Ten Years of English Learning at School 3031327586, 9783031327582

Citation preview

Ten Years of English Learning at School Elsa Tragant · Carmen Muñoz

Ten Years of English Learning at School

Elsa Tragant • Carmen Muñoz

Ten Years of English Learning at School

Elsa Tragant Department of Modern Languages and Literatures and English Studies University of Barcelona Barcelona, Spain

Carmen Muñoz Department of Modern Languages and Literatures and English Studies University of Barcelona Barcelona, Spain

ISBN 978-3-031-32758-2    ISBN 978-3-031-32759-9 (eBook) https://doi.org/10.1007/978-3-031-32759-9 © The Editor(s) (if applicable) and The Author(s), under exclusive licence to Springer Nature Switzerland AG 2023 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Cover credit: Eliane SULLE / Alamy Stock Photo This Palgrave Macmillan imprint is published by the registered company Springer Nature Switzerland AG. The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

To our families for their enduring support

Preface

When Biel1 started learning English at a public school in Barcelona (age 6 grade 1), English was his favourite subject. He loved English and would have liked to learn Chinese as well because, he said, his dad had once been to China. When Biel got to high school (age 12 grade 7), he still liked English but it was no longer his favourite subject. He recognised he often paid little attention in class. He was no longer interested in learning an additional foreign language. By the end of secondary education (age 16 grade 10), Biel had given up on English and commented “I have never experienced English” in reference to a lack of opportunities to practise English outside class. He did not see himself as an adult making use of English for personal purposes “Unless I happen to have an English girlfriend”. Like Biel, Albert started learning English at a public school in Barcelona at the age of 6. His favourite subject back then was maths, but he also liked learning English and thought English and other languages would allow him “to make friends from faraway countries”. Once in high school, Albert started attending a language school and by the end of that year he recognised that he had put a lot of effort into learning English. He was satisfied with how much English he had learned. Once he told us of having spoken to a tourist in the street and feeling that “Now I can  All the names that appear in the book are pseudonyms.

1

vii

viii Preface

communicate with more of the world”. In his last year of high school, Albert participated in a school exchange programme with the Netherlands, and he was thrilled by the experience: “From day one we started to speak [with his Dutch partner] and we never stopped”. He was also very pleased with what he had learned at the language school but demotivated with English lessons at school. When asked to define what learning English meant to him, he said: “An experience (…) a different way of thinking, of behaving (…), something positive for me”. Biel and Albert are two of the participants in the ELLiC (Early Language Learning in Catalonia) project. They belong to a group of learners from five schools in Catalonia (Spain) who participated in a longitudinal study that started when they were 6 years old in their first year of primary education and continued until they reached the end of secondary education. Throughout the six chapters of this book, the reader will get to learn about the project and may reach a better understanding of the complexities of learning English as part of compulsory education in Spain and other countries with similar learning conditions. The study in this book is exceptional in that it spans over a very long period of 10 years. It is also remarkable in its methodological diversity and data richness. We have engaged with numerical data and with statistics to look at group and individual outcomes. The qualitative analyses have put flesh on the bones of those results and have added depth to the quantitative results. By using focal participants within the broader study, we have come close to a multiple case study at some points. The readers of this book will find a very personal book, or rather a bipersonal book, arising from our shared experience of prolonged engagement with a group of learners of English. Because of our extended period of study, we witnessed the development of the focal participants from young children to adolescents and came to understand the reasons why some eventually fared better than others in English. We hope some of our excitement about the many insights we gained from their English learning journey reaches the reader. Barcelona, Spain 

Elsa Tragant Carmen Muñoz

Acknowledgements

This book owes sincere thanks to many people, beginning with the learners, who talked to us and confided in us, their teachers, who opened their classes to us, and their parents, who trusted us. The impetus for this long period of research came from the four years of work we shared with the ELLiE (Early Language Learning in Europe) team. The intellectual energy to continue for six more years was nourished by our research group (GRAL) at the University of Barcelona. Enormous thanks go to all of them and especially to M. Rosa Torras, who teamed with us during the first years and very generously enlightened us with her profound knowledge of primary schools. Thanks also go to our students, Marisa Camuñas and Adriana Sevillano, who dealt with the data very efficiently in their MA theses, and to Bridget Murphy, who read and improved the final text. So many more postgraduate students and research assistants helped us at some point during the 10 years of the study that the list would be too long, but we are grateful to each one of them. Our senior colleagues also had useful suggestions, for which we will always be grateful. This book would simply not have been possible without all of them.

ix

x Acknowledgements

For the use of MLAT-EC we owe thanks to M. Mar Suárez, who developed it, and to Charles W. Stansfield, for his support of her work. We are indebted to CSASE (Departament d’Ensenyament) for granting us access to the official language test marks of our learners. Last but not least, we benefitted from different sources of funding to conduct this research over the years, for which we are very thankful: British Council, Lifelong Learning Programme of the European Commission, and RecerCaixa.

About the Book

Chapter 1 begins with an introduction to longitudinal research. The dearth of such an approach in SLA is highlighted and a number of exceptionally long-­term longitudinal studies are discussed. In the second part of the chapter we introduce the reader to our long-term project, ELLiC, by setting the study in its context, presenting the design, and describing the participants and instruments. Chapter 2 explores the influence of learner and contextual factors on learners’ linguistic outcomes at the end of primary school: general language proficiency, listening and reading comprehension, and writing. The participants are the whole group of learners in grade 6. The explanatory factors include attitudes and motivation, language learning aptitude, first language skills, parents’ educational levels, amount of extracurricular lessons, spoken interaction in English, and out-of-school contact with English. Chapter 3 describes the longitudinal development of the focal learners’ written and oral production from grade 5 and grade 4 respectively and until the end of secondary education (grade 10). It is based on yearly written compositions and oral interactive tasks. Group results are complemented by the analysis of three focal learners to illustrate the individual variability in the group.

Chapter 4 explores three dimensions of the focal learners’ language awareness: language learning awareness, metalinguistic awareness, and crosslinguistic awareness. Through learners’ responses to interview xi

xii 

About the Book

questions, we can observe the changes that FL learners’ awareness experiences over the course of the primary and secondary school years.

Chapter 5 examines the learning conditions (family, extracurricular English, interaction and exposure) a group of focal learners grew up in. We adopt a qualitative approach to show how the factors analysed are closely connected and to better understand how learners ended up having different degrees of success in English at the end of secondary education.

Chapter 6 analyses language learning motivation at the end of secondary education in a group of focal learners. Two profiles are identified for the less and more successful learners each. An inductive approach is followed guided by self-determination theory and the Ideal L2 Self within the L2 Motivational Self System framework. Learners’ visions as future users of English are also examined within the four profiles as well as their motivational trajectories starting in grade 1.

Contents

1 Longitudinal Research: The ELLiC Project  1 2 English at the End of Primary School: Explanatory Factors 29 3 W  ritten and Oral Production Development through Primary and Secondary School 69 4 The Development of Young Learners’ Language Awareness111 5 Levels  of Success with English and Learning Conditions: Same Opportunities?149 6 Motivation  after Ten Years: Learner Profiles with a Time Dimension201 I ndex247

xiii

About the Authors

Carmen Muñoz  is Full Professor of Applied English Linguistics at the University of Barcelona (Spain). Her research interests include the effects of age on second language acquisition, young learners in instructed settings, individual differences, out-of-school exposure, and multimodality in language learning. She is the coordinator of the GRAL research group. She has coordinated several funded research projects, among them the BAF Project and, more recently, the SUBTiLL Project on the effects of captions/subtitles on second language learning. She has edited contributed volumes, e.g., Age and the Rate of Foreign Language Learning (2006), and special issues in international journals. She has published extensively in peer-reviewed journals. Recent papers include: “Audiovisual input for learning L2 vocabulary and grammatical constructions” (co-authored) in Second Language Research (2022) and “Exploring repeated captioning viewing as a way to promote vocabulary learning: Time lag between repetitions and learner factors” (co-authored) in Computer Assisted Language Learning (2022). In 2010 she was granted the ICREA Academia award by the Catalan Government and in 2016 the EuroSLA Distinguished Scholar Award given by EuroSLA (European Second Language Association).

xv

xvi 

About the Authors

Elsa Tragant  is Associate Professor of Applied English Linguistics at the University of Barcelona (Spain). Her research interests include L2 motivation, classroom research and interaction, informal learning and qualitative research methods. She is a member of the GRAL research group. She has coordinated several funded research projects and, more recently, the LLMod project on the role of time distribution and online interaction in different language course modalities. She has published in several peerreviewed journals. Recent papers include: “Measuring the effects of repeated exposure to children’s graded readers” (co-authored) in Language Teaching Research (2021), “The L2 Self and Identity: Exploring what Spanish as a foreign language means for former mixed-­major study abroad sojourners” (co-authored) in Frontiers: The Interdisciplinary Journal of Study Abroad (2022), and “Extending language learning beyond the EFL classroom through WhatsApp” (co-authored) in Computer Assisted Language Learning (2021).

