3.1 Educational standards

In the findings of the first wave of the Programme for International Student Assessment Study (PISA) in 2001, students showed only mediocre results in German reading literacy, mathematics and the natural sciences (Baumert et al. Reference Baumert, Klieme, Neubrand, Prenzel, Schiefele, Schneider, Stanat, Tillmann and Weiss2001). This led to a continuing debate on education policy in Germany, and change is now evident in the field of educational and FL research. Identifying the best possible education for students – how their learning process can be facilitated – has become a challenge for researchers as well as teachers, students and politicians.

As a consequence, the Standing Conference of the Ministers of Education and Cultural Affairs of the Federal States in the Federal Republic of Germany (KMK, Kultusministerkonferenz) resolved to develop national educational standards (NES, Nationale Bildungsstandards) which reflect a change in perspective from input to output orientation (Schröder Reference Schröder2005: 42).

NES define the competences (in alignment with the CEFR – the Common European Framework of Reference for languages) – what students ‘can or will be able to do’ – rather than the topics and the content of what should be taught. According to Weinert (Reference Weinert and Weinert2001: 27, cited in Klieme Reference Klieme2004: 65), competences are ‘the cognitive abilities and skills possessed by or able to be learned by individuals that enable them to solve particular problems, as well as the motivational, volitional and social readiness and capacity to use the solutions successfully and responsibly in variable situations’. This means that NES ‘articulate requirements for school-based teaching and learning. They identify goals for pedagogical work, expressed as desired learning outcomes for students’ (Klieme Reference Klieme2004: 15), which function as a ‘key mechanism in the effort to secure and improve the quality of the work done in schools’ (ibid.). According to Klieme (Reference Klieme2004), empirical research is required to find out whether students meet the expected competence levels, and to ensure quality management within the German school system.

So far, NES have been designed for several subjects (including English and French as FLs) and grade levels (primary school and lower secondary education).

For English and French as FLs, the NES are based on the CEFR (Sekretariat der Ständigen Konferenz der Kultusminister der Länder in der Bundesrepublik Deutschland 2004: 6; Council of Europe 2011). The CEFR operates as a model to describe the communicative language competences of students on a scale as well as verbal descriptors from A1 and A2 (beginners), B1 and B2 (threshold) to C1 and C2 (competent speakers). One of the challenges of implementing NES can be found in the discrepancies between the descriptors for certain levels of competence from the CEFR and the tasks designed to assess students’ language competences (Finkbeiner Reference Finkbeiner2012).

The Institute for Educational Progress in Berlin (Institut zur Qualitätssicherung im Bildungswesen, IQB) is responsible for the empirical investigation of the implementation of the NES and as such is involved in the development of tasks to measure students’ competence levels. Another challenge faced by the NES is to take into account the existing subject-specific findings of FL research (Schröder Reference Schröder2005: 42).

The largest and most significant study in English as a foreign language (EFL) in Germany so far is the DESI study (Assessment of Student Achievements in German and English as a Foreign Language International German English Study), which is particularly concerned with the language competences of students. The findings of the DESI study (Beck & Klieme Reference Beck and Klieme2007; DESI-Konsortium 2008) have attracted considerable attention, which can be seen as a further consequence of the PISA results in Germany in 2000. A team of researchers investigated the language competences of 11,000 students at the beginning and end of the school year 2003–2004 (Klieme Reference Klieme2006: 1). The study was authorized by the KMK and aimed to complement the PISA study's results by identifying and modeling students’ EFL competences and achievement levels.

DESI draws on different research methods and instruments, triangulating quantitative (e.g. achievement tests) with qualitative data (e.g. video recordings) from a large-scale project to investigate a number of sub-topics, such as reading and listening comprehension, learning strategies and text production, and make an in-depth study of the subject matter itself (DESI-Konsortium 2008). Several of the DESI results will be covered in this article. In this section, we will consider the foundation they laid for educational standards.

In his explorative study, Zydatiß (Reference Zydatiß2005: 26) wanted to investigate the interdependence between the C-test as a measure for general language proficiency and selected subsidiary communicative and linguistic skills, such as listening comprehension, text production and asking questions (grammar). At three test intervals between 2000 and 2002, 436 students completed a C-test which was used to define three general competence levels (on the basis of quartiles); the C-tests were compared with tests in the subsidiary skills. Statistical analyses show that this procedure is a valid way to define and measure students’ competences. However, it has been argued that test formats, such as the C-test and multiple-choice, might overly restrict the concept of ‘competence’. How should one take account of such a complex construct, which involves more than mere declarative knowledge about grammatical structures? Besides the empirical data, Zydatiß (Reference Zydatiß2005: 339) provides us with a catalogue of research questions and thoughts for future projects, as well as suggestions for ways to improve the NES.

A different approach is used in one of the German Research Foundation's priority programs on the modeling of student competences (Klieme & Leutner Reference Klieme and Leutner2006). Hartig & Höhler (Reference Hartig and Höhler2010) investigate the use of item response theory models and the interaction of subsidiary skills to solve a task in standardized tests to represent linguistic competences. In contrast to Zydatiß (Reference Zydatiß2005), the theoretical model is based on the assumption that the communicative and linguistic English competences of students can be measured more accurately by focusing on subsidiary skills rather than general linguistic proficiency. Hartig & Höhler's project is based on data from the DESI study (tests on language awareness, i.e. grammar, reading and listening comprehension), and the results show that the inclusion of a theory-driven development of test items alongside the appropriate psychometric measurement techniques can provide additional information on students’ competences (Hartig & Höhler Reference Hartig and Höhler2010: 196). Although this approach to competence modeling by psychologists of learning is an innovative one, it is essential to include the perspective of FL learning and teaching methodology to enable the appropriate implementation of research results in the EFL classroom.

Working within the same program, Pant, Tiffin-Richards & Köller (Reference Pant, Tiffin-Richards and Köller2010) deal with the question of setting standards for competence tests in large-scale assessments. To evaluate NES, it is important that statistical test results of students’ competences can be understood by researchers, teachers and students alike (Pant et al. Reference Pant, Tiffin-Richards and Köller2010: 175). This project investigates the discriminatory power and accuracy of the cut-off scores that separate two adjacent competence levels, such as A1 and A2 or A2 and B1. The database consists of items which have already been tested with 2,932 students and which include tasks on reading and listening comprehension in EFL (Pant et al. Reference Pant, Tiffin-Richards and Köller2010: 179). In a quasi-experimental design, 45 experts in four groups rated the selected items and set cut-off scores for 74 items each (listening and reading tasks). Two groups were homogeneous groups of language teachers while the other groups were mixed, and comprised teachers, researchers from the field of FL teaching, researchers from the field of psychometrics, and government employees. First results show that there are differences between the groups of experts: for example, teachers set lower cut-off scores than mixed groups (Pant et al. Reference Pant, Tiffin-Richards and Köller2010: 181). The feedback of the experts on the standard-setting procedure and methods needs further investigation, for example with the help of the think-aloud procedure to gain deeper insights into their thoughts on standard-setting.

