Saddler, Bruce; Moran, Susan & Mason, Linda. (2008). Teaching spelling in the primary grades: A national survey of instructional practices and adaptations. American Educational Research Journal, 45, 796-825. doi: 10.3102/ 0002831208319722. Greenberg, Karin L. (1992). Validity and reliability issues in the direct assessment of writing. Writing Program Administration, 16/ 1-2, 7-22. Grimm, Hannelore (2001). SETK 3-5: Sprachentwicklungstest für dreibis fünfjährige Kinder: Diagnose von Sprachverarbeitungsfähigkeiten und auditiven Gedächtnisleistun‐ gen. Göttingen: Hogrefe. Griva, Eleni & Chostelidou, Dora (2013). Writing skills and strategies of bilingual immigrant students learning Greek as a second language and English as a foreign language. Reading & Writing, 4/ 1, Art. #31. http: / / dx.doi.org/ 10.4102/ rw.v4i1.31. Gross, Christiane & Gottburgsen, Anja (2013). Gender, soziale Herkunft und Migration: "Intersektionalität" im Erwerb von Mathematikkompetenzen. In: Hadjar, Andreas & Hupka-Brunner, Sandra (Eds.). Geschlecht, Migrationshintergrund und Bildungserfolg. Weinheim: Beltz, 188-212. Haag, Nicole; Kocaj, Alexander; Jansen, Malte & Kuhl, Poldi (2017). Soziale Disparitäten. In: Stanat, Petra; Schipolowski, Stefan; Rjosk, Camilla; Weirich, Sebastian & Haag, Nicole (Eds.). IQB-Bildungstrend 2016. Kompetenzen in den Fächern Deutsch und Mathematik am Ende der 4. Jahrgangsstufe im zweiten Ländervergleich. Münster: Waxmann, 213-235. Haag, Nicole; Böhme, Katrin; Rjosk; Camilla & Stanat, Petra (2016). Zuwanderungsbezo‐ gene Disparitäten. In: Stanat, Petra; Böhme, Katrin; Schipolowski, Stefan; Haag, Nicole (Eds.). IQB-Bildungstrend 2015. Sprachliche Kompetenzen am Ende der 9. Jahrgangsstufe im zweiten Ländervergleich. Münster: Waxmann, 431-479. Haenni Hoti, Andrea (2008). Der Einfluss des Migrationshintergrunds auf die Englisch‐ fertigkeiten von Primarschülerinnen und -schülern. In: Ramseger, Jörg & Wagener, Matthea (Eds.). Chancenungleichheit in der Grundschule: Ursachen und Wege aus der Krise, Wiesbaden: VS Verlag für Sozialwissenschaften, 125-128. Haenni Hoti, Andrea; Heinzmann, Sybille; Müller, Marianne; Oliveira, Marta; Wicki, Werner & Werlen, Erika (2010). Introducing a second foreign language in Swiss primary schools: The effect of L2 listening and reading skills on L3 acquisition. International Journal of Multilingualism, 8/ 2, 98-116. 185 9. References Hakuta, Kenji & Diaz, Rafael M. (1985). The relationship between degree of bilingualism and cognitive ability: A critical discussion and some new longitudinal data. In: Nelson, Keith E. (Ed.). Children’s Language. Volume 5. Hillsdale, NJ: Lawrence, 319-344. Hammarberg, Björn & Williams, Sarah (2009). A study of third language acquisition. In: Hammarberg, Björn (Ed.). Processes in Third Language Acquisition. Edinburgh. Edinburgh University Press, 17-27. Hammarberg, Björn (2018). L3, the tertiary language. In: Bonnet, Andreas & Siemund, Peter (Eds.). Foreign Language Education in Multilingual Classrooms. Benjamins: Amsterdam, 127-150. Hancock, Mark & McDonald, Annie (2000). Teaching writing to school children. IATEFL Young Learners sig Journal CATS, Spring 2000. http: / / hancockmcdonald.com/ ideas/ te aching-writing-school-children. Hansen, Morris H.; Hurwitz, William N.; Marks, Eli S. & Mauldin, Parker, W. (1951). Response errors in surveys. Journal of the American Statistical Association, 46/ 254, 147-190. Harley, Birgit; Hart, Doug & Lapkin, Sharon (1986). The effects of early bilingual schooling on first language skills. Applied Psycholinguistics, 7, 295-322. Hart, Doug; Lapkin, Sharon & Swain, Merrill (1988). Early and Middle Immersion Programmes: Linguistic Outcomes and Social Character. Toronto: The Metropolitan Toronto School Board. Hartig, Johannes & Jude, Nina (2008). Sprachkompetenzen von Mädchen und Jungen. In: DESI-Konsortium (Eds.). Unterricht und Kompetenzerwerb in Deutsch und Englisch. Ergebnisse der DESI Studie. Weinheim: Beltz, 202-206. Hayes, John R. & Flower, Linda S. (1980). Identifying the organisation of writing processes. In: Gregg, Lee W. & Steinberg, Erwin R. (Eds.). Cognitive Processes in Writing. Hillsdale, NJ: Erlbaum, 3-30. Hayes, John R. (1996). A new framework for understanding cognition and affect in writing. In: Levy, C. Michael & Ransdell, Sarah (Eds.). The Science of Writing. Mahwah, NJ: Erlbaum, 1-27. Hayes, John R. (2006). New directions in writing theory. In: MacArthur, Charles; Graham, Steve & Fitzgerald, Jill (Eds.). Handbook of Writing Research. New York: Guilford Press, 28-40. Heine, Lena (2004). Mögliches und Unmögliches: Zur Erforschung von Transfererschei‐ nungen. In: Hufeisen, Britta & Marx, Nicole (Eds.). ‚Beim Schwedischlernen sind Eng‐ lisch und Deutsch ganz hilfsvoll‘. Forschungen zum multiplen Spracherwerb. Tübingen: Narr, 81-96. Heller, Kurt & Geisler, Hans-Jürgen (1983). Kognitiver Fähigkeitstest (Grundschulform). KFT 1-3. Weinheim: Beltz. 186 9. References Heller, Kurt A. & Perleth, Christoph (2000). Kognitiver Fähigkeitstest für 4. bis 12. Klassen. Revision (KFT 4-12+ R). Göttingen: Beltz. Heller, Kurt H.; Kratzmeier, Heinrich & Lengfelder, Angelika (1998). Matrizen-Test-Man‐ ual Band 1. Ein Handbuch mit deutschen Normen zu den Standard Progressive Matrices von J.C. Raven. Göttingen: Beltz. Hempel, Margit; Kötter, Markus & Rymarczyk, Jutta (2018). Fremdsprachenunterricht in der Grundschule in den Bundesländern Deutschlands: Eine Bestandsaufnahme des Status quo und seiner gewünschten Weiterentwicklung. Frankfurt am Main: Peter Lang. Hempel, Margit (2016). Förderung produktiver Sprachkompetenzen - Sind Lehrwerke Teil des Problems oder Teil der Lösung. In: Böttger, Heiner & Schlüter, Norbert (Eds.). Fortschritte im frühen Fremdspachenunterricht. Braunschweig: Westermann, 124-133. Hermas, Abdelkader (2010). Language acquisition as computational resetting: Verb movement in L3 initial state. International Journal of Multilingualism, 7/ 4, 343-362. Hermas, Abdelkader (2014). Restrictive relatives in L3 English: L1 transfer and ultimate attainment convergence. Australian Journal of Linguistics, 34/ 3, 363-387. Hesse, Hermann-Günter; Göbel, Kerstin & Hartig, Johannes (2008). Sprachliche Kompe‐ tenzen von mehrsprachigen Jugendlichen und Jugendlichen nicht-deutscher Erstsprache. In: DESI-Konsortium (Eds.). Unterricht und Kompetenzerwerb in Deutsch und Englisch. Weinheim: Beltz, 208-230. Holobow, Naomi E.; Genesee, Fred & Lambert, Wallace E. (1991). The effectiveness of a foreign language immersion program for children from different ethnic and social class backgrounds: Report 2. Applied Psycholinguistics, 12, 179-198. Homstad, Torild & Thorson, Helga (1996). Using Writing-to-Learn Activities in the Foreign Language Classroom. A Research Grant Report Submitted to the Center for Interdisciplinary Studies of Writing. Technical Report Series No. 14. Minneapolis, MN: University of Minnesota. Hoover, Wesley A. & Gough, Philip B. (1990). The simple view of reading. Reading and Writing: An Interdisciplinary Journal, 2, 127-160. Hopp, Holger (2018). Cross-linguistic influence in the child third language acquisition of grammar: Sentence comprehension and production among Turkish-German and German learners of English. International Journal of Bilingualism, 2018, 1-17. DOI: 10.1177/ 1367006917752523. Hopp, Holger; Kieseier, Teresa; Vogelbacher, Markus & Thoma, Dieter (2018). Einflüsse und Potenziale der Mehrsprachigkeit im Englischerwerb in der Primarstufe. In: Mehl‐ horn, Grit & Brehmer, Bernhard (Eds.). Potenziale von Herkunftssprachen: Sprachliche und außersprachliche Einflussfaktoren. Tübingen: Stauffenburg, 57-80. Hopp, Holger; Kieseier, Teresa; Vogelbacher, Markus; Köser, Sarah & Thoma, Dieter (2017). Mehrsprachigkeit und metalinguistische Bewusstheit im Englischerwerb in der Grundschule. In: Fuchs, Isabell; Jeuk, Stefan & Knapp, Werner (Eds.). Mehrsprachigkeit: 187 9. References Spracherwerb, Unterrichtsprozesse, Schulentwicklung. Stuttgart: Fillibach bei Klett, 55-74. Horst, Sabine (o. J.). Kreatives Schreiben als Chance geschlechtersensibler Schreiberzie‐ hung. https: / / gendersensibel-unterrichten.alp.dillingen.de/ index.php/ unterricht/ prax isbeispiele/ deutsch/ 88-kreatives-schreiben. Howard, Elizabeth R.; Páez, Mariela M.; August, Diane L.; Barr, Christopher D.; Kenyon, Dorry & Malabonga, Valerie (2014). The importance of SES, home and school language and literacy practices, and oral vocabulary in bilingual children’s English reading development. Bilingual Research Journal, 37/ 2, 120-141. Hufeisen, Britta (2018). Models of multilingual competence. In: Bonnett, Andreas & Siemund, Peter (Eds.). Foreign Language Education in Multilingual Classrooms. Am‐ sterdam: Benjamins, 173-190. Hussein, Ali A. & Mohammad, F. Mohammad (2011). Negative L1 impact on L2 writing. International Journal of Humanities and Social Science, 18, 184-195. Hussmann, Anke; Stubbe, Tobias C. & Kasper, Daniel (2017). Soziale Herkunft und Lese‐ kompetenzen von Schülerinnen und Schülern. In: Hussmann, Anke; Wendt, Heike; Bos, Wilfried; Bremerich-Vos, Albert; Kasper, Daniel; Lankes, Eva-Maria; McElvany, Nele; Stubbe, Tobias C. & Valtin, Renate (Eds.). IGLU 2016: Lesekompetenzen von Grundschulkindern in Deutschland im internationalen Vergleich. Münster: Waxmann, 195-218. Hüther, Gudrun (2013). Konzept Frühförderung im Fremdsprachenunterricht: Bloß ein Mythos? Begegnung, 02/ 2013, 8-14. Ioannou-Georgiou, Sophie & Pavlou, Pavlos Y. (2003). Assessing Young Learners. Oxford: Oxford University Press. Ioannou-Georgiou, Sophie (2011). Transition into CLIL: Guidelines for the beginning stages of CLIL. In: Ioannou-Georgiou, Sophie & Pavlou, Pavlos Y. (Eds.). Guidelines for CLIL Implementation in Primary and Pre-Primary Education. Nicosia: Chipre, 34-54. Jaekel, Nils; Schurig, Michael; Florian, Merle & Ritter, Markus (2017). From early starters to late finishers? A longitudinal study of early foreign language learning in school. Language Learning, 67, 631-664. James, Carl & Klein, Kerstin (1994). Foreign language learners’ spelling and proof-reading strategies. Papers and Studies in Contrastive Linguistics, 29, 31-46. Jared, Debra; Levy, Betty A.; Cormier, Pierre & Wade-Woolley, Lesley (2011). Early predictors of biliteracy development in children in French immersion: A 4-year longitudinal study. Journal of Educational Psychology, 103/ 1, 119-139. Jenkins, Joseph R.; Fuchs, Lynn S.; Van den Broek, Paul; Espin, C. & Deno, Stanley L. (2003). Sources of individual differences in reading comprehension and reading fluency. Journal of Educational Psychology, 95/ 4, 719-729. 188 9. References Jenniskens, Tessa; Leest, Bianca; Wolbers, Maarten; Krikhaar, Evelien; Teunissen Cindy; de Graaff, Rick; Unsworth, Sharon & Coppens, Karien (2018). Evaluatie pilot Tweetalig Primair Onderwijs: Vervolgmeting schooljaar 2016/ 17. www.rijksoverheid.nl/ documen ten/ rapporten/ 2018/ 04/ 01/ evaluatie-pilot-tweetalig-primair-onderwijs. Jin, Fufen (2009). Third language acquisition of Norwegian objects: Interlanguage transfer or L1 influence? In: Yan-kit I. Leung (Ed.). Third Language Aquisition and Universal Grammar. Bristol: Multilingual Matters, 144-161. Jochim, Valerie. „Weil Mädchen anders lernen“ - Die Konstruktion von Geschlecht in Grundschulbüchern und ihre heteronormative Wirkmächtigkeit. Eckert. Working Papers, 2014/ 9. www.edumeres.net/ urn/ urn: nbn: de: 0220-2014-00270. Jungbauer-Gans, Monika & Gross, Christiane (2011). More private schools for nonnative students? Migrant performance in private schools of differing national contexts. Education Research International, 2011. doi: 10.1155/ 2011/ 121250/ . Kemper, Thomas (2015). Bildungsdisparitäten von Schülern nach Staatsangehörigkeit und Migrationshintergrund: Eine schulformspezifische Analyse allgemeinbildender Schulen in deutschen Bundesländern anhand von Daten der amtlichen Schulstatistik. Münster: Waxmann. Kennedy, Eithne; Dunphy, Elizabeth; Dwyer, Bernadette; Hayes, Geraldine; McPhillips, Therese; Marsh, Jackie; O’Connor, Maura & Shiel, Gerry. (2012). Literacy in Early Childhood and Primary Education (3-8 Years). Dublin: National Council for Curriculum and Assessment. www.ncca.ie/ en/ Publications/ Reports/ Literacy_in_Early_Childhood_ and_Primary_Education_3-8_years.pdf. Kersten, Kristin (2019). Einflussfaktoren im bilingualen Fremdsprachenerwerb. In: Rohde, Andreas & Steinlen, Anja (Eds.). Sprachenvielfalt als Ressource begreifen. Mehr‐ sprachigkeit in bilingualen Kindertagesstätten und Schulen. Band II. Berlin: Dohrmann, 35-70. Kersten, Kristin; Fischer, Uta; Burmeister, Petra; Lommel, Annette; Schelletter, Christina; Steinlen, Anja K. & Thomas, Shannon (2010b). Immersion in Primary School: A Guide. https: / / files.eric.ed.gov/ fulltext/ ED572883.pdf. Kersten, Kristin; Rohde, Andreas; Schelletter, Christina & Steinlen, Anja K. (Eds., 2010a). Bilingual Preschools. Vol. I: Learning and Development. Vol. II: Best Practices. Trier: WVT. Keßler, Jörg-Uwe & Paulick, Christian (2010). Mehrsprachigkeit und schulisches Eng‐ lischlernen bei Lernern mit Migrationshintergrund. In: Ahrenholtz, Bernt (Ed.). Fachunterricht und Deutsch als Zweitsprache. Tübingen: Narr, 257-278. Kierepka, Adelheid (2020). Mit dem Schriftbild umgehen: Erstes Lesen und Schreiben. In: Heiner Böttger (Ed.). Englisch. Didaktik für die Grundschule. Third Edition. Berlin: Cornelsen, 156-163. 189 9. References Kim, Yoon Kyong; Hutchison, Lindsey A. & Winsler, Adam (2013). Bilingual education in the United States: an historical overview and examination of two-way immersion. Educational Review, 2013. doi: 10.1080/ 00131911.2013.865593. Kishiyama, Mark M.; Boyce, W. Thomas; Jimenez, Amy M.; Perry, Lee M. & Knight, Robert T. (2009). Socioeconomic disparities affect prefrontal function in children. Journal of Cognitive Neuroscience, 21/ 6, 1106-1115. Kissau, Scott & Turnbull, Miles (2008). Boys and French as a second language: A research agenda for greater understanding. Canadian Journal of Applied Linguistics, 11, 151-170. Klemm, Klaus; Hoffmann, Lars; Maaz, Kai & Stanat, Petra (2018). Privatschulen in Deutschland. Trends und Leistungsvergleiche. Berlin: Friedrich-Ebert-Stiftung. http: / / library.fes.de/ pdf-files/ studienfoerderung/ 14189.pdf. KMK (Kultusministerkonferenz, 2004). Bildungsstandards für die erste Fremdsprache (Eng‐ lisch / Französisch) für den Hauptschulabschluss. www.kmk.org/ fileadmin/ Dateien/ veroeffentlichungen_beschluesse/ 2004/ 2004_10_15- Bildungsstandards-ersteFS-Haupt. pdf. KMK (Kultusministerkonferenz, 2013a). Fremdsprachen in der Grundschule - Sachstand und Konzeptionen 2013. www.kmk.org/ fileadmin/ Dateien/ veroeffentlichungen_ beschluesse/ 2013/ 2013_10_17-Fremdsprachen-in-der-Grundschule.pdf. KMK (Kultusministerkonferenz, 2013b). Konzepte für den bilingualen Unterricht - Erfahrungsbericht und Vorschläge zur Weiterentwicklung. www.kmk.org/ fileadmin/ Dateien/ veroeffentlichungen_beschluesse/ 2013/ 201_10_17-Konzepte-_bilingualer-_Unterricht. pdf. Koda, Keiko (2005). Insights into Second Language Reading. New York: Cambridge University Press. Koda, Keiko (2007). Reading and language learning: Constraints on second language reading development. Language, 57/ 1, 1-44. Koda, Keiko (2008). Impacts of prior literacy experience on second-language learning to read. In: Koda, Keiko & Zehler, Annette M. (Eds.). Learning to Read Across Languages. Cross-Linguistic Relationships in Firstand Second Language Literacy Development. New York, NY: Routledge, 68-96. Kolb, Annika (2012). Teaching English in primary school. In: Eisenmann, Maria & Summer, Theresa (Eds.). Basic Issues in EFL Teaching and Learning. Heidelberg: Winter, 30-42. Köller, Olaf; Leucht, Michael & Pant, Hans Anand (2012). Effekte bilingualen Unterrichts auf die Englischleistungen in der Sekundarstufe I. Unterrichtswissenschaft, 40/ 4, 334-350. König, Ekkehard & Gast, Volker (2012). Understanding English-German Contrasts. Third Edition. Berlin: Schmidt. 