List of Figures

Fig. 1.1 Fig. 2.1 Fig. 3.1 Fig. 3.2 Fig. 3.3 Fig. 3.4 Fig. 3.5 Fig. 3.6 Fig. 3.7 Fig. 3.8 Fig. 3.9 Fig. 3.10 Fig. 3.11 Fig. 3.12 Fig. 3.13 Fig. 3.14 Fig. 3.15 Fig. 3.16 Fig. 3.17

The ELLiC design Interaction between parents’ educational level and aptitude Illustration of a written task in G7 Written fluency: Words per text Written lexical diversity: Guiraud’s Index Written syntactic complexity: Clauses per sentence Written syntactic complexity: Coordinate clauses per sentence Written syntactic complexity: Subordinate clauses per sentence Written accuracy: Errors per 100 words Proportion of morphosyntactic and lexical errors Written fluency: Words per text. Three learners Written lexical diversity: Guiraud’s Index. Three learners Written syntactic complexity: Clauses per sentence. Three learners Written syntactic complexity: Coordinate clauses per sentence. Three learners Written syntactic complexity: Subordinate clauses per sentence. Three learners Written accuracy: Errors per 100 words. Three learners Proportion of morphosyntactic errors out of total errors. Three learners Oral fluency: Utterances longer than 2 words Oral lexical diversity: Guiraud’s Index

13 60 75 81 82 82 83 83 84 85 86 86 87 87 88 89 89 91 91 xvii

xviii 

List of Figures

Fig. 3.18 Fig. 3.19 Fig. 3.20 Fig. 3.21

Oral syntactic complexity: Mean length of utterance Oral fluency: Utterances longer than 2 words. Three learners Oral lexical diversity: Guiraud’s Index. Three learners Oral syntactic complexity: Mean Length of Utterance. Three learners Percentages of responses. Evolution of self-concept in primary school Percentages of most frequently mentioned areas of difficulty per grade Percentages of most common responses concerning crosslinguistic differences Question prompt (Color figure online) Percentages of explanations of gender agreement rule per grade Fathers’ and mothers’ levels of education Parents’ English skills

Fig. 4.1 Fig. 4.2 Fig. 4.3 Fig. 4.4 Fig. 4.5 Fig. 5.1 Fig. 5.2

92 93 93 94 117 123 128 132 133 159 160

List of Illustrations

Illustration 1.1 Learner questionaire in G6 (sample items from answer sheet)19 Illustration 1.2 Listening test: a sample item from the G6 test 21 Illustration 1.3 Comic strip 22 Illustration 5.1 Berta’s drawing (MS) 173 Illustration 5.2 Pilar’s drawing (MS) 182 Illustration 5.3 Rita’s word web (MS) 187 Illustration 6.1 Andreu’s (LS) word web “English now” (G10) (Profile 1) 209 Illustration 6.2 Biel’s (LS) teachers’ comments (Profile 2) 214 Illustration 6.3 Antoni’s (LS) word web “English now” (left) and “English in the future” (right) (G10) (Profile 2) 215 Illustration 6.4 Pep’s (MS) word web “English now” (G10) (Profile 3) 220 Illustration 6.5 Cinta’s (MS) word web “English now” (Profile 4) 230 Illustration 6.6 Albert’s answers (MS) to “For you, what is learning English?” in G7 and G10 (Profile 4) 236 Illustration 6.7 Learners’ motivational profiles 238

xix

List of Tables

Table 1.1 Table 1.2 Table 1.3 Table 2.1 Table 2.2 Table 2.3 Table 2.4

The five primary schools in ELLiC 14 The eight secondary schools in the ELLiC project (G7) 16 Overview of instruments/measures 24 Descriptive statistics by gender and school 36 Explanatory factors and data collection times 40 Descriptive statistics of explanatory variables 42 Descriptive statistics of interaction with English speakers. Yearly frequencies and percentages 42 Table 2.5 Descriptive statistics of parents’ educational level. Frequencies and percentages 42 Table 2.6 Descriptive statistics of L2 outcome variables 42 Table 2.7 Correlations of language learning aptitude and multilingual achievement43 Table 2.8 Correlations of fathers’ and mothers’ educational levels 44 Table 2.9 Correlations of parents’ educational level 45 Table 2.10 Correlations between the outcome variables 46 Table 2.11 Summary of results 51 Table 2.12 Estimates of fixed effects of LMMs 59 Table 3.1 Whole group. Significant differences in written and oral measures95 Table 3.2 Estimates. Writing development 103 Table 3.3 Results of RMs Linear Mixed Models. Significant differences in writing development 104 xxi

xxii 

List of Tables

Table 3.4 Table 3.5 Table 4.1 Table 5.1 Table 5.2 Table 5.3 Table 5.4 Table 5.5 Table 5.6 Table 5.7 Table 5.8 Table 5.9 Table 5.10 Table 5.11 Table 5.12 Table 5.13 Table 5.14 Table 6.1 Table 6.2 Table 6.3 Table 6.4 Table 6.5 Table 6.6

Estimates. Oral development Results of RMs Linear Mixed Models. Significant differences in oral development Questions on language awareness and grades Less successful students’ marks More successful students’ marks Less successful students’ extracurricular instruction More successful students’ extracurricular instruction Less successful students’ oral/spoken interaction in English (G1-G10) More successful students’ oral/spoken interaction in English (G1-G10) Less successful students’ leisure activities in English (G4) More successful students’ leisure activities in English (G4) More successful students’ strategies in primary school More successful students’ leisure activities in G7 Synthesis of the less successful students’ contextual factors Synthesis of the more successful students’ contextual factors LS students’ parents’ levels of education and English skills MS students’ parents’ levels of education and English skills A selection of questions on motivation in the G10 interview Profile 1 (LS students): Key features in G10 Profile 2 (LS students): Key features in G10 Profile 3 (MS students): Key features in G10 Profile 4 (MS students): Signs of motivation in G10 Profile 4 (MS students): Key features in G10

105 105 115 156 157 166 167 171 172 180 181 182 185 190 191 195 196 207 212 219 226 228 236

1 Longitudinal Research: The ELLiC Project

Why is longitudinal research essential to advancement of knowledge in the field of SLA? The simple but uncontestable answer is that many questions concerning second language learning are fundamentally questions of time and timing Ortega and Iberri-Shea (2005, p. 27)

1.1 The Story Behind a Long Journey The ELLiC project (Early Language Learning in Catalonia) had its beginning when we joined the ELLiE project (Early Language Learning in Europe). See Enever (2011) for a report of the ELLiE project, which was conceived as a transnational longitudinal initiative in order to give a rich description of language learning and teaching in the first years of foreign language (FL) instruction in primary schools in Europe. When ELLiE ended, after four years of close collaboration with our colleagues in Croatia, England, Italy, the Netherlands, Poland, and Sweden, we felt compelled to continue following our ELLiE learners1 in Catalonia (Spain)  The terms “learner” and “student” are used interchangeably in the book.

1

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 E. Tragant, C. Muñoz, Ten Years of English Learning at School, https://doi.org/10.1007/978-3-031-32759-9_1

1

2 

E. Tragant and C. Muñoz

until the end of primary education. We had been in their English lessons from grade 1 (from hereon G1), spoken to their English teachers, measured their progress from grade to grade and spoken to a subgroup of focal learners year after year. We resisted the idea of putting an end to our longitudinal investigation. After G6 we were clear about not wanting to miss the opportunity to witness how those young learners were going to cope with English in their transition to high school. With that aim in mind, we decided to trace our focal learners from their last year in primary school to their first year in high school (G7). We succeeded and were able to observe them intensely during that “special” year. We did not go back to the high schools for two years until G10. We knew that G10 would be our last chance to contact our focal learners and that was when we collected data from them for the last time. Our ten-year-long “journey” with them ended then. Our experience conducting the ELLiC project for such a long time has been both a gratifying and demanding endeavour. Like Ortega and Iberri-­Shea (2005) and many other SLA (second language acquisition) researchers, like Long and Doughty (2003) or R. Ellis (2021), we also acknowledge the centrality of studying change over time and the need to conduct longitudinal research in our field. It is not uncommon for cross-­ sectional studies to end with a call for longitudinal research or for longitudinal studies to suggest a lengthier follow-up study. Longitudinal research is not the norm in our field and deserves a review in Sect. 1.2 of this chapter as a preamble to the ELLiC project. In this review we first introduce different methodological approaches that are appropriate for the study of change over time and then focus on “classical” longitudinal research with special attention to studies that run over several years in school contexts, which is what the ELLiC project is about. Section 1.3 is an introduction to the ELLiC project and provides the contextual and methodological information that the reader will need later on to understand the empirical studies in this book.

1  Longitudinal Research: The ELLiC Project 

3

1.2 Longitudinal Research One of the first decisions SLA researchers must make when planning research is about its design. If the primary research focus is on change over time, which corresponds to the approach of this project, researchers have four main options to choose from: a simultaneous cross-sectional design, a repeated cross-sectional design, a “classic” longitudinal design and a retrospective study. In studies following a simultaneous cross-­ sectional design, data from learners of different age groups or proficiency levels are recorded. In studies following a repeated cross-sectional design, measures of the same population (but different participants) are recorded at different times. In studies following a longitudinal design, the same participant or participants are recorded at different points in time or continuously over a period of time. Retrospective longitudinal studies, like life history studies, draw on participants’ reporting of past events. The metaphor of a long-distance journey by train with a number of intermediate train stations may be helpful to exemplify the difference between these four designs with a focus on time. In a simultaneous cross-­ sectional study, researchers simultaneously collect data at different train stations (= at different times) from different passengers (usually groups of learners) travelling during the same period of time (e.g., academic year, semester etc.). In a repeated cross-sectional design, the researcher collects data from one single train station once every x time from passengers who are different but have the same profile and make the same journey on different occasions. In a longitudinal study, on the other hand, researchers wait to get on the same train at a number of train stations to be able to trace or take record of one or more passengers as they travel. Finally, in retrospective studies the researcher will meet the passenger(s) at one of the stations or even the final destination and will ask them to rememorate the journey from its beginning or from a certain location or point in time to the present location or train station. On some occasions, a simultaneous cross-sectional design with a focus on time is the preferred option because data can be collected more quickly and can involve large samples. If the study requires data from a large number of participants, it may involve a large team of researchers but this

4 

E. Tragant and C. Muñoz

type of design allows us to have results much sooner than if we had opted for a repeated cross-sectional or a “classic” longitudinal design. Other advantages of simultaneous cross-sectional research with a focus on time is that it is often more likely to lead to generalisation than “classic” longitudinal research, which necessarily involves smaller samples. Cross-­ sectional research of this type also avoids the problem of attrition, inherent in longitudinal studies. Repeated cross-sectional designs can be considered partially longitudinal in nature because data is collected at different rounds from the same population. This type of research is often conducted by institutions to record changes over extended periods of time (i.e., every five years) or to evaluate educational programmes. Whereas this type of research can address important questions regarding the impact of policies, historical changes, and second language (L2) programmes, it may be less attractive for individual researchers or PhD students whose work is evaluated in shorter time scales. Longitudinal research, which by definition has a focus on time, has different advantages and challenges when compared to the two types of cross-sectional research presented above. According to Ortega and Iberri-­ Shea (2005), there are many questions concerning L2 learning that require a truly longitudinal approach, like for example how L2 competence develops over time, the durability of the effects of L2 instructional practices over L2 learning, developmental changes, among others. Longitudinal research may be more manageable for individual researchers or small research teams than simultaneous cross-sectional research since data collection spreads over time. The focus on smaller samples over time also allows for a more situated approach to SLA research. The research design that we adopt in this book can be classified as longitudinal research. If we go back into the history of SLA, it is not difficult to find influential longitudinal studies. In the seventies, for example, Wode (1976) carried out systematic longitudinal observations of German-speaking children learning English as an L2 and English-speaking children learning German as an L2. He was able to question a universal developmental sequence for the acquisition of a language irrespective of whether it is acquired as a first language (L1) or an L2. A good example of longitudinal