Porsch, Tesch & Köller (Reference Porsch2010) focused on the development of items to test the NES for French as a FL. Their results were obtained from three sub-studies. First, a non-representative pilot study was conducted (Porsch & Köller Reference Porsch and Köller2010: 209) to test an item pool which had been developed and tested in a pre-pilot test. All test items were based on the competence levels of the CEFR. Tasks in listening and reading comprehension as well as text production were completed by 2,697 students. Items were tested for their suitability as competence measurements (the focus was clearly methodological). On the basis of a range of statistical analyses, including item response theory and Rasch models, a sufficient number of items were identified as meeting the criteria as test items for communicative French competences (Porsch & Köller Reference Porsch and Köller2010: 224).

The second sub-study focused on the validation of test items and the setting of cut-off scores. In the standard-setting procedure they call the ‘bookmark’ method, 19 experts (researchers on French teaching methodology, French teachers and researchers on psychometrics and test development) identified the level of difficulty of the test items and defined the thresholds between one competence level and the next (Porsch et al. Reference Porsch and Köller2010: 249).

In the areas of listening and reading comprehension, experts identified test items for four cut-off scores by rating all test items according to the levels and descriptors of the CEFR. Cut-off scores were set between A1 and A2, A2 and B1, B1 and B2, and B2 and C1. To set a particular cut-off score, experts had to agree on the level of difficulty of the test items.

Porsch et al. (Reference Porsch, Tesch and Köller2010) related these results to students’ test results and compared them with their French competences, particularly in the sub-skills of listening and reading comprehension. They could thus make clear statements about students’ achievements at a particular point, such as the beginning or end of a school year. Such empirical assessment of student competences is one of the defined functions of NES.

This essential step towards empirical competence modeling having been taken, future studies need to further adapt and develop the bookmark method as an instrument in the standard-setting procedure, make use of other comparable methods and initiate subsequent measures of validation.

Although the bookmark method has proved to be appropriate to its subject matter, experts have criticized the difficulties experienced in reaching a consensus about the test items’ level of difficulty (Porsch et al. Reference Porsch, Tesch and Köller2010: 263). Nor is it clear whether the same criteria were used, or if they were, which were the determining criteria. Moreover, it remains to be investigated how learning scenarios might be designed, on the one hand, to be learner-oriented and, on the other, to depict measurable levels of difficulty.

Test items relating to text production are seen to be more challenging, since coders have to assess the student texts on the basis of a priori criteria, such as a holistic impression of the text, grammatical and lexical aspects, orthography and coherence (Porsch & Köller Reference Porsch and Köller2010: 230).

In the third sub-study, a coding scheme was used to assess text production in French, the FL of 3,602 students (Porsch Reference Porsch2010: 267). Since the overall aim of this research project was not only to provide test items for different competence levels but also the cut-off scores, the focus was on the rating procedure and analysis of the data rather than on the content-related statements about the text production.

Porsch & Köller (Reference Porsch and Köller2010) found that, in making a holistic assessment, raters focus mostly on lexical and grammatical aspects, while the content of text production did not seem to be an indicator of writing competence. However, evaluating a student's text production merely from a linguistic point of view, excluding content-related criteria, does not comply with a functional language learning approach. Future studies will have to address the issue of a more differentiated approach to assessment.

Porsch et al. (Reference Porsch, Tesch and Köller2010) have conducted one of the most comprehensive studies so far on educational standards for FLs, combining the knowledge of outstanding researchers in French teaching methodology with that of experts on the highly innovative quantitative methods of competence modeling. The chapter on empirical efforts also attests to a high degree of critical reflection on the research methods and the questions which will have to be dealt with in future.

Although the current debates on NES show that many challenging research tasks lie ahead (Klieme Reference Klieme2004; Finkbeiner & Knierim Reference Finkbeiner and Knierim2008: 150), little empirical research into task-based language learning is as yet being carried out by German FL researchers.

One interesting approach is that described by Nold & Rossa (Reference Nold and Rossa2006). On the basis both of theoretical assumptions on task-based language learning and of empirical data on the development of test tasks on listening comprehension, they conclude that what is already known about the development of test tasks in general can also be applied to FL learning tasks (Nold & Rossa Reference Nold and Rossa2006: 79). As the same schemes for the analysis of the characteristics of tasks can be used in either context, both can benefit from and take advantage of the positive synergy between these areas of research to validate test tasks. In addition, in the language classroom, appropriate criteria-driven learning tasks can be developed to facilitate students’ learning processes (Nold & Rossa Reference Nold and Rossa2006: 83). However, no empirical study to prove this assumption in a school context has yet been conducted.

Müller-Hartmann & Schocker-v. Ditfurth (Reference Müller-Hartmann and Ditfurth2005) present a collection of works on task-based language learning and teaching and attempt to provide more insight into the area. However, they refer only to theoretical assumptions and practical examples of teaching. It remains to be seen if these findings lead to empirical investigations on the effect of such learning tasks.

The NES have stimulated broad interest in FL assessment and empirical analyses of student competences. A methodological shift can be observed as researchers from different fields, including psychologists and educational scientists, have introduced item response theory and psychometric models into FL research. This means that multiple data from different perspectives on students’ competences must be triangulated to make empirical results valuable for FL teaching practice.

3.2 Teacher education

International studies on students’ learning processes and the current concern with effective ways to foster learning show that the teacher has become a prominent factor in educational research in Germany (cf. Terhart, Bennewitz & Rothland Reference Terhart, Bennewitz and Rothland2011). Research questions include the qualifications teachers need to support their students adequately and, as a consequence, how future teachers can be prepared for their professional lives (Finkbeiner et al. Reference Finkbeiner, Knierim, Smasal and Ludwig2012). While earlier studies in this field dealt with the teacher and his or her personality (cf. Caspari Reference Caspari2003), recent research focuses on professionalism in teaching. Teacher education programs in Germany are currently undergoing change because of the shifts in educational policy mentioned in the preceding section, which underlines the increasing orientation towards student competence.

