Loading...
Title: Evaluating Bilingual Education in Germany

Evaluating Bilingual Education in Germany

CLIL Students’ General English Proficiency, EFL Self-Concept and Interest

by Dominik Rumlich (Author)
©2016 Thesis 582 Pages
English Studies
Series: Mehrsprachigkeit in Schule und Unterricht, Volume 15

Summary

The author uses a theoretical account rooted in TEFL, language acquisition and educational psychology to provide the basis for the development of a comprehensive model of language learning in CLIL. It incorporates prior knowledge, EFL self-concept, interest in EFL classes, verbal cognitive abilities and contact to English. This model is used to estimate the effects of CLIL in the context of high-intensity programmes at German Gymnasien. The statistical evaluation of the quasi-experimental data from 1,000 learners proves the existence of large initial differences due to selection, preparation and class composition effects. After two years, one finds no significant effects of CLIL apart from a minor increase in self-concept, suggesting that the actual effects of CLIL have often been overestimated.

Table Of Contents

  • Cover
  • Title
  • Copyright
  • About the author
  • About the book
  • This eBook can be cited
  • Table of contents
  • List of figures
  • List of tables
  • Abbreviations and explanations of statistical and educational concepts
  • 1. Introduction
  • 2. Conceptual and institutional background of CLIL (in Germany)
  • 2.1 Definition of CLIL and its general characteristics
  • 2.2 The educational context: How CLIL in Germany became what it is today
  • 2.3 CLIL in NRW: The concrete educational context
  • 2.3.1 The development of CLIL provision and participation
  • 2.3.2 The curricular implementation and organisation of CLIL strands in NRW
  • 2.3.3 The characteristics of the CLIL programme to be evaluated in DENOCS
  • 2.4 Selection processes in NRW CLIL strands: The origin of creaming effects
  • 2.5 Students’ preparation for NRW CLIL strands
  • 2.6 Summary and conclusion
  • 3. Learning in EFL CLIL classes/strands: Theoretical background
  • 3.1 Language learning
  • 3.1.1 The connection between CLIL, communicative language teaching and constructivism
  • 3.1.2 Constructivist language learning: Interrelation of content and language
  • 3.1.3 The fulfilment of innatist demands in CLIL
  • 3.1.4 Conscious language learning processes, output orientation and interaction in CLIL
  • 3.1.5 CLIL as a bridge between theories of language learning in conjunction with neuroscientific claims
  • 3.1.6 Authenticity, learning strategies and learner autonomy as essential assets to language learning in CLIL
  • 3.2 Affective-motivational dispositions I: Academic self-concept (ASC)
  • 3.2.1 The importance of the self: Views of the (language) learning sciences
  • 3.2.2 The general importance of (academic) self-concept
  • 3.2.3 Definition of ASC and differentiation from related constructs
  • 3.2.4 The facilitative influence of ASC
  • 3.2.5 Relationships between ASC and achievement: A reciprocal effects model
  • 3.2.6 The structure of ASC
  • 3.2.7 The internal/external frame of reference (I/E) model and its extension to the reciprocal I/E model
  • 3.2.8 Processes related to external frames of reference: Big-fish-little-pond (BFLP), basking-in-reflected-glory (BIRG), Pygmalion and Galatea effects
  • 3.2.9 EFL SC in the light of students’ selection and preparation for CLIL classes
  • 3.2.10 Theoretical considerations on the potential impact of CLIL on EFL SC
  • 3.3 Affective-motivational dispositions II: Interest in EFl classes
  • 3.3.1 The construct: Definition, structure and effects of interest
  • 3.3.2 The development of interest: Mechanisms explaining its decline and changes in the direction of causation between interest and achievement
  • 3.3.3 The influence of CLIL on students’ interest in EFL classes
  • 3.4 Voluntary out-of-school EFL learning: Spare-time/leisure English
  • 3.5 Learning-related reasons for student selection and preparation: Cognitive abilities, prior knowledge, affective-motivational dispositions and thresholds
  • 3.6 Further influences on language learning and achievement
  • 3.6.1 Verbal cognitive abilities (VCA) as an indicator of verbal intelligence
  • 3.6.2 Sex
  • 3.6.3 L1 influences
  • 3.7 Summary and conclusion
  • 4. Learning in EFL CLIL classes/strands: Empirical findings from Germany
  • 4.1 Maintaining a critical stance in the light of a conglomerate of issues
  • 4.2 Overview of the nine largest studies on English CLIL streams at Gymnasien
  • 4.3 Language proficiency