1  Longitudinal Research: The ELLiC Project 

5

research from the mid-1980s is the comparative study conducted by Perdue and colleagues (1993a, b) in the context of the European Science Foundation. The aim of the study was to investigate the spontaneous L2 acquisition in adult immigrants learning English, German, Dutch, French, and Swedish over a two-and-a-half-year period. A later illustration of an influential longitudinal study was carried out by Norton (2000) in her two-year-long study on five recently arrived adult learners in Canada, where she suggests a poststructuralist conception of identity in language learning. In the school context, the Barcelona Age Factor project (BAF) is an often-cited contribution to longitudinal research where a subset of learners who had started learning English at different ages were tracked over time (Muñoz, 2006). Despite the suitability of longitudinal research to answer fundamental problems in SLA, it is surprising to find a dearth of SLA research that follows this approach. In 2005 Ortega and Iberri-Shea conducted a review of 20 major applied linguistics journals over a period of three years (2002–2004) and could only identify 38 longitudinal studies. They also observed that among those 38 studies, only 20% took place in high school or middle school contexts and even fewer involved younger learners. In a later review by Barkaoui (2014) on quantitative longitudinal research, the scarce presence of this approach in published research was also made evident. Only 36 studies could be identified over a ten-year period (2003–2013) from the more than 30 journals under review that quantitatively measured the same dependent variable(s) on the same participant(s) on two or more occasions regardless of study length. Years later, Bridgewater (2020) conducted a review similar to Ortega and Iberri-Shea’s work (2005) and identified studies where the authors self-­ labelled their work as “longitudinal” from a collection of 9 journals over three and a half years (2017–2020). The journals under review were: Applied Linguistics, Language Learning, Language Teaching Research, Second Language Research, TESOL Quarterly, The Modern Language Journal, Annual Review of Applied Linguistics, Journal of Second Language Writing, and Studies in Second Language Acquisition. Out of the roughly 640 articles published in the years under review, 97 were self-labelled as longitudinal. In comparison to Ortega and Iberri-Shea’s review, this figure indicates an increase in the presence of longitudinal research in SLA

6 

E. Tragant and C. Muñoz

journals, yet the proportion continues to be very low (roughly 15% of all published research in those journals). In fact, in a recent brief history of SLA, R. Ellis (2021, p. 196) reiterates that there still exists a lack of longitudinal studies, something that Long and Doughty had already observed in their introduction to The handbook of second language acquisition back in 2003.

1.2.1 Longitudinal Research and Study Length Study length is a characteristic feature of longitudinal research even though it is not a defining feature in definitions of the term such as in Barkaoui (2014), Loewen and Plonsky (2016) or Riazi (2016). This is because study length can vary widely depending on the nature of the study. Ortega and Iberri-Shea (2005) opt to describe typical study length based on the longitudinal studies they review. In the case of studies on L2 development, the time span is between four months and four years. For longitudinal research on L2 programmes outcomes, the study length is typically between four to six years, whereas for investigations of L2 instructional effectiveness there seems to be a preference for eight-week-­ long interventions and post-tests between one and three months later. No temporal references are provided in Ortega and Iberri-Shea for qualitative longitudinal SLA research but they do mention that much of this research does not document change over time even if data is collected over extended periods of time. Variation in study length is also evidenced in Bridgewater’s review of longitudinal classroom research where the shortest study ran over six days (Kunitz & Marian, 2017) and the longest ones spanned 8 years (Kibler, 2017; Kibler & Hardigree, 2017). She also observed that a majority of studies covered relatively short periods of time. Out of the 27 studies that she analysed, there were only four publications that ran over one year and most of them had a duration of a semester or a few weeks. This is especially understandable for research taking place at universities, where the basic unit of tuition (i.e., courses) are organised in semesters or quarters and courses do not tend to have continuity.

1  Longitudinal Research: The ELLiC Project 

7

Longitudinal research taking place over more than a few weeks can be more feasible in a school context than in other contexts because the duration of compulsory education ranges over several years (from 9 to 15 years in Europe, according to the European Commission, 2022). It may be even more feasible in educational systems like Spain’s where courses tend to have a year-long duration or in schools with fairly stable student populations, as is the case of the participating schools in the present project. But even in school contexts, long-term longitudinal research is always a challenge. For example, in Costache et al.’s (2022) three-year-long study taking place in Swiss high schools, one of their eight participating schools dropped out. Something similar happened in Cable et al.’s (2010) larger study in England, where three primary schools (out of a pool of 40) dropped out during their three-year longitudinal study. Attrition at the level of individual learners has also proved to be problematic in school-­ based research, especially in studies that cover the transition from primary to secondary education. De Wilde et al. (2021) could only track 1 in 7 (that is, 13%) of the 800 learners in a study that first tested students in their last year of primary school in Flanders and a second time at the end of the second year of secondary school. Attrition was also unavoidable in the BAF project (Muñoz, 2006) in part because of the learners who were lost in the transition from primary to secondary school and also because learners who began extracurricular English classes during the period of the study or who repeated a school year were excluded from the longitudinal analyses. This meant that only 19% of the 11-year-­ old learners from the initial sample were retained two years later. The proportion was even lower (9%) among 13-year-old learners.

1.2.2 Long-Term Longitudinal Research at Schools In spite of the challenges of conducting longitudinal research, studies do exist in SLA that run over exceptionally long periods of time. In this section, six pieces of research (with a minimum study duration of six years) have been selected to illustrate different affordances of long-term longitudinal research.

8 

E. Tragant and C. Muñoz

Tragant and Victori’s work (2006) on learning strategies and age is a good illustration of long-term descriptive research in the context of the BAF project in Catalonia. In that study, a subset of the longitudinal learners (n = 48) answered the same survey at three points in time over a period of six years (ages 10;9 at Time 1; 12;9 at Time 2 and 16;9 at Time 3). The survey provided evidence of progress in the number of strategies learners reported as they grew older and in the nature of the strategies reported over time. Overall the results from the longitudinal sample complemented those from the larger cross-sectional sample (n = 360) and strengthened the internal validity of the study. A different example of long-term longitudinal research comes from a case study of three Swedish learners (Lindgren & Enever, 2017) who were followed over their six years in primary school. In the thick descriptions of these three learners, who were originally part of the ELLiE project, the reader gets a vivid depiction of how their L2 use in the home environment was not originally the same for the three learners and did not evolve in the same way over time. The reader also learns about how the home environment had a visible effect on learners’ interlanguage development as they grew older. Had this case study been based on a shorter time scale, the reader would have lost depth in understanding the learners’ trajectories with English. A long-term study of a different nature was conducted by Lightbown et  al. (2002) with over 800 French-speaking learners in the Canadian province of Nouveau-Brunswick who started learning English in grade 3 (age 8–9). The learners participated in an experimental programme based on comprehension-based learning that consisted of learners basically learning English from age-appropriate materials of their choice that they would read and listen to independently (30 minutes per day). The programme was originally evaluated after the first three years of its implementation with very promising results (Lightbown, 1992; Lightbown & Halter, 1989). Learners’ attitudes towards the experimental programme were very positive and their performance in English was comparable to that of learners who received audio-lingual instruction (regular ESL programme). Such positive outcomes motivated the continuation of the comprehension-based programme into secondary school and its evaluation six years later with a sample of 73 learners (age 14–15). This

1  Longitudinal Research: The ELLiC Project 

9

follow-up study revealed that secondary school learners’ assessment of their learning was in sharp contrast to findings from these same learners in primary school: they now had different expectations of what an ESL programme should be like. Their performance was still comparable to that of learners in the regular programme regarding measures of comprehension and some measures of oral production but not on measures of written production. In this case, the long-term evaluation of the programme was crucial in detecting the limits of the comprehension-based approach. In Sparks et al. (2009), long-term longitudinal research was used to predict language proficiency in high school students in the US where foreign language study generally starts at the beginning of secondary education. The study started with data collection of L1 skills and L1 academic measures from 156 students in 1st grade, continued during elementary school and carried over into secondary school. The participants were 54 students who had chosen to study a foreign language in high school in 9th grade and had done so for two years (until 10th grade) when L2 proficiency measures were administered. The availability of data from early elementary school allowed the researchers to conclude that differences in L2 proficiency (and L2 aptitude) could be related to students’ lower levels of L1 skills (particularly L1 decoding skills) that had been acquired many years before L2 instruction took place, rather than to low motivation. The ten-year-long period in the design of this study was a necessary element to be able to predict the power of L1 skills in the US school system. Heining-Boynton and Haitema’s work (2007) was also conducted over a 10-year period and consisted of two studies about attitudes towards early foreign language learning. The first one was a large-scale four-year-­ long survey-based investigation of over 20,000 school learners of Spanish and French in elementary schools in the US (second to fifth graders). The second follow-up study was based on structured interviews to 13 high school learners (11th and 12th graders) who had answered the surveys in the first study, were enrolled in a foreign language course in high school and volunteered to be interviewed ten years later. The time interval of this investigation allowed the researchers to conclude that early foreign language instruction had had a positive long-standing impact on