Teacher education research in Germany is therefore a topic of great importance at present, but little of this is subject-specific: teacher education topics are discussed in general without a specific focus on FLs. Subject-specific research, however, could lead to more precise and valuable results that could be implemented more appropriately. Apart from the DESI video sub-study (Helmke et al. Reference Helmke, Helmke, Schrader, Wagner, Nold and Schröder2008), there is little large-scale work. In the main, the studies described here focus on a particular topic and discuss its implications for the education of teachers. What is needed is empirical studies on the effectiveness of FL teacher training programs to gain deeper insights and identify opportunities for future research. However, extending smaller studies to see whether their results apply on a broader scale would result in greater reliability, validity and objectivity.

In the video study of the DESI project (Helmke et al. Reference Helmke, Helmke, Kleinbub, Nordheimer, Schrader and Wagner2007), researchers focused on the interaction between students and teachers in 105 EFL classes, drawing conclusions about the effect of using video transcripts in FL teacher education. Two English lessons from each class were videotaped. Teachers were asked to focus on fostering free speech in all students and to choose one intercultural and one linguistic topic. However, no restrictions were placed on specific content, materials or methods (Helmke et al. Reference Helmke, Helmke, Schrader, Wagner, Nold and Schröder2008: 345). The teachers were interviewed after the lessons. The transcribed lessons were then coded on the basis of the COLT scheme (Communicative Orientation of Language Teaching Observation Scheme, Spada & Fröhlich Reference Spada and Fröhlich1995) along the categories of time (sentence length) and frequency (frequency of teacher feedback) (Helmke et al. Reference Helmke, Helmke, Kleinbub, Nordheimer, Schrader and Wagner2007: 39). Additionally, important aspects of lesson quality were rated by intercoders, and intercoder reliability tests were calculated (Helmke et al. Reference Helmke, Helmke, Schrader, Wagner, Nold and Schröder2008: 348).

Future studies should take advantage of the rich database provided by DESI. Teachers will be able to benefit from the DESI classroom transcripts on different levels, for example by analyzing indicators of lesson quality such as the percentage of teacher talk, classroom language and code switching, feedback and error correction. Students in FL teacher education programs can become more aware of and reflect on classroom situations while intensively observing real classroom discourse.

New media are another focus in FL teacher education research (Grünewald Reference Grünewald2009). In a study on the acceptance and efficiency of blended learning concepts in university teacher education programs in the modern languages English, French, Spanish and German, a mixed method approach was used, involving 186 FL teacher degree program students from three German universities, who filled out a questionnaire before the seminar. Of the 186 students, 95 took part in online activities and completed the evaluation questionnaire. A guided interview was conducted with 53 students. Results show that students had a positive attitude towards teaching with new media and computer-assisted teaching and learning. They appreciated not only the opportunity to learn more about the use of computers and the internet in teaching themselves, but also the innovative character of the blended learning seminar (Grünewald Reference Grünewald2009: 229). On a scale from one to five, a high average mean value of 4.36 was obtained for the value gained from blended learning seminars. Furthermore, 83% of the interviewed students (n = 53) rated blended learning seminars as flexible in terms of time management.

For future research projects, Grünewald (Reference Grünewald2009: 242) suggests training in virtual communication before the actual seminar to foster the future teachers’ own media competences. The results call for further differentiated studies which go beyond students’ self-reports and use different methodological approaches to gain deeper insights into the effectiveness of hybrid learning scenarios. For example, one might suggest comparing treatment groups with control groups in a pre/post design to find out about the effects of two different online activities on student performance and attitudes.

In summary, research on (foreign) language teacher education is still in its infancy, and further studies are required, particularly in the area of FL teachers’ professional development.

3.3 Early FL learning

Early FL learning has a long tradition in Germany. It targets children in kindergarten from the age of two to six years as well as preschoolers and children at primary schools between the ages of five and ten or twelve. It differs from regular FL learning with respect to three variables: the children's age, the stress on speaking, and a high degree of heterogeneity among the children, which arises because they do not learn in separated performance-related groups until grade five (Christ Reference Christ2007: 449). Presumably one of the main challenges of early FL learning is to provide for ample contact time with the FL in order to enhance language acquisition processes that are as natural as possible. Since ‘it is dialogue that constructs linguistic knowledge’ (Swain Reference Swain2000: 98), a strong focus should be placed on verbalization during these processes.

The first large-scale study to describe the landscape of early FL learning on an evidence-based and transnational basis was that put out for tender by the European Commission in 2005 (EAC 89/04, Lot 1) and conducted by Kubanek-German & Edelenbos (Reference Kubanek-German, Edelenbos, Reichel-Wehnert and Schulz2008). The main outputs were meta-analyses of 400 studies on early FL learning in 31 countries, which since 1999 have included Germany. Examples of ‘good practice’ were also collected and categorized with the help of experts, and key pedagogical principles underlying early FL learning were determined. Finally, the outcomes of the application of these pedagogical principles were assessed (Edelenbos, Johnstone & Kubanek Reference Edelenbos, Johnstone and Kubanek2006; Kubanek-German & Edelenbos Reference Kubanek-German, Edelenbos, Reichel-Wehnert and Schulz2008; Kubanek-German & Edelenbos Reference Kubanek-German and Edelenbos2009). The results indicate that, even though an early start can have a substantial advantage for children, it does not guarantee a positive outcome: it takes a supportive environment and continuity in language learning to produce results. All children pass through a number of developmental stages in the acquisition of the target language, in the same chronological order, but at different speeds (see, below, Pienemann, Kessler & Liebner Reference Pienemann, Kessler and Liebner2006). Two learner characteristics have a striking impact on competent language use: motivation and aptitude. The intrinsic motivation of the children arises not only through fun activities and games but also from intellectual challenges. It is also important to introduce young children to reading and writing rather than restricting them to speaking and listening.

Results regarding ‘good practice’ show that there is a huge variety of activities which contribute to the creation of a ‘language-friendly environment at school’ (Edelenbos et al. Reference Edelenbos, Johnstone and Kubanek2006: 8). Research is needed in the following areas: the role of the teacher, the relation between different models of early FL learning and learning results, the development of children's motivation and their learning progress, as well as the relation between teacher profiles and ‘good practice’.

The benefits of this research project lie in its multi-method procedure, which allows the contrastive depiction of the similarities and differences of the situation of early FL learning across various countries and the determining of key principles for the teaching and learning of languages by young children. The study also has value in its combination of meta-analyses and expert-generated principles. Edelenbos et al. (Reference Edelenbos, Johnstone and Kubanek2006) conclude with clear recommendations for teacher education and further research.