  • 4.3.1 Cross-sectional comparisons: Are CLIL students at Gymnasien better than their peers at schools without CLIL streams?
  • 4.3.2 Student selection and preparation
  • 4.3.3 Longitudinal studies: What difference does CLIL actually make?
  • 4.4 EFL SC
  • 4.5 Interest in EFL classes
  • 4.6 Voluntary out-of-school learning: Leisure/spare-time English
  • 4.7 Further influences on learning and achievement
  • 4.7.1 Verbal cognitive abilities (VCA)
  • 4.7.2 Sex
  • 4.7.3 L1 influences
  • 4.8 Summary
  • 5. Research void, study design, (statistical) methods and instruments
  • 5.1 Current academic voids in CLIL research and measures to address them
  • 5.1.1 RQ I: A priori differences (selection, preparation and class composition)
  • 5.1.2 RQ II: Effects of CLIL/non-CLIL environments on general EFL proficiency
  • 5.1.3 RQ III: Effects of CLIL/non-CLIL environments on EFL SC/EFL interest
  • 5.2 Study design and data collection
  • 5.2.1 Entire sample
  • 5.2.2 Effective sample size and composition for present investigations
  • 5.2.3 Non-participation/missings
  • 5.3 Software, estimators, and missing values
  • 5.4 Instruments
  • 5.4.1 General EFL proficiency: C-tests
  • 5.4.1.1 Choice of C-tests, administration and scoring
  • 5.4.1.2 Statistical properties of the C-tests
  • 5.4.2 Scale on EFL SC (academic self-concept)
  • 5.4.2.1 Factor analyses of the EFL SC scale
  • 5.4.2.2 Longitudinal factorial invariance
  • 5.4.2.3 Statistical properties of the overall scale
  • 5.4.3 Scale on interest in EFL classes (subject-related interest)
  • 5.4.3.1 Factor analyses
  • 5.4.3.2 Longitudinal factorial invariance
  • 5.4.3.3 Statistical properties of the overall scale
  • 5.4.4 Spare-time/leisure English
  • 5.4.5 Verbal cognitive abilities (in German)
  • 5.4.6 Individual items for demographic/constitutional data
  • 5.5 Summary
  • 6. A priori differences between CLIL, non-CLIL and regular students due to selection, preparation, and class composition (RQ I; year 6)
  • 6.1 Development of an integrative model
  • 6.2 Demographic/constitutional variables: Age, sex, L1 background
  • 6.2.1 Descriptive and general inferential statistics
  • 6.2.2 Discussion
  • 6.3 Verbal cognitive abilities (in German)
  • 6.3.1 Descriptive and general inferential statistics
  • 6.3.2 Exploratory SEM analyses
  • 6.3.3 Discussion
  • 6.4 Leisure English
  • 6.4.1 Descriptive and general inferential statistics
  • 6.4.2 SEM analyses
  • 6.4.3 Discussion
  • 6.5 EFL SC
  • 6.5.1 Descriptive and general inferential statistics
  • 6.5.2 SEM analyses
  • 6.5.3 Discussion
  • 6.6 Interest in EFL classes
  • 6.6.1 Descriptive and general inferential statistics
  • 6.6.2 SEM analyses
  • 6.6.3 Discussion
  • 6.7 General EFL proficiency
  • 6.7.1 Descriptive and general inferential statistics
  • 6.7.2 SEM analyses
  • 6.7.3 Discussion
  • 6.8 Summary
  • 6.8.1 Overall SEM results
  • 6.8.2 Answers to RQ I
  • 7. The effects of CLIL and non-CLIL environments on general EFL proficiency, EFL SC, and interest in EFL classes (RQ II/III; year 8)
  • 7.1 Development of a longitudinal model
  • 7.2 Demographic variables (age, sex, L1) and verbal cognitive abilities
  • 7.3 Leisure English
  • 7.3.1 Descriptive and general inferential statistics on the year-eight data
  • 7.3.2 Longitudinal changes from year 6 to year 8
  • 7.3.3 SEM analyses
  • 7.3.4 Discussion
  • 7.4 General EFL proficiency
  • 7.4.1 Descriptive and general inferential statistics on the year-eight data
  • 7.4.2 Longitudinal changes from year 6 to year 8
  • 7.4.3 SEM analyses
  • 7.4.4 Discussion
  • 7.5 EFL SC
  • 7.5.1 Descriptive and general inferential statistics on the year-eight data
  • 7.5.2 Longitudinal changes
  • 7.5.3 SEM analyses
  • 7.5.4 Discussion
  • 7.6 Interest in EFL classes
  • 7.6.1 Descriptive and general inferential statistics on the year-eight data
  • 7.6.2 Longitudinal changes
  • 7.6.3 SEM analyses
  • 7.6.4 Discussion
  • 7.7 Summary
  • 7.7.1 Overall SEM results
  • 7.7.2 The interrelatedness of study designs, statistical approaches and results
  • 7.7.3 Profiles of low- and high-proficiency CLIL students
  • 7.7.4 Answers to RQ II and RQ III
  • 8. Final evaluation and future perspectives
  • 8.1 General summary
  • 8.2 Limitations, further analyses and future research
  • 8.3 Implications and conclusion
  • References
  • Appendix
  • A Additional (statistical) details on the year-six data
  • A.1 Alternative models of general EFL proficiency: Changes in strengths and patterns of influences when omitting central variables
  • A.2 Comparison of results obtained from SEM, path and regression models
  • B Additional (statistical) details on the year-eight data
  • Series index