10 

E. Tragant and C. Muñoz

adolescents’ perceptions of foreign language speakers, cultures and their own foreign language education. Another piece of research with a large gap between the major study and the follow-up study is that of Lamb (2018). This study was part of a research project originally conducted when a group of 12 junior high students in Indonesia started their formal study of English. Years later Lamb contacted those same learners again twice when they were in senior high school at the age of 22–24 (between 12 and 13 years after the start of the project). The study was inspired by learners’ spontaneous references to the researcher’s influence on their motivation to learn English in the earlier interviews. In his last contact with these students, Lamb asked them to talk directly about his influence in their language learning trajectory and re-analysed earlier interviews. The study is valuable in that it underscores the relationship that is established with participants when research, particularly interviews, is conducted by the same researcher over such long periods of time. In the case of Lamb, his occasional interactions with learners (fewer than 10 spread over 12–13  years in less than 30-­minute interviews) had a tangible and long-lasting effect on both learners who had been originally chosen because of their high motivation as well as those who were less motivated. After each meeting the relationship between Lamb and his participants became closer, the interviews less structured, and the data more trustworthy. The last piece of research that we would like to bring to the readers’ attention is Pfenninger’s study (2020) on age of onset in an immersive school context in Switzerland. Unlike the studies we have just described above, where data was often collected once a year, this study stands out because data was collected every three months. The many measurements that were gathered throughout the eight years of the study (which ranged from 16 to 32 times depending on the group) allowed the researcher to map out the height and shape of patterns of L2 development between groups. Inspection of individual learner trajectories over the multiple measurements also showed that systematic variability within learners grew with time. In sum, longitudinal research is a logical design when the aim is to explore change over time even though it has never been a very frequent methodological approach in SLA research. This is because, among other

1  Longitudinal Research: The ELLiC Project 

11

reasons, tracing change over time is challenging, with school and participant attrition being a major drawback. There is also a tendency in many longitudinal studies for study length to not go beyond a few weeks. Nevertheless, by going through a selection of long-term longitudinal studies, we have shown that this type of exceptional research is versatile and can be used to describe patterns of change over extended periods of time, reveal individual variability with time, provide thick descriptions of language learners’ trajectories, make long-term predictions of L2 performance, evaluate L2 programmes in the long run and even create fruitful researcher-participant relationships. Long-term longitudinal research is especially attractive in school contexts and is the approach that the ELLiC project has adopted.

1.3 The ELLiC Project The ELLiC project took place in the Barcelona area of Catalonia, where Catalan and Spanish are widely used in the community. Spanish is the majority language in Spain and Catalan is by law the language of communication at school. The use of Catalan to teach at school guarantees that all learners, independently of their family background, end up having a full command of the two official languages (Catalan and Spanish) in Catalonia. The educational system also guarantees that learners have a good command of at least one foreign language by the end of compulsory education, which is almost always English. Compulsory education in Catalonia and the whole of Spain lasts for 10 years. The first six years take place in primary schools (from ages 6–7 to ages 11–12) while the next four years take place in high schools (from ages 12–13 to ages 15–16) and are part of lower secondary education (ESO). Many primary schools also provide preschool education (from ages 3–4 to ages 5–6). Foreign language instruction is required to start no later than G1 in primary schools when students are 6–7 years old even though schools can start foreign language instruction in pre-school. Schools also have the option of starting a second foreign language in the last two years of primary education and teach one or more content subjects in the first foreign language. At the time of the ELLiC project, primary schools had

12 

E. Tragant and C. Muñoz

certain flexibility in the amount of hours devoted to the first foreign language ranging from a minimum of 315 hours to a maximum of 420 hours (in six years). The distribution of the minimum hours of instruction was regulated and increased gradually from a total of 70  hours distributed between G1 and G2 to a total of 105 hours distributed between G3 and G4, and a total of 140 hours distributed between G5 and G6. Foreign language instruction continues to be compulsory in lower secondary education in Catalonia and Spain but the school curriculum for foreign languages is a bit more restrictive than in primary education. At the time of the ELLiC project, learners in Catalonia received a total of 420 hours of instruction in the first foreign language which almost always was English. Foreign language instruction was distributed fairly equally during the four years of compulsory schooling in secondary education. Nevertheless, there was some variability in the instruction of a second foreign language, which could be on offer as an optional subject in some grade levels and high schools, especially in large high schools. Like in primary education, high schools could choose to teach content subjects in English.

1.3.1 The ELLiC Design The ELLiC project started as part of ELLiE and focused on five participant primary schools in the Barcelona area. One or two classes per school were followed year after year until the end of primary education (G6). The ELLiC project continued throughout secondary education and followed a smaller number of the original participants in eight different high schools. Data collection in those high schools was carried out in their first and last years of secondary education only (G7 and G10). Figure  1.1 illustrates the design of the study with black and grey dots standing for primary and secondary education respectively. The blank dots indicate no data collection. Access to primary schools was requested through the schools’ principals and access to secondary schools was requested through the Department of Education (regional government). The ELLiC project includes two samples, a general sample with data from whole groups during primary school, and a focal learner sample

1  Longitudinal Research: The ELLiC Project  Grade 1

Grade 3

Grade 2

Grade 5

Grade 4

Grade 7

Grade 6

13

Grade 9

Grade 8

Grade 10

Fig. 1.1  The ELLiC design

gathered from a few learners from each group in primary and secondary school (see Sects. 1.3.2 and 1.3.3 respectively). The project includes quantitative as well as qualitative studies where multiple instruments are used (see Sect. 1.3.4) and data is gathered from learners of English as well as their English teachers and parents.

1.3.2 The Primary Schools in ELLiC and Participants Five primary schools (A, B, C, D, and E) were selected for participation in this project because they were representative of average state schools in the area of Barcelona. Their students came from a wide range of socioeconomic backgrounds. None of the schools were located in deprived neighbourhoods. The proportion of subsidised lunches was low in all five schools and ranged from 1% to 11%. Immigration was low in all schools even if it ranged from schools with hardly any newcomers to schools with 2 newcomers out of every 10 students. School A was located next to a university’s teacher training school and used to be a site for student teachers to do their practicum. This school together with school C were involved in an institutional project to promote English instruction. Schools D and E were originally independent schools and had a good reputation in their neighbourhoods. School B also had a good reputation for English instruction. Except for school C, where English started in preschool (age 5), in the rest of the schools English instruction started in G1 at the beginning of the ELLiE project. Table 1.1 summarises the main features of the five participating schools. Parental permission was obtained for each participant through the school. From G1 (ages 6–7) to G6 (ages 15–16), some of the data was gathered collectively from the whole group of students. In G1 these classes

14 

E. Tragant and C. Muñoz

Table 1.1  The five primary schools in ELLiC School A School location

School sizea Family profiles

School B

School C

School D

School E

Near Downtown Close to Downtown Residential downtown Barcelona waterfront in Barcelona area in Barcelona Barcelona mid-size town (45 km from Barcelona) 2 2 1 2 1 Well-­ Lower Lower middle, Middle lass Mixed established middle lower class background working class and some class international professionals 11.3% 6% 4.6% 5.1% 1%

Subsidised lunches Newcomers ≤20%

≤10%

≤10%

≤5%

1%

Number of classes per grade level

a

ranged from 20 to 27 students per group. One class participated from each school except for school A in which two groups participated. Initially there were a total of 140 students in the G1 class lists and a considerable number of them stayed in the school until the end of primary education. Comparison of class lists in G1 and G6 yields an attrition of 28%. More information about the whole group or general sample can be found in Chap. 2. Some data was collected individually from focal learners, a subgroup of students in each school (initially 2–4 boys and 2–4 girls per school). These learners were chosen by their English teachers in G1 after telling them we were interested in a variety of learner profiles. Nevertheless, in the long run we came to realise that some of the teachers mostly selected good learners, probably to make a positive impression on us. Our initial focal learner sample in G1 included 30 learners (16 boys and 14 girls). Except for one learner who was born in Argentina, the rest of the students were born in Spain and spoke Catalan (n  =  17), Catalan and Spanish (n = 9), Spanish (n = 2), or Chinese (n = 1) at home.2 Seventeen of the focal learners had at least one parent with a university degree. In  This information was not available from one of the focal learners.

2

1  Longitudinal Research: The ELLiC Project 

15

one of the schools a student left in G3 and was substituted by another student. Two other students left their schools later (in G5 and G6) and were not substituted.

1.3.3 The Secondary Schools in ELLiC and Participants When the ELLiC students moved from primary to secondary education, we succeeded in tracing 28 of the focal learners in G6 in their new high schools in G7. This was possible because the majority of students (n = 25) transferred to the same five high schools, and the remaining three students who attended three other high schools were convenient for us to reach. Six of the eight high schools where we collected data from in G7 were state schools. The other two schools were charter schools, that is, independent schools that were partially funded by the regional government. The proportion of newcomers in these schools ranged from 5% to 20%. See Table 1.2 for further information about the high schools in the ELLiC project. By the time these 28 focal learners were in G10, three students were lost because they were no longer in their initial high school or we were not able to contact them. The final number of focal learners from whom we could collect data during both primary school (G1–G6) and secondary school (G7 and G10) is 25. Six of these learners obtained a low final course mark in English in G10 (≤5 out of a maximum 10), fourteen obtained a high mark (≥8) and the remaining five got either a 6 or 7.