Hypothesizing that primary school children's English language acquisition follows a developmental path, Pienemann, Kessler & Liebner (Reference Pienemann, Kessler and Liebner2006) carried out a profile analysis of speech data from 70 children, 56 of whom were primary school children and 14 were secondary school children. Learners’ utterances were elicited by communicative tasks and examined for 25 developmental characteristics and diverse variation characteristics. The outcome reveals that in primary school English lessons, the learners’ language acquisition proceeded according to the known developmental characteristics. The learners showed the developmental characteristics of the respective lower levels; in other words, higher developmental characteristics included those of lower levels. This points to a continual development process and proves that fundamental psycholinguistic regularities apply to the learning of EFL in the primary school. The results coincide with one of the results in the above-mentioned project by Edelenbos et al. (Reference Edelenbos, Johnstone and Kubanek2006).

Pienemann et al. (Reference Pienemann2006) describe in detail the procedures they used to categorize and evaluate the large data corpus of at least 1,750 measurements (25 developmental linguistic characteristics were measured for 70 students). They provide substantial empirical evidence for the support of Pienemann's processability theory (Pienemann Reference Pienemann1998).

The TAPS study (Testing and Assessing Spoken English in Primary School) was set up (Diehr Reference Diehr2006; Diehr Reference Diehr2007; Diehr, Frisch & Rensch Reference Diehr, Frisch and Rensch2007) as a response to the urgent need for performance assessment in the primary school. Its main goal was to develop, test and evaluate a feasible instrument for the assessment of English speaking performance in primary school. A sample of 216 children from grades 3 and 4 were videotaped in a lesson in which they had to retell a story to each other in small groups. The children's speech performance data were collected and rated with the help of a semi-standardized assessment questionnaire and level descriptors. Results of the quantitative data analysis revealed that the children exceeded the requirements of the curriculum (Bildungsplan). For example, children from grade 3 produced 70 words in a picture description and learners from grade 4 produced as many as 263 words while retelling a story. The qualitative analyses showed that the children were able to deal with the FL productively and reflectively: they not only used grammar rules to formulate sentences but drew to a great extent on expressions they had learned and retrieved from the mental lexicon as prefabricated chunks. These helped them produce complex structures with relatively little effort. Diehr et al. succeeded in developing a reliable instrument which can be used by teachers in the classroom. Inter-rater reliability was satisfactory (Cronbach's α >.70) and the dimensions of speaking performance used were based on a psycholinguistic speech model.

‘Multilingualism in primary school’ was the central interest of Kierepka & Krüger's project (Reference Kierepka and Krüger2007), which concentrated on the acquisition of two or three further languages (English, French and Russian) alongside German. During the implementation of two different models of multilingual FL learning in four primary schools, school administrators and teachers were advised by researchers. Questionnaires for parents and students as well as interviews and tests with students were used to determine how the students coped with the simultaneous learning of several languages. Focus was placed on the development of language and cultural awareness among 140 students by comparing their awareness of the different languages. The tests showed that the children intuitively detected and reflected on similarities between the languages. Moreover, not only the students but their teachers and parents, too, became more language-aware. The hypothesis had been that students who learn several languages in primary school have extended possibilities to gain experience in language acquisition through the comparison of languages, the identification of strengths and weaknesses and the estimation of their own capabilities, and this hypothesis was confirmed.

The project is valuable as it has great practical relevance. In addition, a comprehensive evaluation of the students’ competences was carried out by combining different instruments and triangulating different results. However, the authors themselves mention the lack of a control group. One reason for this was that French was not being taught as first FL in Thuringia at the time of data collection (Kierepka & Krüger Reference Kierepka and Krüger2007: 126). This, in turn, means that even though the results indicate the students’ positive attitude towards other cultures and FLs, it remains unclear how far this can be attributed to the intervention.

Möller, Zaunbauer & Gebauer (see, e.g., Zaunbauer & Möller Reference Zaunbauer and Möller2007; Kuska, Zaunbauer & Möller Reference Kuska, Zaunbauer and Möller2010) are conducting an ongoing longitudinal study on the role of learner variables in immersive learning in FL classrooms (the term immersion is used to describe programs in which the FL is used as language of instruction in all subjects apart from native language lessons). A comparison is being drawn between the performances in mathematics, German and English of primary school students taught in immersive vs. non-immersive classrooms. In the project, the relationships between language development and cognitive variables, language-related variables and motivational variables are analyzed. This project has been funded by the DFG since 2005. First results show that the FL reading comprehension of students in immersive classrooms is considerably higher than that of students taught non-immersively; however, the groups do not differ with regard to competence in first language reading or mathematics.

According to Elsner & Kessler (Reference Elsner and Kessler2010: 2), there is growing interest in the implementation of CLIL programs in primary schools in Germany, and the number of primary schools offering CLIL is increasing. Elsner & Kessler have been studying this topic for the last few years; a comprehensive report has yet to be published but is expected to offer many insights.

To conclude, although the demand for visible and accountable results in early FL learning has led to curricular requirements and standards for increasingly systematic English lessons (Sekretariat der Ständigen Konferenz der Kultusminister der Länder in der Bundesrepublik Deutschland 2005), empirical evidence for the adequacy of the goals that have been set is still lacking (Diehr Reference Diehr2007: 96). Solid evaluation is needed to complement and extend knowledge about early FL learning, which has to a large extent evolved from the practice and subjective theories of teachers (Christ Reference Christ2007: 453).

3.4 Content and language integrated learning (CLIL)

CLIL in Germany did not emerge from established theoretical concepts but evolved as a grass-roots movement out of experimentation by committed teachers (Zydatiß Reference Zydatiß2007a). As a result, in the first 20 years, there was no systematic empirical research (Helbig Reference Helbig2007: 184). But since the publication of a number of empirical studies on the evaluation and effect of CLIL, it has become a hot topic. Most of these, however, have focused on the language aspect of CLIL, while the content aspect – subjects such as geography or biology – has only marginally been taken into account (ibid.). Fehling (Reference Fehling2008) points to the fact that alongside the term ‘content and language integrated learning’, other terms are used in the research literature, sometimes as synonyms, although they in fact denote different models of bilingual learning and teaching. Here, we use the terms ‘CLIL’ and ‘bilingual learning and teaching’ to refer to education in which ‘subjects such as history, geography, politics and biology are taught in a language other than German, generally English or French. That is, a foreign language is used as the instructional language’ (Fehling Reference Fehling2002: 1).