List of figures

Figure 1:     The location of content-based language teaching approaches on a continuum from content-driven to language-driven

Figure 2:     Absolute and relative number of schools, in NRW according to type of school, form of CLIL provision and language of instruction

Figure 3:     The development of absolute CLIL provision (continuous forms) and participants at Gymnasien, Realschulen and Gesamtschulen in NRW from 2005–2014

Figure 4:     The development of English CLIL groups and CLIL lessons taught (continuous forms) in NRW at Gymnasien, Realschulen and Gesamtschulen

Figure 5:     The development of relative English CLIL participation (continuous forms) in NRW at Gymnasien, Realschulen and Gesamtschulen

Figure 6:     The development of absolute English CLIL provision (continuous forms) and student participation in NRW according to type of school

Figure 7:     The development of relative student participation in continuous forms of CLIL (in % of total student population in years 7–12/13) in NRW from 2005 until 2014

Figure 8:     CLIL in the NRW school system: Its implementation and no. of lessons

Figure 9:     Relations of achievement and academic self-concept within and across the general verbal and mathematics domain as predicted by the internal/external frames of reference model

Figure 10:   Appraising the quality of (empirical) research: Schematic illustration of the determinants of best practice/excellence

Figure 11:   Exam grades of CLIL and non-CLIL students who participated in the ministerial evaluation

Figure 12:   Location of the schools in North-Rhine Westphalia

Figure 13:   Composition of the entire sample (N = 1,398)

Figure 14:   Composition of the effective sample (N = 953) used in the thesis at hand

Figure 15:   A basic model of general EFL proficiency, EFL SC and interest in EFL classes

Figure 16:   An extended model of general EFL proficiency, EFL SC and interest in EFL classes including leisure English ← 17 | 18 →