190 9. References Krashen, Stephen (1982). Principles and Practices in Second Language Acquisition. Oxford: Pergamon. Kress, Jacqueline E. & Fry, Edward B. (2015). The Reading Teacher’s Book of Lists: Edition 6. San Francisco: Jossey-Bass. Krüger, Heinz-Hermann; Budde, Jürgen; Kramer, Rolf-Torsten & Rabe-Kleberg, Ursula (Eds., 2010). Bildungsungleichheit revisited. Bildung und soziale Ungleichheit vom Kindergarten bis zur Hochschule. Wiesbaden: Verlag für Sozialwissenschaften. Kruk, Richard S. & Reynolds, Kristina A. (2012). French immersion experience and reading skill development in at-risk readers. Journal of Child Language, 39, 580-610. Kubota, Ryuko (2003). New approaches to gender, class, and race in second language writing. Journal of Second Language Writing, 12, 31-47. Kuschner, Emily S. (2013). Nonverbal intelligence. In: Volkmar, Fred R. (Ed.). Encyclopedia of Autism Spectrum Disorder. New York: Springer, 2037-2041. Kuska, Sandra Kristina; Zaunbauer, Anna C.M. & Möller, Jens (2010). Sind Immer‐ sionsschüler wirklich leistungsstärker? Ein Lernexperiment. Zeitschrift für Entwick‐ lungspsychologie und Pädagogische Psychologie, 42, 143-153. Lambert, Wallace E. & & MacNamera, John (1969). Some cognitive consequences of following a first grade curriculum in a second language. Journal of Educational Psychology, 60, 86-95. Lambert, Wallace E. & Tucker, G. Richard (1972). Bilingual Education of Children: The St. Lambert Experiment. Rowley, MA: Newbury House. Lasagabaster, David (2001). Bilingualism, immersion programmes and language learning in the Basque Country. Journal of Multilingual and Multicultural Development, 22/ 5, 401-425. Lasagabaster, David (2008). Foreign language competence in content and language integrated courses. The Open Applied Linguistics Journal, 1, 30-41. Lawson, Gwen M.; Hook, Cayce J.; Hackman, Daniel A. & Farah, Martha J. (2016). Socioeconomic status and the development of executive functions: Behavioral and neuroscience approaches. In: Griffin, James A.; Freund, Lisa & McCardle, Peggy D. (Eds.). Executive Function in Preschool-age Children: Integrating Measurement, Neuro‐ development, and Translational Research. Washington, DC: American Psychological Association, 259-278. Lee, Patrick (1996). Cognitive development in bilingual children: A case for bilingual instruction in early childhood education. Bilingual Research Journal, 20, 499-522. Legutke, Michael K.; Müller-Hartmann, Andreas & Schocker-von Ditfurth, Marita (2009). Teaching English in the Primary School. Stuttgart: Klett. Leki, Ilona; Cumming, Alister & Silva, Tony (2008). A Synthesis of Research on Second Language Writing in English. New York: Routledge. 191 9. References Lenhard, Wolfgang & Schneider, Wolfgang (2006). ELFE 1-6: Ein Leseverständnistest für Erstbis Sechstklässler. Göttingen: Hogrefe. Lenz, Peter & Studer, Thomas (2007). Lingualevel. Instrumente zur Evaluation von Fremdsprachenkompetenzen. Bern: Schulverlag. Lesaux, Nonie & Geva, Esther (2006). Synthesis: Development of literacy in language-mi‐ nority students. In: August, Diane & Shanahan, Timothy (Eds.). Developing Literacy in Second-Language Learners: Report of the National Literacy Panel on Language-Minority Children and Youth. Mahwah, NJ: Lawrence Erlbaum, 53-74. Leung, Patrick (2015). Fundamental ELT techniques and strategies for CLIL practitioners. In: Pokrivčáková, Silvia (Ed.). CLIL in Foreign Language Education: E-Textbook for Foreign Language Teachers. Nitra: Constantine the Philosopher University, 119-130. www. klis.pf.ukf.sk/ dokumenty/ CLIL/ e-textbook%20CLIL%20in%20FLE%20-%20final.pdf. Lindholm-Leary, Kathryn (2001). Dual Language Education. Avon, UK: Multilingual Matters. Lindholm-Leary, Kathryn & Borsato, Graciela (2006). Academic achievement. In: Gene‐ see, Fred; Lindholm-Leary, Kathryn; Saunders, William & Christian, Donna (Eds.). Educating English Language Learners. New York: Cambridge University Press, 157-179. Lindholm-Leary, Kathryn & Hernández, Ana (2011). Achievement and language profi‐ ciency of Latino students in dual language programmes: Native English speakers, fluent English/ previous ELLs, and current ELLs. Journal of Multilingual and Multicul‐ tural Development, 32/ 6, 531-545. Lindholm-Leary, Kathryn & Howard, Elizabeth (2008). Language and academic achieve‐ ment in two-way immersion programs. In: Fortune, Tara W. & Tedick, Diane J. (Eds.). Pathways to Bilingualism: Evolving Perspectives on Immersion Education. Clevedon, UK: Multilingual Matters, 177-200. Little, David; Simpson, Barbara; Catibusic, Bronagh (2003). PSAK. Primary School Assess‐ ment Kit. Dublin: Integrate Ireland Language and Training. López-Rúa, Paula (2006). The sex variable in foreign language learning: An integrative approach. Porta Linguarum, 6, 99-114. Lynn, Richard & Wilson, R. Graham (1993). Sex differences in second-language ability: An Irish study. School Psychology International, 14, 275-279. Lyster, Roy (2007). Learning and Teaching Languages Through Content. A Counterbalanced Approach. Amsterdam: John Benjamins. MacCoubrey, Sharon J.; Wade-Woolley, Lesly; Klinger, Don & Kirby, John R. (2004). Early identification of at-risk L2 readers. Canadian Modern Language Review, 61, 11-28. Mady, Callie (2014). Immigrants outperform Canadian-born groups in French immersion: Examining factors that influence their achievement. International Journal of Multilin‐ gualism, 12/ 3, 298-311. 192 9. References Mady, Callie (2017). The bilingual advantage for immigrant students in French immersion in Canada: Linking advantages to contextual variables. International Journal of Bilingual Education and Bilingualism, 20/ 3, 235-251. Malicky, Grace V.; Fagan, William T. & Norman, Charles A. (1988). Reading processes of French Immersion children reading in French and English. Canadian Journal of Education, 13/ 2, 277-289. Maluch, Jessica Tsimprea; Kempert, Sebastian; Neumann, Marko & Stanat, Petra (2015). The effect of speaking a minority language at home on foreign language learning. Learning and Instruction, 36, 76-85. Marian, Victoria; Shook, Anthony & Schroeder, Scott R. (2013). Bilingual two-way immersion programs benefit academic achievement. Bilingual Research Journal, 36/ 2, 167-186. Massler, Ute & Burmeister, Petra (Eds., 2010). CLIL und Immersion. Fremdsprachlicher Sachfachunterricht in der Grundschule. Braunschweig: Westermann. Massler, Ute & Ioannou-Georgiou, Sophie (2010). Best practice: How CLIL works. In: Massler, Ute & Burmeister, Petra (Eds.). CLIL und Immersion. Braunschweig: Westermann, 61-75. Massler, Ute & Steiert, Claudia (2010). Implementierung von CLIL-Modulen - die Per‐ spektive von Lehrenden, Kindern, Eltern. In: Massler, Ute & Burmeister, Petra (Eds.). CLIL und Immersion. Erfolgsbedingungen für CLIL in der Grundschule. Braunschweig: Westermann, 11-29. Massler, Ute; Stotz, Daniel & Queisser, Claudia (2014). Assessment instruments for primary CLIL: the conceptualisation and evaluation of test tasks. The Language Learning Journal, 42/ 2, 137-150. Massler, Ute; Stotz, Daniel; Rehm, Seraina & Rimmele, Marisa (2016). Lernstandsbeur‐ teilung im integrierten Englisch- und Sachunterricht: Aufgaben für Unterricht und Assessment. Braunschweig: Westermann. Mather, Nancy; Hammill, Donald D.; Allen, Elizabeth A. & Roberts, Rhia (2004). Test of Silent Word Reading Fluency. Austin, TX: PRO-ED. May, Peter (1994). Rechtschreibregeln für Mädchen - besondere Wörter für Jungen? I. Herausbildung orthographischer Fähigkeiten im Geschlechtervergleich. II. Jungen und Mädchen schreiben „ihre“ Wörter. Zur Rolle der persönlichen Bedeutung beim Lernen. In: Richter, Sigrun & Brügelmann, Hans (Eds.). Mädchen lernen anders lernen Jungen. Geschlechtsspezifische Unterschiede beim Schriftspracherwerb. Bottighofen am Bodensee: Libelle Verlag, 83-98 and 110-120. May, Peter (1994). Regeln sind für Mädchen - Jungen brauchen Sensationen? ! Herausbil‐ dung orthographischer Fähigkeiten im Geschlechtervergleich. www.peter-may.de/ Dok umente/ May_doc/ May94a.pdf. May, Peter (2002). Hamburger Schreib-Probe. Sixth Edition. Hamburg: vpm. 193 9. References May, Peter (2002). Hamburger Schreibprobe für die Klassen 4 und 5. Hinweise zur Durch‐ führung und Auswertung. Hamburg: Verlag für pädagogische Medien. May, Peter (2006). Englisch-Hörverstehen am Ende der Grundschulzeit. In: Bos, Wilfred & Pietsch, Marcus (Eds.). KESS 4. Kompetenzen und Einstellungen von Schülerinnen und Schülern der 4. Klasse an Hamburger Grundschulen. Münster: Waxmann, 