1.3.4 English Instruction in ELLiC Schools A total of 17 primary school teachers taught English to the ELLiC participants for one to up to four school years. Most of these teachers, but not all of them, had been trained as primary school English teachers. The number of English sessions per week and their duration varied from school to school and from year to year within the same school. However, there was a general trend in G1–G2 for English instruction to be provided for two to three 45-minute-long class sessions per week. From G3

Near downtown

School C1 (n = 5)

Residential Close to area outside waterfront city centre State State

School B (n = 5) Close to waterfront

School C2 (n = 1) Downtown

School D1b (n = 5)

a

School E1b (n = 4)

School E2b (n = 1)

Mixed background

Mixed background

Number of classes per grade level, bStudents from secondary schools D1 and D2 had attended primary school D.  The same applies for secondary schools E1 and E2, cNumber of focal learners per high school

2

4

Near Mid-size town Village close downtown (45 km from to mid-size Barcelona) town Charter State State

School D2 (n = 1)

School State Religious State typology charter School 4 4 4 2 5 5 sizea Family Mixed Mixed Mixed Mixed Mixed Middle and profiles background background background background background upper middle

School location

School A (n = 6)c

Table 1.2  The eight secondary schools in the ELLiC project (G7)

16  E. Tragant and C. Muñoz

1  Longitudinal Research: The ELLiC Project 

17

on, the trend was for English instruction to be provided in 45- or 60-minute-long sessions three to four times a week. Whenever possible in some schools, English instruction was taught to half the class at a time for one of the weekly sessions. In primary schools A, C, and E this happened more often than in primary schools D and B.  The methodological approach in the five primary schools varied from teacher to teacher, but there was a tendency for teachers with less training to rely on mechanical practice and a textbook and for better trained teachers to follow a more communicative approach and use a wider variety of teaching materials. Students in schools A and C learned science in English for one hour a week from G3 to G6. A total of twelve different English teachers taught the 28 focal learners in ELLiC when they moved to secondary school. In these schools English was taught for three sessions a week and one of these sessions was taught with half the group of students. In some schools one or two optional quarterly subjects were taught in English. English teachers in these high schools faced a complex situation because of large classes (30–33 students) and because they had very little knowledge of what or how their “new” students had been learning English previously in primary schools. This was due to a lack of communication with English teachers in feeder schools. Another source of complexity was the range of learners’ levels of proficiency because of feeder schools with different levels of English and some learners having attended or attending private English language schools in addition to their school lessons. The general trend for these English teachers, who had a good level of grammar knowledge as English philologists, was to follow a form-focused approach and start teaching grammar from zero in G7  in an attempt to balance learners’ levels of proficiency. For more information about the transition from primary to secondary school in the ELLiC project see Muñoz et  al. (2015) and Camuñas (2013).

18 

E. Tragant and C. Muñoz

1.3.5 Instruments Interviews, Questionnaires and Other Instruments This section describes data elicited from interviews, questionnaires and other instruments in order to capture perceptions from primary and secondary school English teachers, learners and their parents.

Teacher Interviews and Questionnaires Data was elicited from primary school English teachers in G1–G6 and high school teachers in G7. In G10 it was not possible to gather any data from the focal learners’ English teachers. Once every year teachers answered our questions and completed a questionnaire about their teaching methodology and the focal learners. In G7 teachers were also asked about the topic of transition from primary to secondary school in an additional interview. All teacher interviews, which were conducted in Catalan or Spanish and often lasted for 30–45  minutes, were audio recorded and transcribed.

Learner Interviews and Questionnaires Focal learners were interviewed individually at least once a year in primary school, twice in G7 (in December and May/June) and one last time in G10. During these interviews, which were originally developed as part of the ELLiE project, learners were encouraged to express their perceptions of their own learning process and their attitudes towards the classroom. In the G7 and G10 interviews, the ELLiE interview script was expanded to better fit the high school context. From G6 on a second interview was developed as part of the ELLiC project with a focus on metalinguistic and crosslinguistic awareness. All the interviews were conducted in Catalan (the school language). The length of the interviews

1  Longitudinal Research: The ELLiC Project 

19

tended to increase over time. For example, in G1–G2 most of them lasted between 5 and 6 minutes while in G10 they lasted from 12 to 25 minutes. An L2 motivation/attitudes questionnaire was administered to all students in class during primary education and to focal learners only in G7 and G10. It was created as part of the ELLiE project (Enever, 2011) and included questions like: “Do you like English this year?”; “Is English easy or difficult for you this year?”. Information about extracurricular instruction was also obtained in this questionnaire (except in G1 and G2 when the teacher provided us with this information). Illustration 1.1 shows the answer choice in the G2–G6 questionnaire. The questionnaire was always written in Catalan and administered towards the end of each school year. In G7 it was also administered in December. Five additional items were added in G10 about parental encouragement (Brady, 2015).

Parents’ Questionnaires Two pencil and paper questionnaires were sent home with the children in G2 and G4 as part of the ELLiE project (Enever, 2011). The questionnaires enquired about the parents’ bio data, their children’s out-of-school contact with English, observations of their children’s learning of English, etc. A total of 73 and 175 parents’ questionnaires were collected from G2 and G4 respectively. 1











Dislike a lot

Dislike

Neither like nor dislike

Like

Like a lot









Neither easy nor difficult

Easy

Very easy

2



Very difficult Difficult

Illustration 1.1  Learner questionaire in G6 (sample items from answer sheet)

20 

E. Tragant and C. Muñoz

Other Instruments In an attempt to use age-appropriate data collection methods, two instruments with a strong visual component were used in G4 and G10. This modality of data made sense given the centrality of visual images in children’s and teenagers’ lives. In G4 learners were asked to produce a drawing in connection with English (“When you hear the word ‘English’, which image comes to your mind?”). In G10 learners were asked to complete two word webs with the prompts “English now” and “English in the future”.

Linguistic and Cognitive Measures In this section we provide a general description of the following linguistic measures: two general proficiency measures, two comprehension tests and two production tasks. The comprehension tests and the oral production tasks were originally developed as part of the ELLiE project (Enever, 2011). A language aptitude test is also described at the end of this section.

General Proficiency Measures Two indicators of general English proficiency were: the scores obtained from an official external examination in G6 and final course marks. The official examination assessed L2 listening and reading comprehension (14 and 24 multiple choice items respectively) and L2 written production (between 8 and 10 lines). The exam is considered a reliable indicator of learners’ English language achievement at the end of primary school. Scores from the Catalan and Spanish tests in the official external examination were also collected. All tests in the official examination were standardised tests. See Chap. 2 for additional information about these exams. The second English measure was the final course marks for English, which were collected every year.

1  Longitudinal Research: The ELLiC Project 

21

Listening and Reading Comprehension Tests The listening test included a selection of items from the test used in the BAF project (Muñoz, 2006) and it consisted of a picture selection task in which learners were expected to match oral prompts with one of three visuals (see Illustration 1.2). The test, which was administered from G1 to G6, never required any reading or writing from learners but the number and difficulty of the items increased over time to adapt to learners’ increasing level of proficiency. The test was administered collectively during class time at the end of the school year. The reading test, which was administered from G4 to G6, consisted of a picture comic strip for children. Learners were asked to read short sentences and match them to the illustrations (see Illustration 1.3). In G6 additional true/false items were added to increase the difficulty of the test. The reading test, like the listening test, never required learners to do any writing and was administered collectively during class time at the end of the school year.

Production Tasks Learners were asked to perform a written and an oral production task towards the end of every school year. The written task was administered from G5 on while the oral task was used from G4 on. The written task was about writing a letter to a pen friend from a different country.

Illustration 1.2  Listening test: a sample item from the G6 test Instructions: Circle the correct picture Oral prompt 8: “Mum puts John’s lunch on the table”

22 

E. Tragant and C. Muñoz

Illustration 1.3  Comic strip. (Pesquis I Baliga-CF1119 ©Cavall Fort 2009, by Viladoms – All rights reserved) Instructions: Write a letter in every bubble. There is an extra letter

Learners completed the task in ten minutes (maximum) by handwriting during class time in G5 and G6 and individually outside class in G7 and G10. The oral task, which was always administered individually, included short questions about themselves, their family, and friends and a guessing game based on a picture in which many different people appeared. The scene of the picture was a classroom scene in G4, G6, and G7, a supermarket in G5 and a scene at the beach in G10 but all scenes always included a similar number of people (21–24). During the guessing game, the researcher and the learner took turns guessing a character of their choice.

1  Longitudinal Research: The ELLiC Project 

23

Language Aptitude Test Learners’ language learning aptitude was measured by means of the MLAT-E (elementary version of the Modern Language Aptitude Test) in its Catalan adaptation (Suárez, 2010). The test was administered to intact classes in G6. It consists of four parts: Hidden Words (ability to associate sounds and symbols), Matching Words (sensitivity to grammatical structure), Finding Rhymes (ability to hear speech sounds), and Number Learning (auditory alertness and ability to remember).

Lesson Observations Two types of lesson observations were conducted: group and focal learner observations. At least two group observations per teacher were carried out each year except in G7 when observations took place in only some schools. It was not possible to conduct lesson observations in G10. A template was used while the lesson was going on in which a chronological description of the stages of the lessons plus field notes were recorded. After each lesson, a post-observation sheet was completed that included structured and open questions. From G1 to G4, observations of the focal learners were also conducted in which three students were observed at a time once or twice per school year. The templates were originally developed as part of the ELLiE project (Enever, 2011) and were gradually modified during the course of the study.

Summary of Instruments Table 1.3 lists all the instruments/measures in the ELLiC project with an indication of when they were used.

24 

E. Tragant and C. Muñoz

Table 1.3  Overview of instruments/measures Instruments/ measures Interviews, Teacher interview questionnaires and and other instruments questionnaire Learner ELLiE interview Meta- and cross-ling awareness interview Learner questionnaire Parents’ questionnaire Learners’ drawings Learners’ word webs Linguistic and Official external cognitive measures examination scores Course marks (English) Listening test Reading test Written task Oral task Language aptitude test Observations Classroom observations Focal learners’ observations