Figure 17:   An extended model of general EFL proficiency, EFL SC and interest in EFL classes including indirect influences

Figure 18:   A model of general EFL proficiency, EFL SC, interest in EFL classes, the effects of CLIL and non-CLIL environments

Figure 19:   Hypothesised relationships among outcome variables including the influence of streaming/preparation

Figure 20:   Measurement model of EFL SC (see p. 278 for items) and interest in EFL classes in year 6 (see p. 289 for items)

Figure 21:   Complete model in year 6

Figure 22:   Main and indirect effects involving KFT sum scores

Figure 23:   Relative frequencies of Rasch EFL proficiency scores according to groups in year 6

Figure 24:   SEM paths with statistically significant β coefficients in year 6

Figure 25:   Complete model in year 8

Figure 26:   Relative frequencies of longitudinal gains in Rasch general EFL proficiency scores according to groups

Figure 27:   Average Rasch general EFL proficiency scores in years 6, 7 and 8 according to groups

Figure 28:   SEM paths with statistically significant β coefficients in year 8

Figure 29:   Illustration of significant main effects (p < .05) and relationships among variables

Figure 30:   Model 1 with group affiliation (CLIL/non-CLIL class) as the only predictors

Figure 31:   Model 2 with three to four predictors per outcome variable

Figure 32:   Model 3 with four to five predictors per outcome variable

Figure 33:   Graphic illustration of model 4

Figure 34:   Paths of SEM with statistically significant β coefficients in year 6 for the entire sample

Figure 35:   Paths of SEM with β coefficients in year 6 for the entire sample including indirect effects

Figure 36:   Measurement model of EFL SC (see p. 278 for items) and interest in EFL classes in year 8 (see p. 289 for items)

Figure 37:   Longitudinal measurement model of EFL SC (see p. 278 for items) and interest in EFL classes (see p. 289 for items)

Figure 38:   Model 1 with CLIL environment and non-CLIL class as the only predictors

Figure 39:   Model 2 with prior measures in addition to CLIL environment and non-CLIL class as predictor ← 18 | 19 →

Figure 40:   Model 3 includes the main predictor of the respective outcome variables in addition to prior measures, CLIL environment and non-CLIL class as predictors

Figure 41:   Model 5 equals the cross-sectional model of year 6 (omission of prior measures)

Figure 42:   Complete model in year 8

Figure 43:   Paths of SEM with statistically significant β coefficients in year 8 ← 19 | 20 → ← 20 | 21 →

List of tables

Table 1:      Systematic overview of the CLIL programme to be evaluated and alternative configurations

Table 2:      Selection processes at Gymnasien in NRW: Changes in student numbers in relation to the starting cohort that moved to Gymnasien after primary school at the beginning of the school year 2006/2007

Table 3:      Overview of students’ spare-time English; percentage of students at respective types of school who get in contact with or do activities involving English at least several times a month (year 9, ~15 years of age)

Table 4:      The nine largest studies on EFL CLIL streams at Gymnasien (by CLIL sample size)

Table 5:      Differences in average general EFL proficiency at the end of year 6 between CLIL and non-CLIL students taught in the same English class

Table 6:      Overview of Bredenbröker’s longitudinal results

Table 7:      EFL achievement at the beginning of year 8

Table 8:      Overview of the constructs incorporated in the student questionnaire

Table 9:      Student participation in years 6 and 8 according to groups

Table 10:    Comparison of key outcome variables (yr 8) according to groups and status of missingness in yr 6

Table 11:    Comparison of key outcome variables (yr 6) according to groups and status of missingness in yr 8

Table 12:    Demographic/constitutional properties of the sample and missings according to years

Table 13:    Descriptive measurement properties of C-test batteries according to years

Table 14:    Measurement properties of C-test #A (anchor test; 24 items)