203-224. McElvany, Nele; Becker, Michael & Lüdtke, Oliver (2009). Die Bedeutung familiärer Merkmale für Lesekompetenz, Wortschatz, Lesemotivation und Leseverhalten. Zeitschrift für Entwicklungspsychologie und Pädagogische Psychologie, 41/ 3, 121-131. McElvany, Nele; Kessels, Ursula; Schwabe, Franziska & Kasper, Daniel (2017). Geschlecht und Lesekompetenz. In: Hussmann, Anke; Wendt, Heike; Bos, Wilfried; Bremerich- Vos, Albert; Kasper, Daniel; Lankes, Eva-Maria; McElvany, Nele; Stubbe, Tobias C. & Valtin Renate (Eds.). IGLU 2016. Lesekompetenzen von Grundschulkindern in Deutsch‐ land im internationalen Vergleich. Münster: Waxmann, 177-194. Mehisto, Peeter; Marsh, David & Frigols, María J. (2008). Uncovering CLIL: Content and Language Integrated Learning in Bilingual and Multilingual Education. Oxford: Macmillan. Merga, Margaret (2017). Do males really prefer non-fiction, and why does it matter? English in Australia, 52/ 1, 27-35. Merkel, Christina (2015). Mädchen lernen Fremdsprachen besser als Jungen. Nürnberger Zeitung: February 10, 2015. www.nordbayern.de/ 2.209/ 2.207/ madchen-lernen-fremd sprachen-besser-als-jungen-1.4182741. Met, Myriam (1999). Content-Based Instruction: Defining Terms, Making Decisions. NFLC Reports. Washington, DC: The National Foreign Language Center. www.carla.umn. edu/ cobaltt/ modules/ principles/ decisions.html. Met, Myriam (2002). Expanding on think, pair, share: An instructional framework for scaffolding oral language use by integrating oral and written modalities. Presentation to Twin Cities Immersion Teachers, April 2002. http: / / carla.umn.edu/ cobaltt/ modules/ strategies/ TPS_expanded.pdf. Ministerium Baden-Württemberg für Kultus, Jugend und Sport (2004). Bildungsstandards für Englisch Grundschule: Klassen 2, 4. www.bildungsplaene-bw.de/ site/ bildungsplan/ get/ documents/ lsbw/ Bildungsplaene/ Bildungsplaene-2004/ Bildungsstandards/ Grundschule_Bildungsplan_Gesamt.pdf. Ministerium Baden-Württemberg für Kultus, Jugend und Sport (2008). Niveaukonkretisierung für Englisch Klasse 4: Bilinguales Lernen (Bildungsplan 2004 Grund‐ schule). www.bildungsplaene-bw.de/ site/ bildungsplan/ get/ documents/ lsbw/ Bildungs plaene/ Bildungsplaene-2004/ Niveaukonkretisierung/ Grundschule/ NK-GS-E.pdf. Ministerium Baden-Württemberg für Kultus, Jugend und Sport (2016a). Bildungsstan‐ dards für Englisch Grundschule: Klassen 2, 4. http: / / www.bildung-staerkt-men schen.de/ service/ downloads/ Bildungsstandards/ GS/ GS_E_bs.pdf. 194 9. References Ministerium Baden-Württemberg für Kultus, Jugend und Sport (2016b). Bildungsplan der Grundschule. www.bildungsplaene-bw.de/ ,Lde/ 6555804#FaecherGS. Ministerium Baden-Württemberg für Kultus, Jugend und Sport (2020). Bildungsplan der Grundschule. Englisch (ab Klasse 3/ 4). www.bildungsplaene-bw.de/ site/ bildungsplan/ get/ documents/ lsbw/ export-pdf/ depot-pdf/ ALLG/ BP2016BW_ALLG_GS_E34.pdf. Ministerium für Bildung, Wissenschaft, Forschung und Kultur des Landes Schles‐ wig-Holstein (2007). Lehrpläne für die Primarstufe: Rahmenlehrplan Englisch. www.lehrplan.lernnetz.de/ index.php? wahl=4. Ministerium für Bildung, Wissenschaft, Forschung und Kultur des Landes Schles‐ wig-Holstein (1997). Lehrpläne für die Primarstufe: Heimat- und Sachunterricht. www. lehrplan.lernnetz.de/ index.php? wahl=154. Möller, Christine (2015). Young L2 Learners’ Narrative Discourse: Coherence and Cohesion. Tübingen: Narr. Möller, Jens; Fleckenstein, Johanna; Hohenstein, Friederike; Preusler, Sandra; Paulick, Isabell & Baumert, Jürgen (2018). Varianten und Effekte bilingualen Lernens in der Schule. Zeitschrift für Erziehungswissenschaft, 21, 4-28. Möller, Jens; Hohenstein, Friederike; Fleckenstein, Johanna; Köller, Olaf & Baumert, Jürgen (Eds., 2017a). Erfolgreich integrieren - die Staatliche Europa-Schule Berlin. Münster: Waxmann. Möller, Jens; Hohenstein, Friederike; Fleckenstein, Johanna & Baumert, Jürgen (2017b). Formen und Effekte des Fremdsprachenerwerbs und der bilingualen Beschulung. In: Möller, Jens; Hohenstein, Friederike; Fleckenstein, Johanna; Köller, Olaf & Baumert, Jürgen (Eds.). Erfolgreich integrieren - die Staatliche Europa-Schule Berlin. Münster: Waxmann, 25-48. Morvay, Gabriela (2015). The role of L1 reading ability, L2 proficiency and nonverbal intelligence in L2 reading comprehension. Argentinian Journal of Applied Linguistics, 3/ 1, 26-47. Mossenson, Leila; Stephanou, Andrew; Forster, Margaret; Masters, Geoffrey; McGregor, Margaret; Anderson, Prue; Hill, Peter (2003). Tests of Reading Comprehension (TORCH). Second Edition. Melbourne: ACER Press. Mullis, Ina V. S.; Martin, Michael O.; Foy, Peter & Hooper, Martin (2017). PIRLS 2016 International Results in Reading. Chestnut Hill, MA: TIMSS & PIRLS International Study Center, Boston College. http: / / timssandpirls.bc.edu/ pirls2016/ internationalresults/ . Mussli, Schulz, Ronald (w/ o year). Gender-sensible Lehrplangestaltung für die Grundschule. Beispiele für den Unterricht: Deutsch. www.ph-freiburg.de/ filead min/ dateien/ sonstige/ gleichstellung/ Gender-sensible_Lehrplangestaltg_GS_Unterric htsbeispieleDeutsch.pdf. 195 9. References Myhill, Debra & Fisher, Ros (2010). Editorial: Writing development: cognitive, sociocul‐ tural, linguistic perspectives. Journal of Research in Reading, 33/ 1, 1-3. Na Ranong, Sirirat & Leung, Yan-kit I. (2009). Null objects in L1 Thai-L2 English-L3 Chinese: An empirical take on a theoretical problem. In: Leung, Yan-kit I. (Ed.). Third Language Acquisition and Universal Grammar. Bristol: Multilingual Matters, 162-191. Naglieri, Jack A. (2001). Do ability and reading achievement correlate? Journal of Learning Disabilities, 34, 304-305. NEPS (National Educational Psychology Service, Ireland, 2019). Effective Interventions for Struggling Readers. Second Edition. Dublin. www.education.ie/ en/ Schools-Colleges/ Information/ Resources-Guidance/ Effective-Interventions-for-Struggling-Readers-A- Good-Practice-Guide-for-Teachers.pdf. Nicolay, Anne-Catherine & Poncelet, Martine (2013). Cognitive advantage in children enrolled in second-language immersion elementary school program for three years. Bilingualism: Language and Cognition, 16/ 3, 597-607. Nicolay, Anne-Catherine & Poncelet, Martine (2015). Cognitive benefits in children enrolled in an early bilingual immersion school: A follow up study. Bilingualism: Language and Cognition, 18/ 4, 789-795. Nijakowska, Joanna (2010). Dyslexia in the Foreign Language Classroom. Bristol: Multi‐ lingual Matters. Nikolova, Roumiana & Ivanov, Stanislav (2010). Englischleistungen. In: Bos, Wilfried (Ed.). KESS 8. Kompetenzen und Einstellungen von Schülerinnen und Schülern am Ende der Jahrgangsstufe 8. Münster: Waxmann, 49-66. Nikula, Tarja & Mård-Miettinen, Karita (2014). Language learning in immersion and CLIL classrooms. In: Östman, Jan-Ola & Verschueren, Jef (Eds.). Handbook of Pragmatics. Amsterdam: Benjamins, 1-24. Nold, Daniela (2010). Sozioökonomischer Status von Schülerinnen und Schülern 2008. Ergebnisse des Mikrozensus. Wirtschaft und Statistik 2/ 20, 138-149. www.destatis.de/ DE/ Publikationen/ WirtschaftStatistik/ BildungForschungKultur/ StatusSchueler_22 010.pdf ? __blob=publicationFile. Nold, Günter; Hartig, Johannes; Hinz, Silke & Rossa, Henning (2008). Klassen mit bilin‐ gualem Sachfachunterricht: Englisch als Arbeitssprache. In: DESI-Konsortium (Eds.). Unterricht und Kompetenzerwerb in Deutsch und Englisch. Ergebnisse der DESI-Studie. Weinheim: Beltz, 451-457. O’Brien, Beth A. & Wallot, Sebastian (2016). Silent reading fluency and comprehension in bilingual children. Frontiers in Psychology, 7, Article 1265. doi.org/ 10.3389/ fpsyg. 2016.01265. Odlin, Terence (1989). Language Transfer: Cross-Linguistic Influence in Language Learn‐ ing. Cambridge: Cambridge University Press. 