Historically, the initiation of bilingual educational programs can be traced back to the German-French cooperation treaty (Élysée Treaty) of 1963, when the instructional language in CLIL was mainly French. This changed in favor of English with the transition from the European Community to the European Union in 1992/93; 500 out of 700 schools now offer English CLIL programs. CLIL is seen as useful for the effective support of FL competences (Zydatiß Reference Zydatiß2007a: 32). Until now, CLIL programs in Germany have mostly been offered in secondary high schools and mainly in geography, history and politics (Helbig Reference Helbig2007: 181; Zydatiß Reference Zydatiß2007a: 30). CLIL is usually organized in individual schools as bilingual tracks or special programs, such as modules (Fehling Reference Fehling2008) for which students have to apply.

The following account of several studies will illustrate the effect CLIL can have on the acquisition of language-specific and subject-specific competences.

In a sub-study of the DESI study already mentioned (see section 3.1), the performance of 958 students in CLIL programs was compared with 987 students outside CLIL programs (Nold et al. Reference Nold, Hartig, Hinz and Rossa2008). English competences such as text reconstruction, listening comprehension, reading comprehension, grammar, socio-pragmatics and writing were tested. Overall, the outcomes showed that CLIL comprehensively achieves its aims: on average the ninth-graders attending CLIL programs achieved a language proficiency level not reached by non-CLIL students until grade eleven (A-levels) (Nold et al. Reference Nold, Hartig, Hinz and Rossa2008: 456). Nold et al.'s research was the first evidence produced in Germany to confirm the effectiveness of CLIL programs on the basis of a very large sample. However, this result is perhaps to be expected, as the time allocated to the target language differs significantly between CLIL and non-CLIL classes and thus is very likely to have an effect on the learning outcome (Fehling Reference Fehling2008).

In the DEZIBEL-project (Deutsch-Englische Züge in Berlin, German-English Tracks in Berlin), the English language performance of 180 learners in CLIL programs in grammar schools (Gymnasien) in Berlin was tested and compared (Zydatiß Reference Zydatiß2007b; Reference Zydatiß2007c). The analyses showed that CLIL students had a higher language proficiency level than students taught in regular classes. Moreover, significant differences existed in CLIL students’ favor as regards the qualitative characteristics of text-type related writing tasks. Strikingly, around a third of all CLIL students leave these classes at the end of grade ten or eleven. Zydatiß suggests two school-system-related challenges to explain this. First, CLIL students’ higher achievements are not sufficiently acknowledged, as the rigid traditional grading system does not allow for a differentiated recognition of their performance. Second, the requirements of the course system for the upper secondary level (Oberstufe) are not flexible enough. Schools with CLIL classes need to resolve the conflict between the development of a school profile on the one hand and the problem of fairness towards the students on the other. The study shows that CLIL programs can have an effect on students’ language proficiency. Something that is often neglected, however, is the higher pre-existing level of language proficiency of CLIL students compared to non-CLIL students at the entrance level (Fehling Reference Fehling2008). The most noteworthy part of Zydatiß’ study is his critical depiction of what he calls the ‘fairness trap of bilingual programs’ (Zydatiß Reference Zydatiß2007c: 161), which corroborates the positive outcomes. He proposes ‘counter strategies’ and concludes with recommendations to ensure high quality bilingual education in Germany.

In a project on subject-specific learning and (foreign) language learning with a focus on task-based cognition, communication and cooperation funded by the DFG, Vollmer (Reference Vollmer2007b, Reference Vollmer2009) tested the written geographical competence of students in a CLIL (n = 84) and a traditional regular geography class (n = 90). Discourse functions such as describing, explaining and assessing (Vollmer Reference Vollmer2009: 171) were measured on a holistic scale comprising six levels. According to the findings, CLIL geography students were on the same subject-specific competence level as traditional geography learners. However, both groups showed insufficiently developed subject-specific literacy, that is, genre writing specific to the subject.

Vollmer's study sheds light on the other side of the coin of CLIL programs: students’ content-specific proficiency. Though his research is representative for only one of Germany's federal states (Lower Saxony), it points to a significant need for CLIL classes across Germany: the promotion of scientific literacy.

Weitz & Rohde (Reference Weitz and Rohde2010) investigated variation in the receptive English vocabulary of German-speaking children in three different bilingual kindergarten programs. Ethnographic observations as well as questionnaires were used to gain information about the organization of the programs and L2 use in the different contexts (Weitz & Rohde Reference Weitz and Rohde2010: 65). The British Picture Vocabulary Scale (BPVS II, Dunn, Whetton & Burley Reference Dunn, Whetton and Burley1997) was used to measure the FL proficiency of 49 German-speaking children. Comparison between the programs showed that the language proficiency attained strongly depended on how the recommended principles for the organization of such programs were put into practice. The consequent use of the ‘one-person-one-language’ principle as well as how native speakers and the children interacted influenced the success of such bilingual programs in kindergarten.

Weitz & Rohde's (Reference Weitz and Rohde2010: 68) findings are not generalizable as only 49 children were tested; they also varied widely in age (1–6 years old) and the duration of their L2 contact (2–46 months). Small groups with comparable characteristics therefore had to be analyzed separately. Nevertheless, the results hint at the success indicators crucial for bilingual kindergarten programs. A large-scale implementation of the specifically adapted instruments used in this study, for example the BPVS II, would provide a foundation for the necessary improvement measures in bilingual kindergartens.

New vistas for intercultural learning are opened up through CLIL, as it provides a frame in which pedagogic and linguistic aspects of intercultural learning can be linked together (see section 4.4).

4.1 Motivation and interest

In this section, we look at how motivation and interest in the learning and teaching of languages are investigated in the context of learner characteristics. Until recently there has been little interest in non-cognitive factors such as these, but researchers now agree that these factors have to be taken seriously, as they might impact on learning success (Finkbeiner Reference Finkbeiner2005; Riemer Reference Riemer2006; Biel Reference Biel2007).

Until the early 1990s, motivation research in the field of foreign and second language learning was dominated by the socio-psychological approach of Gardner, Lambert, Clément and colleagues (Biel Reference Biel2007: 2). In contrast, there has been a much longer tradition of research in motivation and interest in the field of educational science and psychology (Finkbeiner Reference Finkbeiner2005: 38).

The constructs ‘motivation’ and ‘interest’ are often not clearly defined and are taken to comprise a range of concepts, such as motivation to perform, volition, anxieties, motives, expectations, abilities and self-concepts (Solmecke Reference Solmecke1983; Preiser & Sann Reference Preiser and Sann2006). As part of the effort to unify and itemize all the relevant factors, the development of different models has been pivotal to research in this field.