Table 15:    Collection of items for the measurement of EFL SC

Table 16:    Statistical properties of the EFL SC scale in years 6 and 8

Table 17:    Results of the tests of factorial invariance of the eight-item EFL SC scale

Table 18:    Descriptive statistics of the EFL SC scale in years 6 and 8

Table 19:    Collection of items for the measurement of students’ subject-related interest ← 21 | 22 →

Table 20:    Statistical properties of the scale “interest in EFL classes” in years 6 and 8

Table 21:    Results of the tests of factorial invariance of the seven-item scale on interest in EFL classes

Table 22:    Descriptive statistics of the interest scale in years 6 and 8

Table 23:    Descriptive statistics of KFT-scores in year 7

Table 24:    Pearson correlations among outcome variables and types of leisure English in the overall sample from year 6

Table 25:    Pearson correlations among outcome variables and types of leisure English in the overall sample from year 6 (continued)

Table 26:    Demographic variables according to groups in year 6

Table 27:    Point-biserial correlations among demographic predictor variables in the overall sample from year 6

Table 28:    Descriptive statistics of students’ KFT sum scores according to groups in year 7

Table 29:    Percentage of KFT sum scores equal to or above median of overall sample according to groups

Table 30:    Descriptive statistics of students’ leisure English sum scores according to groups in year 6

Table 31:    Percentage of students’ leisure English sum scores equal to or above median of overall sample according to groups

Table 32:    β coefficients for predictors of leisure English scores in incrementally more complex SEMs

Table 33:    Descriptive statistics of students’ mean EFL SC scores according to groups in year 6

Table 34:    Percentage of mean EFL SC scores equal to or above median of the overall sample according to groups

Table 35:    β coefficients for predictors of EFL SC in incrementally more complex SEMs

Table 36:    Descriptive statistics of students’ mean interest (in EFL classes) scores according to groups in year 6

Table 37:    Percentage of mean interest (in EFL classes) scores equal to or above median of overall sample according to groups

Table 38:    β coefficients for predictors of interest in EFL classes in incrementally more complex SEMs

Table 39:    Descriptive statistics of Rasch EFL proficiency scores according to groups in year 6

Table 40:    Distribution of Rasch EFL proficiency scores according to groups in year 6 after a median split of the overall sample ← 22 | 23 →

Table 41:    β coefficients for predictors of Rasch general EFL proficiency scores in incrementally more complex SEMs

Table 42:    Unstandardised B coefficients for predictors of EFL proficiency with corresponding advantage of future CLIL students expressed in Rasch general EFL proficiency scores

Table 43:    Unstandardised B coefficients for predictors of EFL proficiency with corresponding advantage of non-CLIL students expressed in Rasch general EFL proficiency scores

Table 44:    Overview of direct comparisons without covariates between subgroups in year 6

Table 45:    Demographic variables according to groups

Table 46:    Descriptive statistics of students’ leisure English sum scores according to groups in year 8

Table 47:    Percentage of students’ leisure English sum scores equal to or above median of overall sample according to groups

Table 48:    Results of Wilcoxon signed-rank tests for longitudinal changes in leisure English scores from years 6 to 8

Table 49:    β coefficients for predictors of leisure English scores in incrementally more complex SEMs

Table 50:    β coefficients for predictors of leisure English scores from SEMs in years 6 and 8

Table 51:    Descriptive statistics of students’ Rasch general EFL proficiency scores according to groups in year 8

Table 52:    Percentage of students’ general EFL proficiency scores above the median of the overall sample according to groups

Table 53:    Dependent-samples t tests for longitudinal changes in general EFL proficiency scores from years 6 to 8

Table 54:    β coefficients for predictors of Rasch general EFL proficiency scores in incrementally more complex SEMs

Table 55:    Unstandardised B coefficients for predictors of EFL proficiency with corresponding advantage of CLIL students expressed in Rasch general EFL proficiency scores

Table 56:    Unstandardised B coefficients for predictors of EFL proficiency with corresponding disadvantage of non-CLIL students expressed in Rasch general EFL proficiency scores