196 9. References OECD (Organisation for Economic Co-Operation and Development, 2000). PISA 2000: Literacy Skills for the World of Tomorrow. Paris: OECD Publishing. OECD (Organisation for Economic Co-Operation and Development, 2019). PISA 2018 Results (Volume I). What Students Know and Can Do. Paris: OECD Publishing, www. oecd.org/ publications/ pisa-2018-results-volume-i-5f07c754-en.htm. Ogle, Donna M. (1986). K-W-L: A teaching model that develops active reading of expository text. Reading Teacher, 39, 564-570. Ossner, Jakob (2008). Sprachdidaktik Deutsch. Second Edition. Paderborn: Schöningh. Oxley, Emily & de Cat, Cécile (2019). A systematic review of language and literacy interventions in children and adolescents with English as an additional language (EAL). The Language Learning Journal. doi: 10.1080/ 09571736.2019.1597146. Parlindungan, Firman (2018). What research has to say about spelling instruction for English language learners. International Seminar on English Language Teaching and Research, 1/ 1, Universitas Islam Malang, 310-325. http: / / eltar.inggrisppsunisma.ac.id/ wp-content/ uploads/ 2018/ 04/ 33.-Firman-Parlindungan-310-325.pdf. Peal, Elizabeth & Lambert, Wallace. (1962). The relation of bilingualism to intelligence. Psychological Monographs, 76/ 546, 1-23. Peregoy, Suzanne F. & Boyle, Owen F. (2013). Reading, Writing, and Learning in ESL. A Resource Book for K-12 Teachers. Edition 6. Boston: Pearson Education. Pérez-Cañado, Maria L. (2012). CLIL research in Europe: past, present, and future. International Journal of Bilingual Education and Bilingualism, 15/ 3, 315-341. doi: 10.1080/ 13670050.2011.630064. Petermann, Franz & Petermann, Ulrike (2008). HAWIK-IV. Kindheit und Entwicklung, 17, 71-75. Pinter, Annamaria (2006). Teaching Young Language Learners. Oxford: Oxford University Press. Pires, Heldemeria S.; Candeias, Adelina A.; Grácio, Luisa; Galindo, Edgar & Melo, Mada‐ lena (2017). The influence of family support according to gender in the Portuguese language course achievement. Frontiers in Psychology, 8, 1610. doi.org/ 10.3389/ fpsyg. 2017.01610. Piske, Thorsten (2010). Positive and negative effects of exposure to L2 orthographic input in the early phases of foreign language learning: A review. In: Diehr, Bärbel & Rymarczyk, Jutta (Eds.). Researching Literacy in a Foreign Language among Primary School Learners - Forschung zum Schriftspracherwerb in der Fremdsprache bei Grund‐ schülern. Frankfurt am Main: Peter Lang, 37-50. Piske, Thorsten (2015). Zum Erwerb der CLIL-Fremdsprache. In: Rüschoff, Bernd; Sudhoff, Julian & Wolff, Dieter (Eds.). CLIL Revisited: Eine kritische Analyse zum gegenwärtigen Stand des bilingualen Sachfachunterrichts. Frankfurt am Main: Peter Lang, 101-125. 197 9. References Piske, Thorsten (2017a). Old dogs and new tricks? Aktuelle Erkenntnisse der Spracher‐ werbsforschung zur Rolle des Alters, des Geschlechts und eines Migrationshinter‐ grunds beim Erlernen von Fremdsprachen. Praxis Englisch 4/ 2017, 44-46. Piske, Thorsten (2017b). The earlier, the better? Some critical remarks on current EFL teaching to young learners and their implications for foreign language teacher education. In: Wilden, Eva & Porsch, Raphaela (Eds.). The Professional Development of Primary EFL Teachers. National and International Research. Münster, New York: Waxmann, 45-57. Piske, Thorsten (2018). Erkenntnisse der Spracherwerbsforschung. In: Heinz, Susanne; Riedel, Andrea & Riecke-Baulecke, Thomas (Eds.). Basiswissen Lehrerbildung. Englisch unterrichten. Seelze: Kallmeyer / Klett, 21-42. Piske, Thorsten (2020). Bilingualer Unterricht an Grundschulen in Deutschland: Was man schon erforscht hat und was wir noch nicht beantworten können. In: Bayerisches Staatsministerium für Unterricht und Kultus. Bilinguale Grundschule Französisch. Leitfaden. Vorläufige Fassung. München: Staatsinstitut für Schulqualität und Bildungs‐ forschung, 13-25. Piske, Thorsten & Burmeister, Petra (2008). Erfahrungen mit früher, englischer Immer‐ sion an norddeutschen Grundschulen. In: Schlemminger, Gerald (Ed.). Erforschung des bilingualen Lehrens und Lernens: Forschungsarbeiten und Erprobungen von Unterrichtskonzepten und -materialien in der Grundschule. Baltmannsweiler: Schneider Verlag Hohengehren, 131-151. Poarch, Gregory J. (2018). Multilingual language control and executive function: A repli‐ cation study. Frontiers in Communication, 3, Article 46. doi: 10.3389/ fcomm.2018.00046. Pokrivčáková, Silvia (2015, Ed.). CLIL in Foreign Language Education: E-Textbook for Foreign Language Teachers. Nitra: Constantine the Philosopher University. www.klis. pf.ukf.sk/ dokumenty/ CLIL/ e-textbook%20CLIL%20in%20FLE%20-%20final.pdf. Rathvon, Natalie (2004). Early Reading Assessment: A Practitioner’s Handbook. New York, NY: Guilford. Rauch, Dominique; Jurecka, Astrid & Hesse, Hermann-Günter (2010). Für den Drittspra‐ cherwerb zählt auch die Lesekompetenz in der Herkunftssprache. Untersuchung der Türkisch-, Deutsch- und Englisch-Lesekompetenz bei Deutsch-Türkisch bilingualen Schülern. In: Allemann-Ghionda, Cristina; Stanat, Petra; Göbel, Kerstin; Röhner, Charlotte (Eds.). Migration, Identität, Sprache und Bildungserfolg. Weinheim: Beltz, 78-100. Raven, John C. (1976a). Colored Progressive Matrices (CPM). Sets A, AB & B. Prepared by JC Raven. San Antonio, TX: Harcourt. Raven, John C. (1976b). Standard Progressive Matrices (SPM). Sets A, B, C, D & E. Prepared by JC Raven. San Antonio, TX: Harcourt. 198 9. References Reeder, Kenneth; Buntain, Jennifer & Takakuwa, Mitsunori (1999). Intensity of L2 instruction and biliterate proficiency in the intermediate years of a French immersion program. The Canadian Modern Language Review, 56, 49-72. Reichart-Wallrabenstein, Maike (2004). Kinder und Schrift im Englischunterricht der Grundschule. Eine theorie- und empiriegeleitete Studie zur Diskussion um die Integration von Schriftlichkeit. Berlin: dissertation.de. Reinhardt, Paula (2018). Written HL Skills of Turkish Children Attending a German Primary School. University Erlangen-Nürnberg: Zulassungsarbeit / Manuscript. Richter, Sigrun & Brügelmann, Hans (1994). Mädchen lernen besser - zu lesen und zu schreiben. Eine Einführung. In: Brügelmann, Hans & Richter, Sigrun (Eds.). Mädchen lernen anders lernen Jungen. Geschlechtsspezifische Unterschiede beim Schriftspracher‐ werb. Konstanz: Libelle, 9-12. Ringbom, Hakan (1987). The Role of the First Language in Foreign Language Learning. Clevedon: Multilingual Matters. Ringbom, Hakan (2001). Lexical transfer in L3 production. In: Cenoz, Jasone; Hufeisen, Britta & Jessner, Ulrike (Eds.). Cross-Linguistic Influence in Third Language Acquisition: Psycholinguistic Perspectives. Clevedon: Multilingual Matters, 59-68. Rischawy, Nina (2016). Bilinguale Züge an Realschulen in Bayern. Vergleichsstudien zum Zuwachs englischsprachiger Fertigkeiten und Kompetenzen im Rahmen des Modellver‐ suchs. Universität Eichstätt-Ingolstadt: Unpublished Dissertation. Rolff, Hans-Günter; Leucht, Michael & Rösner, Ernst (2008). Sozialer und familialer Hintergrund. In: DESI-Konsortium (Eds.). Unterricht und Kompetenzerwerb in Deutsch und Englisch. Ergebnisse der DESI-Studie. Weinheim und Basel: Beltz, 283-300. Roquet, Helena & Pérez-Vidal, Carmen (2017). Do productive skills improve in Content and Language Integrated Learning contexts? The case of writing. Applied Linguistics, 38/ 4, 489-511. Roquet, Helena; Llopis, Jaume & Perez-Vidal, Carmen (2016). Does gender have an impact on the potential benefits learners may achieve in two contexts compared: formal instruction and formal instruction + content and language integrated learning? International Journal of Bilingual Education and Bilingualism, 19/ 4, 370-386. Rothman, Jason (2015). Linguistic and cognitive motivations for the typological primacy model (TPM) of third language (L3) transfer: Timing of acquisition and proficiency considered. Bilingualism: Language and Cognition, 18/ 2, 179-190. Rubin, Hyla; Turner, Anne & Kantor, Miriam (1991). Fourth grade follow-up of reading and spelling skills of French immersion students. Reading and Writing: An Interdisci‐ plinary Journal, 3/ 1, 63-73. Rubin, Renée & Galván Carlan, Veronica (2005). Using writing to understand bilingual children’s literacy development. The Reading Teacher, 58/ 8, 728-739. 