In 2006 Küster (Reference Küster2007) conducted a survey on the motivation of 290 learners of French as a FL from grades 9 to 11 in grammar schools (Gymnasien) in Berlin. According to the respondents’ answers, the subject was primarily chosen instrumentally: because of its usefulness. A personal connection to France and its culture was claimed by 41% of the learners. However, more than a third (n = 99) said they had given up the subject due to low performance assessment, and 59% considered themselves unable to express their thoughts in the FL. The learners expressed a desire for innovative methods such as creative writing and cooperative learning, and for teachers with solid language and cultural skills, methodological competence and subject interest.

Although the structure of Küster's questionnaire is described in detail, it remains unclear how ‘motivation’ and ‘interest’ were defined and how these single-dimension constructs were expressed in the questionnaire, particularly as the terms were used interchangeably. However, the outcomes of this study are of interest for researchers in this field because of the number of learners surveyed, even though the results are representative only of the federal state of Berlin.

In a mainly qualitative case study, Abendroth-Timmer (Reference Abendroth-Timmer2007) explored the acceptance of bilingual and multilingual modules by students and their motivational development. The modules were in a range of subjects (such as sociology, drama and physical education) which were taught in English, French or Spanish. Through action research, she collected data by means of questionnaires, diaries, motivation graphs and guided interviews. Quantitative analyses of the data from 77 students were combined with qualitative case studies of six students. The results showed that students with an interest in languages can be motivated to study the subject content through bilingual and multilingual modules, as can students who are interested in the subject content. The students generally welcomed new forms of learning. Various learner characteristics, such as the student's concept of ‘self’, have a greater impact than context characteristics. One very important motivational factor is the student's awareness of their own success in language comprehension, as students tend to underestimate this.

Using methodological triangulation, Abendroth-Timmer was able to provide significant insights into FL learners’ motivation, not all of which can be described here. The strength of her study lies in the single case studies which illustrate the individuality of learners’ motivational structures.

Research on the interaction between interest and reading comprehension was conducted by Finkbeiner (Reference Finkbeiner2005): her study will be described in the next section.

4.2 FL reading competence

In a practical handbook on how to teach English literature, Nünning & Surkamp (Reference Nünning and Surkamp2006: 12–;13) point to a recent shift in the significance of teaching literature in FL classrooms, in which literary texts are seen as useful tools to achieve learning objectives, such as the development of empathy, understanding the other and the adoption of perspectives. This shift is connected to the notion of reading as an interactive process between the reader and the text, as a result of which interest in FL literacy studies is increasing. The field has also advanced considerably through the DESI study (DESI-Konsortium 2008, see also section 3.1 for detailed information). In one part of the study, which focused on reading comprehension (Nold, Rossa & Chatzivassiliadou Reference Nold, Rossa and Chatzivassiliadou2008), different facets of reading comprehension were captured in tasks relating to various text types and diverse linguistic and text-pragmatic levels. These tasks were designed with reference to three curriculum-based competence levels (A, A+, B and C), which had been developed by experts beforehand. One of the main outcomes is that 44% of the 11,000 ninth-graders had considerable problems understanding English texts and were not capable of solving reading tasks at the lowest level (A); 34% of the students reached level A, and 22% levels B and C. These results hint at two deficiencies. Firstly, the competence levels and resulting expectations formulated in the curricula were unattainable for a large group of students, a fact from which useful conclusions were drawn for the definition of national educational standards (NES, see also section 3.1). Secondly, there is an evident need for more research in this area and the development of support programs for students.

The extent to which socio-affective variables such as emotions and interest influence the reading process of EFL learners was investigated in the following two studies by Burwitz-Melzer (Reference Burwitz-Melzer2008) and Finkbeiner (Reference Finkbeiner2005).

Burwitz-Melzer (Reference Burwitz-Melzer2008) conducted an empirical case study on emotions in EFL literature classes, the assumption being that they should be observable in FL reading. According to Burwitz-Melzer (Reference Burwitz-Melzer2008: 32), an ‘emotion’ consists of different reactions to a stimulus, such as affective appraisal, assessment, cognitive appraisal and monitoring. To encapsulate this processual character of emotions, she uses the term ‘emotional process’. A group of 14 tenth-graders from a secondary high school (Gymnasium) near Frankfurt engaged in several creative tasks on a US-American short story. Four of six successive lessons were videotaped, followed by retrospective guided interviews with the teacher and several students. Analyses of transcripts show that emotional processes are not only experienced by the learners but ‘also verbalized and discussed and used for the exploration of the meaning of the text as well as for an evaluation of the learning process’ (Burwitz-Melzer Reference Burwitz-Melzer2008: 27). Based on these results, Burwitz-Melzer (Reference Burwitz-Melzer2008) regards her assumptions as justified and concludes that along with intercultural competences, the competence to express emotions can be developed through active engagement with intercultural texts. Her study provides a good theoretical background followed by a practically oriented study. How the interviews contribute to the stated results and conclusions, and how the transcripts were analyzed, are not explained. Since the analytical criteria are not specified, the study is not fully transparent.

The main focus of Finkbeiner (Reference Finkbeiner2005) is the interplay of reading interests and reading strategies and their effect on learning during the processes of reading and understanding English texts. She triangulated data in a multi-trait/multi-method approach and conducted both a quantitative and a qualitative study in several phases: the first part of the main study consisted of a semi-standardized survey with four questionnaires on reading interest (ITEF), reading strategies (LEFT), reading comprehension (CTEF) and the concept of ‘self’ among 287 students of grade 9, and an open questionnaire with 13 EFL teachers. Factor analyses, correlations, regressions and LISREL analyses provided insight into the determining factors of FL reading (Finkbeiner Reference Finkbeiner2005: 257). Elaboration strategies play a major role in understanding English texts (Finkbeiner Reference Finkbeiner2005: 372). In the second part of the main study, 57 of the 287 students were systematically sampled and interviewed. This was followed by a qualitative comparative culture study with ten students from Taiwan, ten from the USA and ten from Germany.

Finkbeiner (Reference Finkbeiner2005: 113) found empirical evidence for a distinct connection between elaboration strategies (relating new information to prior knowledge) and reading interests. Personal elaboration (relating new information to personal experience) plays a major role in FL reading comprehension and, as a strategy, is dependent on interest.