Secondary education

Primary education

G1 G2 G3 G4 G5 G6 G7 x x x x x x x

G10

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x x x x

x

x

x

x

x

x

x

x

x

x x x

x x x x

x x x x x

x

x

x

x

x

x

x

x

x

x

x

x

x x

x x

x

1  Longitudinal Research: The ELLiC Project 

25

References Barkaoui, K. (2014). Quantitative approaches for analysing longitudinal data in second language research. Annual Review of Applied Linguistics, 34, 65–101. https://doi.org/10.1017/S0267190514000105 Brady, I. K. (2015). The ideal and ought-to L2 selves in Spanish learners of English [Doctoral dissertation]. Universidad de Murcia. TESEO. https://digitum. um.es/xmlui/handle/10201/45976 Bridgewater, C. (2020). Longitudinal classroom-based research: A synthesis of L2 studies [Master’s thesis]. Universitat de Barcelona. Retrieved February 9, 2023, from http://hdl.handle.net/2445/193380 Cable, C., Driscoll, P., Mitchell, R., Sing, S., Cremin, T., Earl, J., et al. (2010). Language learning at key stage 2: A longitudinal study (DCSF Research Reports DCSF-RR198). Department for Children, Schools, and Families. Camuñas, M. (2013). The provision of English as a foreign language at transition from primary to secondary school in the classrooms of Catalonia [Master’s thesis]. Universitat de Barcelona. Retrieved February 9, 2023, from http://hdl.handle.net/2445/193307 Costache, O., Becker, E.  S., & Goetz, T. (2022). Is English the culprit? Longitudinal associations between students’ value beliefs in English, German, and French in multilingual Switzerland. Modern Language Journal, 106(2), 313–327. https://doi.org/10.1111/modl.12774 De Wilde, V., Brysbaert, M., & Eyckmans, J. (2021). Young learners’ L2 English after the onset of instruction: Longitudinal development of L2 proficiency and the role of individual differences. Bilingualism: Language and Cognition, 24, 439–453. https://doi.org/10.1017/S1366728920000747 Ellis, R. (2021). A short history of SLA: Where have we come from and where are we going? Language Teaching, 54, 190–205. https://doi.org/10.1017/ S0261444820000038 Enever, J. (Ed.). (2011). ELLiE. Early language learning in Europe. British Council. European Commission, European Education, and Culture Executive Agency, Motiejūnaitė-Schulmeister, A., Sicurella, A., & Birch, P. (2022). The structure of the European education systems 2022/2023: Schematic diagrams. Publications Office of the European Union. https://data.europa.eu/doi/10.2797/21002 Heining-Boynton, A. L., & Haitema, T. (2007). A ten-year chronicle of student attitudes toward foreign language in the elementary school. The Modern Language Journal, 91(2), 149–168. http://www.jstor.org/stable/4625998

26 

E. Tragant and C. Muñoz

Kibler, A. K. (2017). Becoming a “Mexican feminist”: A minoritized bilingual’s development of disciplinary identities through writing. Journal of Second Language Writing, 38(October), 26–41. https://doi.org/10.1016/j. jslw.2017.10.011 Kibler, A.  K., & Hardigree, C. (2017). Using evidence in L2 argumentative writing: A longitudinal case study across high school and university. Language Learning, 67(1), 75–109. https://doi.org/10.1111/lang.12198 Kunitz, S., & Marian, K.  S. (2017). Tracking immanent language learning behaviour over time in task-based classroom work. TESOL Quarterly, 51(3), 507–535. https://doi.org/10.1002/tesq.389 Lamb, M. (2018). When motivation research motivates: Issues in long-term empirical investigations. Innovation in Language Learning and Teaching, 12(4), 357–370. https://doi.org/10.1080/17501229.2016.1251438 Lightbown, P. M. (1992). Can they do it themselves? A comprehension-based ESL course for young children. In R. Courchêne, J. S. John, C. Thérien, & J.  Glidden (Eds.), Comprehension-based language teaching: Current trends (pp. 353–370). University of Ottawa Press. Lightbown, P. M., & Halter, R. (1989). Evaluation of ESL learning regular and experimental programs in four New Brunswick school districts: 1987–88 (Vol. 2). Research report prepared for the Ministry of Education of New Brunswick. Lightbown, P. M., Halter, R. H., White, J., & Horst, M. (2002). Comprehension-­ based learning: The limits of “do it yourself ”. Canadian Modern Language Review, 58, 427–464. https://doi.org/10.3138/cmlr.58.3.427 Lindgren, E., & Enever, J. (2017). Employing mixed methods for the construction of thick descriptions of early language learning. In J.  Enever & E.  Lindgren (Eds.), Early language learning. Complexity and mixed methods (pp. 201–221). Multilingual Matters. Loewen, S., & Plonsky, L. (2016). An A-Z of applied linguistics in research methods. Palgrave. Long, M., & Doughty, C. (2003). SLA and cognitive science. In C. Doughty & M. Long (Eds.), The handbook of second language acquisition (pp. 866–870). Blackwell. Muñoz, C. (Ed.). (2006). Age and the rate of foreign language learning. Multilingual Matters. Muñoz, C., Tragant, E., & Camuñas, M. (2015). Transition: Continuity or a fresh start? APAC ELT Journal, 80, 11–16. Retrieved January 17, 2023, from https://static1.squarespace.com/static/594fcf2ca5790afdb0a78fb6/t/605 0c7aaf88b82622ea045ad/1615906762195/APAC+ELT+Journal+n.80.pdf

1  Longitudinal Research: The ELLiC Project 

27

Norton, B. (2000). Identity and language learning: Gender, ethnicity, and educational change. Longman. Ortega, L., & Iberri-Shea, G. (2005). Longitudinal research in second language acquisition: Recent trends and future directions. Annual Review of Applied Linguistics, 25, 26–45. https://doi.org/10.1017/S0267190505000024 Perdue, C. (Ed.). (1993a). Adult language acquisition: Cross-linguistic perspectives. Vol. I, Field methods. Cambridge University Press. Perdue, C. (Ed.). (1993b). Adult language acquisition: Cross-linguistic perspectives. Vol. II, The results. Cambridge University Press. Pfenninger, S. E. (2020). The dynamic multicausality of age of first bilingual language exposure: Evidence from a longitudinal content and language integrated learning study with dense time serial measurements. The Modern Language Journal, 104, 662–686. https://doi.org/10.1111/modl.12666 Riazi, A.  M. (2016). The Routledge encyclopedia of research methods in applied linguistics. Quantitative, qualitative, and mixed-methods research. Routledge. Sparks, R., Patton, J., Ganschow, L., & Humbach, N. (2009). Long-term relationships among early first language skills, second language aptitude, second language affect, and later second language proficiency. Applied Psycholinguistics, 30(4), 725–755. https://doi.org/10.1017/S0142716409990099 Suárez, M. (2010). Language aptitude in young learners: The elementary modern language aptitude test in Spanish and Catalan [Doctoral dissertation]. University of Barcelona. Retrieved January 30, 2022, from http://www.tesisenred.net/handle/10803/38244 Tragant, E., & Victori, M. (2006). Reported strategy use and age. In C. Muñoz (Ed.), Age and the rate of foreign language learning (pp.  208–236). Multilingual Matters. Wode, H. (1976). Developmental sequences in naturalistic L2 acquisition. Working Papers on Bilingualism, 1(1), 1–13.

2 English at the End of Primary School: Explanatory Factors

… taking a look at second-language acquisition is … like watching a feather drop from an airplane, buffeted by winds, weighted by moisture, and slowed by pressure. Just as observing the feather in a real and changing atmosphere teaches us about winds and other environmental factors, studying how one acquires a second language holds out the promise of helping us to understand the role of the diverse conditions under which human learning occurs. Bialystok and Hakuta (1994, p. 4)

2.1 Introduction The present chapter aims to deepen our understanding of the diverse conditions under which second language learning in primary school occurs. Specifically, we aim to identify the factors that may predict learners’ language outcomes at the end of primary school, both those contextual factors shifting the learner in different directions at various points as well as the internal factors that interact with the external ones in the learner’s progression. Most research concerned with the effect of individual differences on second language learning has been conducted with adult learners, and there is a scarcity of research with primary school © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 E. Tragant, C. Muñoz, Ten Years of English Learning at School, https://doi.org/10.1007/978-3-031-32759-9_2

29

30 

E. Tragant and C. Muñoz

children. Likewise, most previous research on individual differences in primary school learners has singled out one variable (e.g., motivation) or type of variables (e.g., affective variables), providing only a partial view of the forces that determine second language (L2) learning (but see Berthele and Udry [2021], for a recent exception). In this chapter we include affective, cognitive, and contextual factors. Our assumption is that this multi-factor influence is dynamic, and that high variation characterises the linguistic (and non-linguistic) development of primary school learners. Just a snapshot of that process at such a significant point as the end of primary school has the potential to enrich our knowledge about young school learners’ L2 development. In tune with the general aim of this book - to offer a comprehensive view of L2 learners’ development  - we focus on a variety of linguistic dimensions measured by tests of listening comprehension, reading comprehension, and general proficiency. The receptive skills targeted are listening and reading skills. In addition, we include the analysis of a productive skill, writing, and we measure learners’ written performance by means of a set of indices of complexity and fluency. The next section presents the affective, cognitive, and contextual variables that are examined in this chapter, followed by the research questions that guided the study concerned with the associations among those potentially predictive factors and with their relative influence on the linguistic outcomes.

2.2 Affective, Cognitive, and Contextual Variables Individual differences among young language learners have not been widely explored, probably because young learners are considered to be similar to one another (Mihaljević Djigunović, 2015) or, more precisely, because an early age of learning is believed to be a unifying key to general success (but see Muñoz, 2006b, 2008). However, both research and classroom evidence tell us that young language learners are at least as different from one another as older learners. In the present analyses we have included affective, cognitive, and contextual variables in the

2  English at the End of Primary School: Explanatory Factors 

31

understanding that they all impact L2 learning with different strengths at different points in time. Attitudes towards English learning is a well-researched affective variable. Attitudes have been conceptualised and defined in different ways. In relation to L2 learning specifically, Wenden’s (1991) definition of attitude comprises three components: a cognitive component involving perceptions and beliefs, an evaluative component provoking like or dislike, and a behavioural component prompting learners to develop specific learning behaviours. Attitudes are seen as closely linked to motivation, with attitudes traditionally included in more complex models of motivation (see, for example, Gardner’s [2001] socio-educational model), and these two affective variables are often investigated together (e.g., Heinzmann, 2013). Research has shown that learner attitudes are shaped over time by factors such as gender and developmental changes with age (Myles, 2022), as well as by contextual factors such as the school and the wider environment (e.g., Cable et al., 2010; Kissau, 2006). Researchers of the pioneer modern language teaching programmes in elementary schools in Europe identified the development of positive attitudes and motivation as the main gains of an early start to language learning (Blondin et al., 1998; Edelenbos et al., 2006). At that time, it was generally expected that an early introduction to foreign language learning would ensure learners’ continuing engagement at higher educational levels. A number of studies have since shown that the learners’ initial positive attitudes and motivation wane with age and classroom experience or, in other words, that an early introduction of a foreign language (FL) does not by itself generate long-lasting motivation and positive attitudes. Empirical research has also shown that while motivation and attitudes appear at first linked to the teacher and the classroom processes (e.g., the fun activities characteristic of primary school; see Muñoz, 2017b), these factors later become associated to learner-internal variables such as a sense of competence and progress (e.g., Chambers, 2016; Courtney et  al., 2017; Graham et al., 2016; Nikolov, 1999). The decrease in motivation and the more negative attitudes in secondary school have been associated to changes in pedagogies (e.g., less singing and games and more grammar exercises) and issues related to an abrupt transition between primary and secondary school (Chambers, 2019; Muñoz et al., 2015). An earlier drop