199 9. References Rumlich, Dominik (2016). Evaluating Bilingual Education in Germany: CLIL students’ General English Proficiency, EFL Self-Concept and Interest. Frankfurt: Peter Lang. Rymarczyk, Jutta (2008). Früher oder später? Zur Einführung des Schriftbildes in der Grundschule. In: Böttger, Heiner (Ed.). Fortschritte im Frühen Fremdsprachenlernen. Ausgewählte Tagungsbeiträge Nürnberg 2007. München: Domino Verlag, 170-182. Rymarczyk, Jutta (2010). Früher Schriftspracherwerb in der ersten Fremdsprache Eng‐ lisch bei Kindern mit Migrationshintergrund. Forum Sprache. Die Online-Zeitschrift für Fremdsprachenforschung und Fremdsprachenunterricht. 4/ 2010, 60-78. www.hueber. de/ seite/ pg_hefte2010_fos#id191606. Rymarczyk, Jutta (2011). „Lautes Lesen = mangelhaft / Leises Lesen = sehr gut? “. Diskrepanzen in den Leseleistungen von Erst- und Drittklässlern im Fremdsprache‐ nunterricht Englisch. In: Kötter, Markus & Rymarczyk, Jutta (Eds.). Fremdsprachenun‐ terricht in der Grundschule. Forschungsergebnisse und Vorschläge zu seiner weiteren Entwicklung. Frankfurt am Main: Peter Lang, 49-67. Rymarczyk, Jutta (2016). Early reading and writing in the foreign language classroom - maybe not as controversial as one might think. Babylonia, 2/ 2016, 53-57. Rymarczyk, Jutta & Musall, Annika (2010). Reading skills of firstgraders who learn to read and write in German and English. In: Diehr, Bärbel & Rymarczyk, Jutta (Eds.). Researching Literacy in a Foreign Language among Primary School Learners. Frankfurt am Main: Peter Lang, 69-88. Sachverständigenrat deutscher Stiftungen für Integration und Migration (2016). Dop‐ pelt benachteiligt? Kinder und Jugendliche mit Migrationshintergrund im deutschen Bildungssystem. Eine Expertise im Auftrag der Stiftung Mercator. www.stiftung-mer cator.de/ media/ downloads/ 3_Publikationen/ Expertise_Doppelt_benachteiligt.pdf. Sagasta Errasti, Maria P. (2003). Acquiring writing skills in a third language: The positive effects of bilingualism. International Journal of Bilingualism, 7, 27-42. Samuels, Douglas D. & Griffore, Robert J. (1979). The Plattsburgh French language immersion program: Its influence on intelligence and self-esteem. Language Learning, 29, 45-52. Sanchez Perez, Laura (2015). The acquisition of written competence in learners of L3 English in a Spanish/ German Content and Language Integrated Learning Context. In: Marsh, David; Perez Cañado, Maria Luisa & Raez Padilla, Juan (Eds.). CLIL in Action: Voices from the Classroom. Cambridge: Cambridge Uuniversity Press, 64-81. Scheerer-Neumann, Gerheid (1998). Stufenmodelle des Schriftspracherwerbs - Wo ste‐ hen wir heute? In: Balhorn, Heiko (Ed.). Schatzkiste Sprache. Von den Wegen der Kinder in die Schrift. Frankfurt am Main: Arbeitskreis Grundschule DGLS, 54-62. Schipolowsky, Stefan; Wittig, Julia; Weirich, Sebastian & Böhme, Katrin (2017). Geschlechtsbezogene Disparitäten. In: Stanat, Petra; Schipolowski, Stefan; Rjosk, Ca‐ milla; Weirich, Sebastian; Haag, Nicole (Eds.). IQB-Bildungstrend 2016. Sprachliche 200 9. References Kompetenzen am Ende der 4. Jahrgangsstufe im zweiten Ländervergleich. Münster: Waxmann, 187-211. Schmenk, Barbara (2002). Geschlechtsspezifisches Fremdsprachenlernen? Zur Konstruktion geschlechtstypischer Lerner- und Lernbilder in der Fremdsprachenforschung. Tübingen: Stauffenburg. Schmid-Schönbein, Gisela (2008). Didaktik und Methodik für den Englischunterricht. Kompakter Überblick. Ziele - Inhalte - Verfahren. Für die Klassen 1 bis 4. Berlin: Cornelsen. Schneider, Wolfgang; Blanke, Iris; Faust, Verena & Küspert, Petra (2011). Würzburger Leise Leseprobe - Revision. Göttingen: Hogrefe. Schoonen, Rob (2019). Are reading and writing building on the same skills? The relationship between reading and writing in L1 and EFL. Reading and Writing, 32/ 3, 511-535. doi.org/ 10.1007/ s11145-018-9874-1. Schröder, Konrad; Harsch, Claudia & Nold, Günter (2006). DESI - Die sprachpraktischen Kompetenzen unserer Schülerinnen und Schüler im Bereich Englisch. Zentrale Be‐ funde. Neusprachliche Mitteilungen aus Wissenschaft und Praxis, 59/ 3, 11-32. www. the-english-academy.de/ wp-content/ uploads/ 2016/ 08/ DESI-Sprachpraktische_Kom‐ petenzen.pdf. Schroeder, Christoph & Şimşek Yazgul (2014). Das Türkische. In: Krifka, Manfred; Błaszczak, Joanna; Leßmollmann, Annette; Meinunger, Andre; Stiebels, Barbara; Tracy, Rosemarie & Truckenbrodt, Hubert (Eds.). Das mehrsprachige Klassenzimmer. Über die Muttersprachen unserer Schüler. Berlin: Springer, 115-134. Schwab, Götz (2013). Bili für alle? Ergebnisse und Perspektiven eines Forschungsprojekts zur Einführung bilingualer Module in einer Hauptschule. In: Breidbach, Stefan & Viebrock, Britta (Eds.). Content and Language Integrated Learning (CLIL) in Europe. Research Perspectives on Policy and Practice. Frankfurt am Main: Peter Lang, 297-313. Schwippert, Knut; Wendt, Heike & Tarelli, Irmela (2012). Lesekompetenzen von Schü‐ lerinnen und Schülern mit Migrationshintergrund. In: Bos, Wilfred; Tarelli, Irmela; Bremerich-Vos, Albert & Schwippert, Knut (Eds.). IGLU 2011: Lesekompetenzen von Grundschulkindern in Deutschland im internationalen Vergleich. Münster: Waxmann, 191-207. Segeritz, Michael; Stanat, Petra & Walter, Oliver (2010). Muster des schulischen Erfolgs von Mädchen und Jungen mit Migrationshintergrund. In: Allemann-Ghionda, Cris‐ tina; Stanat, Petra; Göbel, Kerstin & Röhner, Charlotte (Eds.). Migration, Identität, Sprache und Bildungserfolg. Weinheim: Beltz, 165-186. Senatsverwaltung Berlin für Bildung, Jugend und Sport (2006). Rahmenlehrplan für die Grundschule und die Sekundarstufe I: Englisch. mdb-sen-bildung-schulorganisa‐ tion-lehrplaene-gr_englisch.pdf. 201 9. References Senatsverwaltung Berlin für Bildung, Jugend und Sport (2016). Ausführungsvorschriften für bilingualen Unterricht an weiterführenden allgemeinbildenden Schulen. www.berlin. de/ sen/ bildung/ schule/ rechtsvorschriften/ av_bilingualer_unterricht.pdf. Sendlmeier, Walter & Oertel, Alexandra (2015). Rechtschreibdidaktiken im ersten Schul‐ jahr. Eine psychologische und sprachwissenschaftliche Einordnung und Bewertung. Berlin: Logos. Sheridan, Margaret A.; Sarsour, Khaled; Jutte, Douglas; D’Esposito, Mark & Boyce, W. Thomas (2012). The impact of social disparity on prefrontal function in childhood. PloS ONE, 7/ 4, e35744. doi: 10.1371/ journal.pone.0035744. Shin, Joan Kang (2017). Literacy instruction for young EFL learners: A balanced approach. Boston, MA: National Geographic Learning. https: / / ngl.cengage.com/ ourworldtours / OurWorld/ media/ Downloads/ Literacy-Instruction-for-Young-EFL-Learners_White- Paper_2017.pdf. Siegert, Manuel (2008). Schulische Bildung von Migranten in Deutschland: Integrationsreport Teil 1 des Bundesamts für Migration und Flüchtlinge. www.bamf.de/ SharedDocs/ Anlagen/ DE/ Publikationen/ WorkingPapers/ wp13-schulische-bildung.pdf ? __blob= publicationFile. Skutnabb-Kangas, Tove & Toukomaa, Pertti (1976). Teaching Migrant Children’s Mother Tongue and Learning the Language of the Host Country in the Context of the Sociocultural Situation of the Migrant Family. Helsinki: The Finnish National Commission for UNESCO. Snow, Marguerite Anne (1990). Instructional methodology in immersion foreign lan‐ guage education. In: Padilla, Amado M.; Fairchild, Halford H. & Valadez, Concepcion M. (Eds.). Foreign Language Education: Issues and Strategies. Newbury Park, CA: Sage, 156-171. Somers, Thomas (2017). Content and Language Integrated Learning and the inclusion of immigrant minority language students: A research review. International Review of Education, 63, 495-520. doi.org/ 10.1007/ s11159-017-9651-4. Sorenson Duncan, Tamara & Paradis, Johanne (2018). How does maternal education influence the linguistic environment supporting bilingual language development in child second language learners of English? International Journal of Bilingualism, 27, doi.org/ 10.1177/ 1367006918768366. Sprenger-Charolles, Liliane & Casalis, Séverine (1996). Lire - Lecture et écriture: acquisi‐ tion et troubles du développement. Paris: Presses Universitaires de France. Stanat, Petra (2006). Disparitäten im schulischen Erfolg: Analysen zur Rolle des Migrationshintergrunds. Unterrichtswissenschaft, 34, 98-124. Stanat, Petra; Böhme, Katrin; Schipolowski, Stefan & Haag, Nicole (Eds., 2016). IQB-Bil‐ dungstrend 2015. Sprachliche Kompetenzen am Ende der 9. Jahrgangsstufe im zweiten Ländervergleich. Münster: Waxmann. 