A LISREL model demonstrates that personal elaboration (Finkbeiner Reference Finkbeiner2005: 372) and self-regulated learning are predictors for in-depth processing during FL reading. Activating strategies play a key role as intervening variables when it comes to the effect of interest in FL reading on in-depth processing. The exploratory part of the comparative culture study yielded cultural differences in reader response tasks. The study's most valuable insight is gained through the LISREL analyses, in which the constructs ‘interest in foreign language reading’, ‘personal elaboration’, ‘self-regulated learning’ and ‘in-depth processing’ are interrelated.

Peyer & Kaiser (Reference Peyer and Kaiser2010) compared the objective test results with self-assessment results of 506 French- and Italian-speaking university students of German regarding their reading competences and reading performance. A linked question was how far learners were aware of their reading comprehension problems in German as a FL. The reading comprehension and competence were assessed in three ways: a questionnaire on global self-assessment reading competence, a reading test and a placement test. A qualitative study, in which students’ verbalized thought processes whilst working on texts were audio recorded, yielded complementary results. The students were quite critical in their self-assessment, partly because German has the reputation of being a difficult language. The majority of learners assessed their German reading competence one level below their overall German proficiency (Peyer & Kaiser Reference Peyer and Kaiser2010: 79). In contrast, they felt confident when answering reading comprehension questions. This confidence increased with the students’ level of German proficiency: advanced learners, in particular, tended to have a false sense of security and thus often failed to realize the existence of comprehension problems. Peyer & Kaiser (Reference Peyer and Kaiser2010) suggest that self-assessment cannot be seen as an alternative or substitute for objective measurement but rather as a useful addition. Questions concerning self-assessment and confidence can also support the students’ awareness of comprehension problems when reading FL texts.

With their solid research study, Peyer & Kaiser make a vital contribution to the debate on FL reading competence. Assessment and awareness of one's own reading comprehension are critical to enhancing reading competence. Peyer & Kaiser show how this may be facilitated.

The studies described in this section show that research on reading is often combined with research on other topics, such as intercultural learning (Burwitz-Melzer), learner variables such as strategies and interest (Finkbeiner) and assessment methods (Peyer & Kaiser). From the accounts given, it can be inferred that FL reading serves as a vehicle for the development of various competences and that literacy plays a crucial role in FL acquisition in general. Despite the growing interest in, and the obvious need for, the investigation of FL reading, research in the area of FL reading is still rather rare.

4.3 Learning strategies

While the focus of learning strategy research has long been on the question of terminology and the identification and classification of strategies, studies now focus on factors which influence the quality of strategy use, and hence learner success.

The DESI study mentioned earlier uses three scales adapted from the Strategy Inventory for Language Learning (SILL) questionnaire by Oxford & Burry-Stock (Reference Oxford and Burry-Stock1995) and several scales from the PISA study, to find empirical evidence that general and area-specific learning strategies, such as reading strategies, contribute to the prediction and explanation of students’ achievements (Schrader et al. Reference Schrader, Helmke, Wagner, Nold and Schröder2008: 270).

The quantitative analyses show strong correlations between (language-specific) metacognitive learning strategy use and English proficiency although, in contrast to Finkbeiner's findings (Reference Finkbeiner2005), general learning strategies (elaboration strategies) are found to account only for a weaker influence on the learning outcome (Schrader et al. Reference Schrader, Helmke, Wagner, Nold and Schröder2008: 280). Students who enjoy reading in the English language improve their performance over the course of the school year in inferring the meanings of new words and paraphrasing them. However, for students who could already do this well at the beginning of the school year, little change could be found in their reading activities by the end.

It is debatable if reports by students (questionnaires) alone can tell us anything about the influence of learning strategies on learners’ success (Schrader et al. Reference Schrader, Helmke, Wagner, Nold and Schröder2008: 280). Finkbeiner et al. (Reference Finkbeiner, Knierim, Ludwig and Wilden2008) therefore used coded videotapes and transcripts to analyze the quality of learning strategies and their effect on students’ success.

The focus of the DFG-funded ADEQUA (Adequacy of learning strategy use and teacher support actions) video study was on the adequate use of learning strategies in EFL reading in a cooperative, autonomous learning environment (Finkbeiner et al. Reference Finkbeiner, Knierim, Ludwig and Wilden2008; Finkbeiner et al. Reference Finkbeiner, Knierim, Smasal and Ludwig2012). The study had two phases. In the first video study (n = 164), the ninth grade (all tracks) students’ learning strategies were videotaped, coded and analyzed while the students were working together in dyads on a reading task. The video data were supplemented by learner data on constructs, such as demographic information, EFL proficiency, interest and self-regulatory competence. The main research questions focused on a) the identification of the applied learning strategies according to an expanded taxonomy of strategies (cf. O'Malley & Chamot Reference O'Malley and Chamot1990) and b) the quality of strategies used by students ‘naturally’ without specific training (adequacy and success with respect to the given situation).

Prior studies mainly relied on questionnaires, that is, self-reports on strategy use. However, the ADEQUA project took a different approach in which the learning strategies actually applied were analyzed on the basis of the videos.

One of the results is that adequate situation-specific strategy use is more important than frequent strategy use, as it has significant influence on the learning outcome (Finkbeiner et al. Reference Finkbeiner, Knierim, Ludwig and Wilden2008; Finkbeiner et al. Reference Finkbeiner, Knierim, Smasal and Ludwig2012). The use of a particular learning strategy is adequate when it can help to solve a text comprehension problem. For example, if there are unknown words in a text, it is not situationally adequate for students to look up every single word in a dictionary, as some words can be inferred from the context and others might not be crucial for text comprehension at all. A second result is that successful strategy use is not correlated with frequent strategy use: quality counts, not quantity.

These results led to a second video study in which teachers (n = 11) were observed in L2 classes while supporting their students’ (n = 273) autonomous use of learning strategies and helping them use situationally adequate strategies. The findings show that, probably because of a hidden curriculum, teachers ‘[. . .] predominantly support their students by carrying out actions for them instead of supporting the students in carrying out these actions themselves. In particular, these teachers summarize (parts of) the text and paraphrase or translate unknown words’ (Finkbeiner et al. Reference Finkbeiner, Knierim, Smasal and Ludwig2012). ‘Furthermore, with respect to all teacher support actions, only 35% focused directly on the learning process, and out of those, only 6% could be identified as metacognitive reflection’ (Finkbeiner et al. Reference Finkbeiner, Knierim, Smasal and Ludwig2012: 73). ADEQUA shows that teachers need to learn how to support their students adequately. More empirical studies are therefore required to reproduce these findings and to use the results in FL teacher training.