32 

E. Tragant and C. Muñoz

in motivation has also been observed in the later years of primary school (Cable et  al., 2010), likely associated also to pedagogical and developmental changes. On the other hand, perceptions of achievement and progress may result in more positive attitudes and higher motivation over time, as we will see in Chap. 6 (see also Muñoz, 2017a). Motivation together with language learning aptitude have consistently been found to be the best predictors of L2 learning (Skehan, 2002). Moreover, these two variables make unique contributions to language learning as indicated by the results of a meta-analysis showing that they are uncorrelated (Li, 2016). Language learning aptitude generally refers to the specific talent for learning a foreign or second language (Carroll, 1981). From its inception, language learning aptitude could be considered as an umbrella term consisting of a set of cognitive abilities: phonetic coding ability, grammatical sensitivity, inductive language learning ability and associative memory (Carroll & Sapon, 1967), although the traditional conception of aptitude treated it as a unitary construct. A long-standing contention is whether language learning aptitude is innate and stable or trainable. The partially innate nature of aptitude finds support in its association with first language (L1) development (Skehan & Ducroquet, 1988) and in recent neurobiological research (e.g., Turker et al., 2021). The modifiable nature of aptitude has been highlighted by SLA researchers who contend that aptitude may be trained (e.g., McLaughlin, 1990) or changed over time or through the interaction with learners’ affective and social variables (e.g., Dörnyei, 2010). The association evidenced in research between L1 development and L2 learning aptitude has led to the consideration of the latter as a remnant of aptitude for L1 learning (Skehan & Ducroquet, 1988), not dissimilar to Carroll’s (1973) speculation that aptitude might be the residue of an L1 learning ability, though for Carroll this ability fades at different rates in different people. Similarly, according to The Linguistic Coding Difference Hypothesis (LCDH) (Sparks, 2022; Sparks & Ganschow, 1991), learners’ L1 skills are related to, and consistent with, their aptitude for L2 learning, and research in this line has shown L1 proficiency as a significant predictor of L2 proficiency years later, with L2 aptitude as the best predictor (e.g., Sparks et al., 2009).

2  English at the End of Primary School: Explanatory Factors 

33

Not many studies have focused on primary school learners’ L2 aptitude (e.g., Muñoz, 2014; Suárez & Muñoz, 2011). A recent exception is the study by Roehr-Brackin and Tellier (2019) with learners age 8–9, where they found that aptitude and metalinguistic awareness – and particularly, language analytic ability at the core of both constructs – significantly predicted children’s achievement in the L2. In another study with learners age 8–11 in a similar classroom context, Kasprowicz et al. (2019) found that aptitude (grammatical sensitivity and deductive and inductive analytic abilities) was a significant predictor of morphosyntactic development under explicit instruction. Another recent study investigated the impact of individual difference variables and environmental factors in 10–12-year-old children over two years (Berthele & Udry, 2021), paying particular attention to language aptitude. Aptitude measures for language analysis (grammatical sensitivity and inductive ability) were predictive of achievement in L2 English and grammatical sensitivity was also predictive of L1 German proficiency (Schneider, 2021). The present study can shed more light on the role of language learning aptitude in young learners and its relationship with other variables, particularly with L1 skills. Evidence of linguistic interdependence between our bilingual participants’ two first languages and the L2 may yield rich insights into the multiple influences that play a role in children’s language achievement. Another learner factor that will be considered in the present analyses is gender. Gender differences in linguistic and non-linguistic outcomes have often been found in applied linguistics research (e.g., Cadierno et al., 2022; Courtney et al., 2017; Heining-Boynton & Haitema, 2007; Jaekel et al., 2017; MacIntyre et al., 2002; Muñoz et al., 2018; Sundqvist & Wikström, 2015; Sylvén & Sundqvist, 2012), and they have been ascribed to various factors such as motivation and attitudes, language learning aptitude or, more recently, to differences in time spent interacting with the target language out of school (i.e., gaming). But research results have been inconsistent, which may indicate a likely dependence on the tasks used and the dimensions assessed in each study. The attested inconsistent results also highlight the situational nature of those differences and the influence of educational and cultural factors. The influence of context on learner outcomes is investigated in the present study by means of several factors. The first factor is the family

34 

E. Tragant and C. Muñoz

socio-cultural status, measured by means of parents’ educational level. Previous research has indicated that families’ socio-cultural status are more strongly related to children’s achievement than their economic status (Caldas & Bankston, 1997). Studies on FL learning have widely confirmed the influence of parents’ educational level on children’s achievement and motivation (e.g., Butler & Le, 2018; Iwaniec, 2018); in particular, maternal education has been found a significant predictor in learners’ language development (e.g., Bornstein & Bradley, 2003). Most relevant to the current chapter is a study from the ELLiE project with a sample of 865 children from seven European countries in their fourth year of FL instruction (see Chap. 1). In that study, parents’ educational level was a strong predictor of reading comprehension scores (Lindgren & Muñoz, 2013). Likewise, and although proficiency was only self-reported, a large recent survey in Catalonia uncovered a much lower percentage of fifth graders considering their English level deficient when they came from families with university-level education than when they came from families with the lowest educational level (25.2% vs. 43.5%) (Síndic de Greuges, 2022). Similarly, in another large survey in the Region of Madrid, ninth graders from lower socioeconomic backgrounds reported lower levels of self-confidence in English than students from higher socioeconomic backgrounds (Shepherd & Ainsworth, 2017). Other contextual factors that may explain differences in linguistic outcomes in our study involve learners’ use of and exposure to English. One of these factors is the use of English in conversations (with native or non-­ native speakers of English), that is, the amount of real interaction in the target language out of the classroom. Another factor is the amount of out-of-school exposure to English (mainly through media). This variable was shown to have the greatest influence, together with language distance, on the listening and reading scores in the large cross-country study mentioned above (Lindgren & Muñoz, 2013). Another input-related variable included in our study is learners’ extracurricular English lessons (in private language schools or at home). This factor is of relevance here because a very large proportion of school pupils in Catalonia, as in the whole of Spain, enrol in private English tuition (de la Rica & González de San Roman, 2013). Because of the heterogeneous and “unofficial” nature of such English teaching practice, its quantitative effects on

2  English at the End of Primary School: Explanatory Factors 

35

primary school learners’ actual achievement have been difficult to observe, although positive effects have been reported in secondary school students (Síndic de Greuges, 2022) (see Chap. 5). Furthermore, such effects are hard to separate from the effects of parents’ educational level since children of parents with higher educational levels tend to take extracurricular English lessons more frequently than children of parents with lower levels of education. For example, the above-mentioned survey in Catalonia exemplified this association with a sample of ninth graders: the percentage of students from families with the highest educational level attending private extracurricular English lessons was three times as much as the percentage of students from families with the lowest educational level (44.5% vs 16.4%) (Síndic de Greuges, 2022). Finally, another potentially influential contextual variable is school. In the present study we did not seek to find differences among schools per se, but we included this factor in the analyses (as a random factor) because learners belonged to five different schools. All the schools in the study are state-funded, which guarantees a widely mixed family influx in terms of socioeconomic status (SES) and parents’ education in this region, but schools still differ in those respects as a function of location in different neighbourhoods. Likewise, although state-funded schools cannot freely hire their own teaching staff, teachers show high variation in terms of linguistic and methodological expertise and teaching approach (see Chap. 1). The study in the present chapter was guided by two research questions. The first asks how the different learner-internal factors (affective and cognitive) and contextual factors are interrelated, with the assumption that they are not totally independent of one another. The second and main research question strives to identify the variables that have stronger effects on different dimensions of language achievement at the end of primary school. The two research questions are formulated as follows: RQ1: In what ways do the different learner-internal and contextual factors associate with each other? RQ2: What learner-internal and contextual factors contribute the most to young learners’ English language achievement at the end of primary school?

36 

E. Tragant and C. Muñoz

2.3 Method 2.3.1 Participants: ELLiC Sample The sixth graders’ total sample was composed of 116 learners, 54 of them female and 62 males, age 11–12 years. However, not all pupils completed all the tests in grade 6 (hereafter G6) and neither did all the families return the take-home questionnaire in G4 (some children may have enrolled later in the school) or answer all the questions. Thus, the sample size varied in the different analyses based on those variables. The learners came from six intact classes in five state-funded schools. They were all Catalan-Spanish bilinguals, and these two languages are considered first languages (L1s) in the present study (though language dominance may vary in different individuals). The children’s families came from a variety of socioeconomic classes, from blue-collar workers to professionals. As seen in Chap. 1, learners in four of these schools were introduced to English in G1 and in one school they started English classes in pre-school (at the age of 5). Because of the organisation of primary education into three cycles of two years each, English teachers tend to stay with the same class for the two years of a cycle, so that children from each school had commonly been taught by three different teachers over the six years of primary education. Table 2.1 displays the descriptive statistics of the sixth graders’ total sample by gender and school. Table 2.1  Descriptive statistics by gender and school School A School B School C School D School E Total

Females

Males

Total

8 9 10 9 18 54

16 9 10 14 13 62

24 18 20 23 31 116

2  English at the End of Primary School: Explanatory Factors 

37

2.3.2 Instruments and Procedure Language Tests The listening comprehension test administered in G6 was an adaptation of a test that had been piloted and used in a previous study (the BAF project; see Muñoz, 2006a) with similar learners. The adapted test consisted of 17 items in increasing order of difficulty that had also been satisfactorily used in G5 with the same learners. It was a picture selection task in which learners were required to match the word or utterance they heard from an audiotape to one picture in a set of three pictures. The first item was a  vocabulary recognition item while the remaining 16 items consisted of full sentences, easier sentences at the beginning of the task and more complex sentences at the end. The language used in the stimuli was simple and referred to topics common in the school activities and daily lives of the learners. Learners were given one point for each correct answer. The reading comprehension test administered in G6 had a cartoon format with eight speech bubbles. The same test had been used in G5, so they were familiar with the test format. Learners were asked to read a number of short expressions in English to themselves and decide which should be inserted in each of the empty bubbles to successfully complete the storyline. Then learners had to decide whether 15 sentences in English about the story were true or false. The maximum score was 23. The task was not timed but learners completed it in less than 15 minutes. Learners’ general English proficiency was measured by means of an English proficiency test that was administered in G6 at the end of primary education by the regional educational authorities. The test included three components: listening comprehension (14 multiple choice items), reading comprehension (24 multiple choice items) and writing (between 30–45 words). The sections on listening and reading comprehension were based on two texts each and questions measured the ability to extract and interpret information. The evaluation of students’ writing involved both linguistic and discourse competence. The maximum score was 100. The time allotted for this test was 50 minutes.