202 9. References Stanovich, Keith; Cunningham, Anne & Feeman, Dorothy (1984). Intelligence, cognitive skills, and early reading progress. Reading Research Quarterly, 19/ 3, 278-303. Statistisches Bundesamt (2012). Bevölkerung mit Migrationshintergrund: Ergebnisse des Mikrozensus 2011. Fachserie 1, Reihe 2.2. Wiesbaden. www.destatis.de. Statistisches Bundesamt (2020a). Bevölkerung mit Migrationshintergrund: Ergebnisse des Mikrozensus 2019. Fachserie 1, Reihe 2.2. Wiesbaden. www.destatis.de/ DE/ Themen/ Gesellschaft-Umwelt/ Bevoelkerung/ Migration-Integration/ Publikationen/ _publikati onen-innen-migrationshintergrund.html; jsessionid=AAC2F6503C501D32338529F3 8C3E2217.internet8722. Statistisches Bundesamt (2020b). Bildung und Kultur: Allgemeinbildende Schulen. Schul‐ jahr 2019/ 20. Fachserie 11 Reihe 1. Wiesbaden: destatis. www.destatis.de/ DE/ Themen / Gesellschaft-Umwelt/ Bildung-Forschung-Kultur/ Schulen/ Publikationen/ Downloads-Schulen/ allgemeinbildende-schulen-2110100207005.html. Steiert, Claudia & Massler, Ute (2011). Guidelines for evaluating and developing CLIL materials. In: Ioannou- Georgiou, Sophie & Pavlou, Pavlos (Eds.). Guidelines for CLIL Implementation in Primary and Pre-primary Education. Nicosia: Chipre, 98-113. www. schools.ac.cy/ klimakio/ Themata/ Anglika/ teaching_material/ clil/ guidelines forclilimplementation1.pdf. Steinig, Wolfgang & Huneke, Hans-Werner (2002). Sprachdidaktik Deutsch (Grundlagen der Germanistik, 38), Second Edition. Berlin: Schmidt. Steinlen, Anja K. (2016). Primary school minority and majority language children in a partial immersion program: The development of German and English reading skills. Journal of Immersion and Content-Based Language Education, 4/ 2, 198-224. Steinlen, Anja K. (2017). The development of English grammar and reading comprehen‐ sion by majority and minority language children in a bilingual primary school. Studies in Second Language Learning and Teaching, 7/ 3, 419-442. Steinlen, Anja K. (2018a). Grundschulen mit verschiedenen Englischprogrammen: Geschlecht und Migration auf dem Prüfstand. Zeitschrift für Fremdsprachenforschung, 29/ 1, 3-26. Steinlen, Anja K. (2018b). The development of English and German writing skills in a bilingual primary school in Germany. Journal of Second Language Writing, 39/ 1, 42-52. Steinlen, Anja K. & Gerdes, Ulrike (2015). Eine Pilotuntersuchung zur sprachlichen Entwicklung von Kindern mit und ohne Migrationshintergrund in einer Grundschule mit bilingualem Sachfachunterricht. In: Méron-Minuth, Sylvie & Özkul, Senem (Eds.). Fremde Sprachen lehren und lernen - Aktuelle Fragen und Forschungsaufgaben. Frank‐ furt am Main: Peter Lang, 81-106. Steinlen, Anja K. & Piske, Thorsten (2013). Academic achievement of children with and without migration backgrounds in an immersion primary school: A pilot study. Zeitschrift für Anglistik und Amerikanistik, 61/ 3, 215-244. 203 9. References Steinlen, Anja K. & Piske, Thorsten (2015). Zur Entwicklung der Schulleistungen von Kindern mit und ohne Migrationshintergrund in einer bilingualen Grundschule: Eine Pilotstudie. In: Kötter, Markus & Rymarczyk, Jutta (Eds.). Englischunterricht auf der Primarstufe. Neue Forschungen - weitere Entwicklungen. Frankfurt am Main: Peter Lang, 123-150. Steinlen, Anja K. & Piske, Thorsten (2016a). Minority language students as at-risk learners: Myth or reality? Findings from an early German-English partial immersion program. In: Ehland, Christoph; Mindt, Ilka & Tönnies, Merle (Eds.). Anglistentag Paderborn 2015: Proceedings (37). Trier: Wissenschaftlicher Verlag Trier, 9-28. Steinlen, Anja K. & Piske, Thorsten (2016b). Wortschatz und Leseverständnis des Englischen von Kindern mit und ohne Migrationshintergrund in einer bilingualen Grundschule. In: Steinlen, Anja K. & Piske, Thorsten (Eds.). Wortschatzentwicklung in bilingualen Schulen und Kindertagesstätten. Forum Angewandte Linguistik. Frankfurt am Main: Peter Lang, 123-166. Steinlen, Anja K. & Piske, Thorsten (2018a). Deutsch- und Englischleistungen von Kindern mit und ohne Migrationshintergrund im bilingualen Unterricht und im Fremdsprachenunterricht: Ein Vergleich. In: Ballis, Anja & Hodaie, Nazli (Eds.). Per‐ spektiven auf Mehrsprachigkeit: Individuum - Bildung - Gesellschaft. Berlin: Mouton De Gruyter, 85-97. Steinlen, Anja K. & Piske, Thorsten (2018b). Die Entwicklung des englischen und deut‐ schen Lesens bei ein- und mehrsprachigen Kindern in Grundschulen mit bilingualen Angeboten. In: Massler, Ute & Kutzelmann, Sabine (Eds.). Fach- und Sprachgrenzen überwinden: Mehrsprachige Leseförderung aus der Perspektive der schulsprachlichen und fremdsprachlichen Didaktik. Studienbuch. Narr: Tübingen, 51-62. Steinlen, Anja K. & Piske, Thorsten (2018c). Abschlussbericht für die Platanus Schule Berlin: Wissenschaftliche Begleitung im Schuljahr 2016/ 2017. Manuskript: Berlin & Erlangen. Steinlen, Anja K. & Piske, Thorsten (2020). *breakfist, *breackfest, *brakefast oder *brackfust? Orthographische Kompetenzen im Englischen in einem bilingualen Grundschulprogramm in Deutschland. Bulletin Suisse de Linguistique Appliquée, No spécial, III-IV, 61-80. https: / / doc.rero.ch/ record/ 11876? ln=fr. Steinlen, Anja K. & Piske, Thorsten (i. pr.). “Does speaking a foreign language at school make your kids smarter? ” The development of nonverbal intelligence in a primary school with an immersion and a mainstream foreign language program. In: Piske, Thorsten & Steinlen, Anja K. (Eds.). Cognition and Second Language Acquisition. Tübingen: Narr. Steinlen, Anja K. & Piske, Thorsten (i. prep.). Dyslexic students in different English language primary schools programs in Germany. 204 9. References Steinlen, Anja K.; Kersten, Kristin & Piske, Thorsten (2019). Mehrsprachige Jungen als Problemfall in bilingualen Kitas? Zur Rolle von Geschlecht und sprachlichem Hintergrund. In: Rohde, Andreas & Steinlen, Anja (Eds.). Sprachenvielfalt als Ressource begreifen. Mehrsprachigkeit in bilingualen Kindertagesstätten und Schulen. Band II. Berlin: Dohrmann, 71-88. Stiftung Bildungspakt Bayern (w/ o year). Bilinguale Grundschule Französisch. https: / / bildungspakt-bayern.de/ bilinguale-grundschule-franzoesisch/ . Stiftung Bildungspakt Bayern (w/ o year). Lernen in zwei Sprachen - Bilinguale Grund‐ schule Englisch. https: / / bildungspakt-bayern.de/ lernen-in-zwei-sprachen-bilingualegrundschule-englisch/ . Swain, Merrill (1975). Writing skills of grade three French immersion pupils. Working Papers on Bilingualism, 7, 1-38. Swain, Merrill & Lapkin, Sharon (1982). Evaluating Bilingual Education: A Canadian Case Study. Clevedon: Multilingual Matters. Swain, Merrill & Lapkin, Sharon (1991). Heritage language children in an English-French bilingual program. The Canadian Modern Language Review, 47, 635-641. Swain, Merrill & Lapkin, Sharon (1995). Problems in output and the cognitive processes they generate: A step towards second language learning. Applied Linguistics, 16/ 3, 371-391. Swain, Merrill; Lapkin, Sharon; Rowen, Norman & Hart, Doug (1990). The role of mother tongue literacy in third language learning. Vox, 4, 111-121. Tamm, Carolin (2010). Eine Schule macht sich auf den Weg: Einführung eines bilingualen Zuges an der Grundschule an der Hügelstraße. In: Massler, Ute & Burmeister, Petra (Eds.). CLIL und Immersion. Fremdsprachlicher Sachfachunterricht in der Grundschule. Braunschweig: Westermann, 30-37. Tel2L (Teacher Education by Learning through Two Languages, 2000). MBE (Mainstream Bilingual Education) Issues. www.unavarra.es/ tel2l/ eng/ MBEIssues.htm#Teacher. Thaler, Engelbert (2012). Englisch unterrichten. Grundlagen - Kompetenzen - Methoden. Berlin: Cornelsen. Thomas, Wayne P. & Collier, Virginia P. (2012). Dual Language Education for a Trans‐ formed World. Albuquerque, NM: Dual Language Education of New Mexico/ Fuente Press. Thomas, Wayne P. & Collier, Virginia P. (2017). Validating the Power of Bilingual Schooling: Thirty-Two Years of Large-Scale, Longitudinal Research. www.thomasandco llier.com/ assets/ aral-2017-(typed).pdf. Treutlein, Anke; Landerl, Karin & Schöler, Hermann (2013). (Frühe) Schrifteinführung im Englischunterricht - Überlegungen zu Zeitpunkt und Methode auf Grundlage von psycholinguistischen Studien. Zeitschrift für Fremdsprachenforschung, 24/ 1, 3-27. 