Table 57:    β coefficients for predictors of general EFL proficiency scores from SEMs in years 6 and 8

Table 58:    Descriptive statistics of students’ EFL SC scores according to groups in year 8 ← 23 | 24 →

Table 59:    Dependent-samples t tests for longitudinal changes in EFL SC scores from years 6 to 8

Table 60:    β coefficients for predictors of EFL SC scores in incrementally more complex SEMs

Table 61:    Unstandardised B coefficients for predictors of EFL SC scores with corresponding advantage of CLIL students expressed in EFL SC scores

Table 62:    β coefficients for predictors of EFL SC scores from SEMs in years 6 and 8

Table 63:    Descriptive statistics of students’ interest scores according to groups in year 8

Table 64:    Dependent-samples t tests for longitudinal changes in interest scores from years 6 to 8

Table 65:    β coefficients for predictors of interest in EFL classes scores in incrementally more complex SEMs

Table 66:    Unstandardised B coefficients for predictors of interest in EFL classes scores with corresponding advantage of CLIL students expressed in interest scores

Table 67:    β coefficients for predictors of interest in EFL classes scores from SEMs in years 6 and 8

Table 68:    Comparison of SEM and regression model estimates of β coefficients and their p values in year 8 with varying study designs

Table 69:    Profiles of low- and high-proficiency CLIL students

Table 70:    Profiles of low-proficiency CLIL students and matched regular students

Table 71:    Regression analysis of low-proficiency CLIL and matched medium-proficiency regular students with general EFL proficiency in year 8 as dependent variable

Table 72:    Overview of direct comparisons without covariates between subgroups in year 8

Table 73:    Overview of significance tests for longitudinal changes (from year 6 to 8) in outcome variables according to groups

Table 74:    The typical/average CLIL, regular and non-CLIL student in years 6 and 8

Table 75:    The typical/average CLIL, regular and non-CLIL student in years 6 and 8

Table 76:    Overview of direct comparisons without covariates between subgroups in years 6 ← 24 | 25 →

Table 77:    SEM β coefficients (STDYX standardisation) and p values for main effects and indirect/mediation effects for the entire sample and according to groups in year 6

Table 78:    SEM B coefficients (unstandardised) and p values for main effects and indirect/mediation effects for the entire sample and according to groups in year 6

Table 79:    Correlations (p values) among covariates (SEM estimates; model 4) in the overall sample from year 6

Table 80:    Descriptive statistics of students’ Rasch general EFL proficiency scores according to students’ L1 background and groups in year 6

Table 81:    Descriptive statistics of students’ EFL SC scores according to students’ L1 background and groups in year 6

Table 82:    Descriptive statistics of students’ interest in EFL classes scores according to students’ L1 background and groups in year 6

Table 83:    β coefficients for predictors in incrementally more complex SEMs

Table 84:    Predictors of Rasch EFL proficiency scores in alternative SEMs with varying combinations of affective-motivational and cognitive factors

Table 85:    β coefficients from SEM, path and regression models for the sample in year 6

Details

Pages
582
Publication Year
2016
ISBN (ePUB)
9783631694152
ISBN (MOBI)
9783631694169
ISBN (PDF)
9783653064605
ISBN (Hardcover)
9783631671290
DOI
10.3726/978-3-653-06460-5
Language
English
Publication date
2016 (July)
Published
Frankfurt am Main, Berlin, Bern, Bruxelles, New York, Oxford, Wien, 2016. 582 pp., 96 b/w tables, 36 b/w ill., 7 coloured ill.
Product Safety
Peter Lang Group AG

Biographical notes

Dominik Rumlich (Author)

Dominik Rumlich studied English, geography and educational science in Germany and New Zealand. His areas of expertise include CLIL, assessment, motivation, learning strategies, quantitative research, heterogeneity and individual learner characteristics.

Previous

Title: Evaluating Bilingual Education in Germany