38 

E. Tragant and C. Muñoz

Last, the writing task, administered in G6, consisted of a personal letter to a penfriend in another country, which had also been completed in G5. Children were instructed to write about themselves (family, friends, town, hobbies, favourite music, school) and ask questions about their penfriends. Learners were given a sheet of paper with 10 lines as a guide regarding length. They were told to write as much as they could, to make themselves understood, and not to worry too much about errors. The time allotted for this task was 10 minutes, but most learners completed the task in a shorter time. The handwritten productions of the learners were transcribed and coded in CHAT format (MacWhinney, 2000) by a research assistant, and one of the researchers coded 10% of the transcripts and resolved any discrepancies. Transcripts were analysed by means of three different measures: fluency, lexical diversity, and syntactic complexity. The index of written fluency chosen was words per text (W/Tx), that is, the total number of tokens (L1 words and proper names excluded). Although the task was not strictly timed because learners were given ample time to complete it, this measure has been found to give a good approximation of learners’ written fluency (Torras et  al., 2006; Wolfe-Quintero et  al., 1998). The index of lexical diversity selected was Guiraud’s Index (GI), which is an improvement on the type/token ratio that takes into account text length (Treffers-Daller et  al., 2018). For syntactic complexity we used the ratio of clauses per sentence (Cl/S). This measure is widely used and considered to be an index of overall complexity; it has also proved to successfully reflect the emergence of complexity in the interlanguage of beginners (Torras et al., 2006). See Chap. 3 for a more extended description of these measures.

Explanatory Factors An L2 motivation/attitudes questionnaire was administered in G6 to the intact classes to gauge learners’ motivation and attitudes towards learning English (hereafter Affect). It asked learners whether they liked learning English words, listening to English words, singing in English, whether they found English easy or difficult, etc. It had been used every year since

2  English at the End of Primary School: Explanatory Factors 

39

G1 and had been adapted to children’s increasing age (after performing item internal consistency tests). In G6 the questionnaire had 7 items with a 5-point scale (represented by smiley face emoticons). The maximum score was 35. The task was not timed but learners completed it in less than 5 minutes. The MLAT-E (elementary version of the Modern Language Aptitude Test) in its Catalan adaptation (Suárez, 2010) was administered to intact classes in G6 to gauge learners’ language learning aptitude. The MLAT-E is a simplified version that was adapted to children age 8–12 (Carroll & Sapon, 1967). Like the MLAT, it consists of four parts: Hidden Words, Matching Words, Finding Rhymes, and Number Learning. Hidden Words corresponds to Spelling Clues of the MLAT, and it aims at measuring knowledge of vocabulary, as well as sound-symbol association ability. Matching Words corresponds to Words in Sentences in the MLAT and measures grammatical sensitivity. Finding Rhymes, a subtest that is not in the MLAT, attempts to measure the ability to hear speech sounds by asking the test taker to select words that rhyme. Finally, Number Learning, as in the MLAT, aims at measuring the memory component (rote memory) by asking the examinee to learn the names of numbers in an artificial language. Suárez (2010) translated and adapted the MLAT-ES (the Spanish version for children developed by Stansfield and Reed in 2005) into Catalan to elaborate the Catalan version (MLAT-EC) that was used in this study. In the MLAT-EC, the maximum possible score of the Hidden Words and Matching Words subtests is 30, that of the Finding Rhymes subtest is 38 and of Number Learning is 25. The maximum total score is 123. The Cronbach alpha of MLAT-EC was 0.735. The time allotted for this test was 50 minutes. A Catalan language test and a Spanish language test were used to measure learners’ L1 literacy skills. These tests were administered and evaluated by the regional educational authorities in G6, together with the English language test above. Both tests had the same format: a word dictation, a reading comprehension task (mostly multiple choice items as well as open responses) elicited by two texts (a narrative and an informative text), and a composition. The evaluation focused on multiple aspects: gathering information and text interpretation for reading comprehension, lexis and morphosyntax, spelling and formal aspects,

40 

E. Tragant and C. Muñoz

and coherence and adequacy for writing. The time allotted for each test was 50 minutes. A take-home questionnaire gathered information from parents in G4 about type and amount of support for FL learning the learners generally had at home and out of school. Five of the questions provided data for the present analysis. Mother and father were asked to report the highest level of education they had completed. Parents were also asked to provide information about their children’s amount of  spoken interaction with English speakers (at home or abroad), and more specifically the number of occasions per year when the children used English to interact with somebody “who did not speak Spanish or Catalan but did speak English”. Another exposure-related question asked for the amount of time the children spent on five different activities in a typical school week. The activities were: watching English films, cartoons and/or series on TV or the Internet; playing English video/computer games; listening to English music; reading English books, magazines, comics, etc.; and speaking English with someone. Finally, information about learners’ extracurricular tuition (hours per week) was captured from the questionnaire up to G4. This information was also gathered from the teachers and from the students themselves every year. The present study used the total accumulated amount of extracurricular hours in primary education as an explanatory factor. Table 2.2 displays the explanatory variables and the times when data were collected. Table 2.2  Explanatory factors and data collection times Learner variables

Contextual variables

Gender Affect G6 Language aptitude G6 L1s skills G6

School Parents’ educational level G4 Spoken interaction in English G4 Exposure to English G4 Amount of extracurricular English lessons G1–G6

2  English at the End of Primary School: Explanatory Factors 

41

2.4 Analysis and Results In this section we first present the descriptive statistics of the variables in this study. Then we present a series of correlational analyses that were conducted to address the first research question, concerned with the associations among the explanatory variables. This is followed by the correlational analysis of the outcome variables and the inferential tests that were conducted to answer the second research question aiming at discovering the strongest predictors of English language achievement at the end of G6. We used the statistical software programme SPSS (version 27).

2.4.1 Variables Description Descriptive statistics were calculated for the explanatory variables: motivation and attitudes towards English (Affect), language learning aptitude (Aptitude), Catalan language test (Catalan), Spanish language test (Spanish), yearly interaction with English speakers (Interaction), weekly exposure to English outside of school (Exposure), amount of hours of extracurricular lessons (Extra), mother’s educational level (M education), and father’s educational level (F education). Descriptive statistics were also calculated for the outcome variables: listening comprehension scores (Listen), reading comprehension scores (Read), English general proficiency (Proficiency), written fluency as measured by words per text (W/Tx), written lexical diversity as measured by the Guiraud’s Index (GI), and written syntactic complexity as measured by clauses per sentence (Cl/S).1 Table 2.3 displays descriptive statistics for the continuous variables. The variable of interaction with English speakers was not linearly distributed and was re-categorised into three levels (Table 2.4). Table 2.5 displays descriptive statistics for parents’ education; only 3 parents had doctoral degrees, so they were merged with the parents holding bachelor degrees. Data about parents’ education was obtained from approximately 80% of children. Table 2.6 displays descriptive statistics for the L2 outcome variables.  The shortened variable labels shown above are used in the tables in this section.

1

42 

E. Tragant and C. Muñoz

Table 2.3  Descriptive statistics of explanatory variables Affect (n = 116) Aptitude (n = 111) Catalan (n = 108) Spanish (n = 110) Extra (n = 113) Exposure (n = 95)

Mean(SD)

Median

Min-Max

CI

25.09(4.55) 101.68(22.52) 84.48(12.39) 79.15(12.67) 1.99(2.65) 6.12(2.84)

25.00 109.00 86.78 81.22 1.00 6

10–33 12–122 35.95–100 47.85–100 0–10 0–16

24.26–25.93 97.45–105.92 82.12–86.84 76.76–81.55 1.50–2.48 5.54–6.69

Table 2.4 Descriptive statistics of interaction with English speakers. Yearly frequencies and percentages

Table 2.5  Descriptive statistics of parents’ educational level. Frequencies and percentages

None: 0 times Low: 1–15 times High: 50+ times Total

97 (83.6%) 14 (12.1%) 5 (4.3%) 116

M education F education Primary 5 (5.4%) Secondary 37 (39.8%) Tertiary 51 (54.8%) Total 93

11 (12.1%) 43 (47.3%) 37 (40.6%) 91

Table 2.6  Descriptive statistics of L2 outcome variables Listen (n = 116) Read (n = 116) Proficiency (n = 103) W/Tx (n = 116) GI (n = 116) Cl/S (n = 116)

Mean(SD)

Median

Min-Max

CI

12.34(2.66) 15.53(3.98) 80.67(13.41) 68.01(23.09) 4.55(0.70) 1.38(0.45)

13.00 16.00 81.99 63.00 4.54 1.25

5–17 5–23 36.02–100 16–156 2.69–6.45 0.70–3.29

11.86–12.83 14.79–16.26 78.05–83.29 63.76–72.25 4.42–4.67 1.30–1.47

2.4.2 Correlational Analysis of Explanatory Variables A series of (two-tailed) Pearson product-moment tests were run among the study variables to discover associations among them. Those correlations that were statistically significant2 are presented below in meaningful  Bivariate correlations are interpreted based on statistical significance (p