205 9. References Turnbull, Miles; Hart, Doug & Lapkin, Sharon (2003). Grade 6 French immersion students’ performance on large-scale reading, writing, and mathematics tests: Building explan‐ ations. The Alberta Journal of Educational Research, 49/ 1, 6-23. Turnbull, Miles; Lapkin, Sharon & Hart, Doug (2001). Grade three immersion stu‐ dents’ performance in literacy and mathematics: Province-wide results from Ontario (1998-99). The Canadian Modern Language Review, 58/ 1, 9-26. Uhl, Patricia (2016). Potential gender differences in learning French as a foreign language. In: Saleh, Salmiza & Yakob, Nooraida (Eds.). Papers Presented at USM-FAU International Conference. Pulau Pinang: Universiti Sains Malaysia, 47-56. Uhl, Patricia; Schwanke, Katrin & Piske, Thorsten (2020). Französisch in der Grund‐ schule - Erste empirische Ergebnisse aus einem bayernweiten Modellprojekt. In: Böttger, Heiner; Festman, Julia & Müller, Tanja (Eds.). Language Education & Acquis‐ ition Research. Focusing Early Language Learning. Konferenzband. Bad Heilbrunn: Klinkhardt. 62-81. UNESCO (United Nations Educational, Scientific and Cultural Organization, 2012). ISCED 2011. International Standard Classification of Education. Paris: UNESCO Institute of Statistics. http: / / uis.unesco.org/ sites/ default/ files/ documents/ international-standa rd-classification-of-education-isced-2011-en.pdf. Ur, Penny (1996). A Course in Language Teaching. Cambridge: Cambridge University Press. Van der Pol, Janneke; Volman, Monique & Beishuizen, Jos (2010). Scaffolding in teacher-student interaction: A decade of research. Educational Psychology Review, 22/ 3, 271-296. Van der Leij, Aryan; Bekebrede, Judith & Kotterink, Mieke (2010). Acquiring reading and vocabulary in Dutch and English: the effect of concurrent instruction. Reading and Writing, 23/ 3-4, 415-434. Van Dongera, Rixt; Van der Meer, Cor & Sterk, Richt (2017). Research for CULT Committee - Minority Languages and Education: Best Practices and Pitfalls. European Parliament, Policy Department for Structural and Cohesion Policies, Brussels. www.europarl.eur opa.eu/ RegData/ etudes/ STUD/ 2017/ 585915/ IPOL_STU%282017%29585915_EN.pdf. VBW (Vereinigung der Bayerischen Wirtschaft, 2009). (2009). Geschlechterdifferenzen im Bildungssystem. Jahresgutachten 2009. Wiesbaden: Verlag für Sozialwissenschaften. Verhoeven Ludo & Van Leeuwe, Jan (2012). The simple view of second language reading throughout the primary grades. Reading and Writing, 25/ 8, 1805-1818. Wesche, Marjorie Bingham (2002). Early French immersion: How has the original Canadian model stood the test of time? In: Burmeister, Petra; Piske, Thorsten & Rohde, Andreas (Eds.). An Integrated View of Language Development - Papers in Honor of Henning Wode. Trier: Wissenschaftlicher Verlag Trier, 357-379. 206 9. References Whyte, Shona & Schmid, Euline Cutrim (2018). Classroom technology for young learners. In: Garton, Sue & Copland, Fiona (Eds.). The Routledge Handbook of Teaching English to Young Learners. London: Routledge, 338-355. Wilden, Eva & Porsch, Raphaela (2015). Die Hör- und Leseverstehensleistungen von Kindern im Fach Englisch am Ende der Grundschulzeit unter besonderer Berücksich‐ tigung lebensweltlicher Ein- und Mehrsprachigkeit. In: Kötter, Markus & Rymarczyk, Jutta (Eds.). Englischunterricht auf der Primarstufe: Neue Forschungen - weitere En‐ twicklungen. Frankfurt am Main: Peter Lang, 59-80. Wilden, Eva; Porsch, Raphaela & Ritter, Markus (2013). Je früher desto besser? - Frühbeginnender Englischunterricht ab Klasse 1 oder 3. Zeitschrift für Fremdsprachen‐ forschung, 24/ 2, 171-201. Williams, Sara & Hammarberg, Björn (1998). Language switches in L3 production: Implications for a polyglot speaking model. Applied Linguistics, 19, 295-333. Wode, Henning (1995). Lernen in der Fremdsprache: Grundzüge von Immersion und bilingualem Unterricht. Ismaning: Hueber. Wode, Henning (2009). Frühes Fremdsprachenlernen in bilingualen Kindergärten und Grundschulen. Braunschweig: Westermann. Wolff, Dieter (1997). Bilingualer Sachfachunterricht: Versuch einer lernpsychologischen und fachdidaktischen Begründung. In: Vollmer, Hans J. & Thürmann, Eike (Eds.). Englisch als Arbeitssprache im Fachunterricht: Begegnungen zwischen Theorie und Praxis. Soest: Landesinstitut für Schule und Weiterbildung, 50-62. Wolff, Dieter & Südhoff, Julian (2015). Zur Definition des Bilingualen Lehrens und Lernens. In: Rüschoff, Bernd; Südhoff, Julian & Wolff, Dieter (Eds.). CLIL Revisited: Eine kritische Analyse zum gegenwärtigen Stand des bilingualen Sachfachunterrichts. Frankfurt am Main: Peter Lang, 9-42. Wood, David; Bruner, Jerome S. & Ross, Gail (1976). The role of tutoring in problem-solv‐ ing. Journal of Child Psychology and Psychiatry and Allied Disciplines, 17, 89-100. Wright, Irene (1995). Patterns of Writing Development in French by Students in French Im‐ mersion. Vancouver: Simon-Fraser University: MA Thesis. summit.sfu.ca/ system/ files/ iritems1/ 6735/ b17572770.pdf. Yadollahi, Hadis; Piske, Thorsten & Steinlen, Anja (2020). Effects of bilingual programs and digital tools on minority language children’s and adult refugees’ emerging L2 and L3 literacy skills. In: Eisenmann, Maria (Ed.). Beiträge zur Fremdsprachenforschung (BFF), Band 15. Baltmannsweiler: Schneider Verlag Hohengehren, 131-144. Zaunbauer, Anna C.M. & Jens Möller (2006). Schriftsprachliche und mathematische Leis‐ tungen in der Erstsprache: Ein Vergleich monolingual und teilimmersiv unterrichteter Kinder der zweiten und dritten Klassenstufe. Zeitschrift für Fremdsprachenforschung, 17, 181-200. 207 9. References Zaunbauer, Anna C.M. & Möller, Jens (2007). Schulleistungen monolingual und immersiv unterrichteter Kinder am Ende des ersten Schuljahres. Zeitschrift für Entwicklungs‐ psychologie und Pädagogische Psychologie, 39, 141-153. Zaunbauer, Anna C.M. & Möller, Jens (2010). Schulleistungsentwicklung immersiv unterrichteter Grundschüler in den ersten zwei Schuljahren. Psychologie in Erziehung und Unterricht, 57, 30-45. Zaunbauer, Anna C.M.; Bonerad, Eva-Marie & Möller, Jens (2005). Muttersprachliches Leseverständnis immersiv unterrichteter Kinder. Zeitschrift für Pädagogische Psycho‐ logie, 19, 233-235. Zaunbauer, Anna C.M.; Gebauer, Sandra K. & Möller, Jens (2012). Englischleistungen immersiv unterrichteter Schülerinnen und Schüler. Unterrichtswissenschaft: Zeitschrift für Lernforschung, 40/ 4, 315-333. Zaunbauer, Anna C.M.; Gebauer, Sandra K. & Möller, Jens (2013). Bilinguale Grundschu‐ len: Auswirkungen auf das Sachfachwissen am Beispiel Deutsch und Mathematik. In: Steinlen, Anja & Rohde, Andreas (Eds.). Mehrsprachigkeit in bilingualen Kindertages‐ stätten und Schulen: Voraussetzungen - Methoden - Erfolge. Berlin: Dohrmann Verlag, 96-106. Zaunbauer, Anna C.M.; Gebauer, Sandra K.; Retelsdorf, Jan & Möller, Jens (2013). Moti‐ vationale Veränderung von Grundschulkindern in Englisch, Deutsch und Mathematik im Immersions- und Regelunterricht. Zeitschrift für Entwicklungspsychologie und Pädagogische Psychologie, 45/ 2, 91-102. Zöller, Isabelle & Roos, Jeanette (2009). Einfluss individueller Merkmale und familiärer Faktoren auf den Schriftspracherwerb. In: Roos, Jeanette & Schöler, Herrmann (Eds.). Die Entwicklung des Schriftspracherwerbs in der Grundschule. Wiesbaden: Verlag für Sozialwissenschaften, 47-108. Zumhasch, Clemens (2010). Schulleistungen, Selbstkonzepte sowie unterrichtsklimatische Einstellungen deutscher und italienischer Schüler: Quer- und Längsschnittbefunde zu einem bilingualen Schulversuch. Bad Heilbrunn: Klinckhardt. Zydatiß, Wolfgang (2007). Deutsch‐Englische Züge in Berlin (DEZIBEL). Eine Evaluation des bilingualen Sachfachunterrichts an Gymnasien. Kontext, Kompetenzen, Konsequen‐ zen. Frankfurt am Main: Peter Lang. Zydatiß, Wolfgang (2009). Die Gerechtigkeitsfalle bilingualer Bildungsgänge. In: Caspari, Daniela & Hallet, Wolfgang (Eds.). Bilingualer Unterricht macht Schule. Frankfurt am Main: Peter Lang, 161-173. 208 9. References