4.4 Intercultural competence, language awareness and intercultural awareness

The CEFR identifies intercultural competence as the ability which distinguishes FL learners from monolingual native speakers (Council of Europe 2011: 43). Intercultural competence does not only refer to knowledge about cultures and languages but also calls for the development of cultural awareness and language awareness (Svalberg Reference Svalberg2007).

The increased awareness of intercultural aspects of language learning arises from increasing cultural pluralism, personal mobility and transnational migration. As a consequence, students with highly diverse linguistic experiences are expected to acquire language-related competences and knowledge (Gogolin Reference Gogolin2007: 101).

The Gießen postgraduate program (Graduiertenkolleg) ‘Didactics of understanding others’ (Didaktik des Fremdverstehens), funded by the DFG over nine years, has increased interest in the debate on intercultural learning (Bredella, Christ & Legutke Reference Bredella, Christ and Legutke2000). Those involved in the program have thoroughly investigated the teaching and learning of intercultural understanding and developed models for intercultural FL learning and teaching (Krumm Reference Krumm2007: 140), and a number of researchers have emerged from the program to continue the work across the state and across different languages. Apart from the research conducted around the Gießen research group, a number of independent studies have been carried out in the field.

In a qualitative ethnographic interpretive study undertaken between 2000 and 2002, Fäcke (Reference Fäcke2006) wanted to discover (1) what mental processes relating to transculturalityFootnote ² take place during reading in FL literature classes and (2) the impact of mono- and bicultural socialization on the students’ mental processes and approaches to literary texts. Think-aloud protocols and reading diaries as well as follow-up interviews were collected from 19 adolescents from grades 11 and 12 during and after the students’ work on a text. Fäcke presents ten case studies and shows that biographic, sociopolitical, academic and literary contexts play an important role in the students’ engagement with the texts. Biographic experiences which refer to the students’ mono- or bicultural socialization prove to be highly relevant. Based on the data from three adolescents, Fäcke shows the crucial role of ‘transcultural openings’ for FL learners. She concludes that transculturality should be considered not only in theoretical debates but also in practical FL learning and teaching literature classes. Fäcke succeeds not only in analyzing mental processes with regard to transculturality, but goes a step further and connects these processes with biographical, sociopolitical, educational and literary contexts, thus achieving highly complex results.

Hu (Reference Hu2006) also examined the subjective significance of language learning for an individual's learning biography. She analyzed language memoirs and autobiographical texts written by seven bilingual or multilingual authors with special reference to linguistic and cultural identity, the coexistence of different languages and the perception of languages (2006: 183). The analyses showed that, in order to explore linguistic learning processes thoroughly, the individual's biographical experiences – their historical, political and idiosyncratic environment – must be taken into account. Furthermore, the learning of languages in the context of migration is an identity-related process. It can lead to changes and expansion of identities, to liberation as well as to conflicts (Hu Reference Hu2006: 197). Since these features are a central component of learners’ biographies, language learning in school should be reconsidered. Specifically, the interaction between the language that was acquired first and languages acquired later must be related to language learning in the school context.

Hu uses highly reflective texts by professional writers and encourages a similar approach to the investigation of school-based language learning. Her study should serve as an inspiration for similar research in the school context.

Schmidt & Finkbeiner's The ABC's of cultural understanding and communication (2006) is a professional model for intercultural learning and teaching which uses an autobiographical approach, as suggested by Hu. It aims at the development of language awareness and intercultural awareness through literacy activities. The procedure includes several steps: writing an Autobiography, conducting an interview with a partner and writing their Biography and Comparing A and B with respect to similarities and differences (C1), doing a cross-cultural in-depth analysis (C2) and developing ideas for culturally responsive teaching or for a culturally responsive workplace (C3). From 2009 to 2011 an ABC project was conducted with 12 EU and US partners funded by the EU (EAEAC) and the US government within the framework of the FIPSE/Atlantis program. The model was disseminated through university teacher education and business programs and applied in a range of subjects, socio-cultural contexts and durations. The results demonstrate that the model can be successfully implemented and that it contributed to students’ awareness of both self and the other. More detailed results can be obtained from the individual reports written by each instructor (Finkbeiner & Lazar Reference Finkbeiner and Lazar2013).

Göbel (Reference Göbel2009; Göbel & Helmke Reference Göbel and Helmke2010) analyzed data collected in the DESI project mentioned above to determine the effect of instruction on 11,000 ninth-graders’ intercultural learning results. Multi-level analyses were used to measure the influence of the variables ‘teacher competences’ and ‘classroom characteristics’. In addition, exploratory video analyses of lessons given by ten teachers were carried out to find out about the role that teachers’ intercultural experience (contact with the English-speaking culture) played in class. It turned out that the teachers’ ‘intercultural experience’, ‘discipline’ (defined as well-organized classroom management) and a ‘positive error culture’ (defined as ‘constructive treatment of errors in classroom discourse as perceived by students’) (Göbel & Helmke Reference Göbel and Helmke2010: 1574) in class had a significant influence on the students’ results. This confirmed existing research outcomes which found that intercultural learning results are dependent on didactic features. Moreover, ‘comparing less and more interculturally experienced teachers’ in the exploratory sub-study ‘showed that the latter were more likely to focus on (. . .) topics of subjective culture’ [cultural meaning and interpretation rather than surface culture and cultural products (= objective culture)], ‘and to give students the opportunity to share ideas and intercultural experiences than the former’ (Göbel & Helmke Reference Göbel and Helmke2010: 1579). This, in turn, had a positive effect on the students’ intercultural interest. Teachers generally had difficulties focusing on cultural and intercultural meaning construction in class, and teachers with little intercultural experience hardly dealt with intercultural topics at all.

The outcomes of this study so far elucidate the crucial role the teacher's competences and own experiences might play in the classroom. Complementing the statistical results with qualitative in-depth analyses, Göbel tried to demonstrate how important teachers’ intercultural competences can be for students’ intercultural learning. Teachers with rich intercultural experience seem to be more successful in eliciting students’ cultural awareness during their lessons (Göbel & Helmke Reference Göbel and Helmke2010: 1576), but as that study was of an exploratory nature only, more research is needed to obtain further insight into this important area. Furthermore, the notion of the construct of ‘culture(s)’ needs careful examination and differentiation with respect to ‘classroom culture’ and ‘target culture’ in any further research.

The idea of ‘Linguistic Awareness of Cultures’ (LAC) was developed by Müller-Jacquier (Reference Müller-Jacquier2000) during the 1990s; he argued for a systematic consideration of the communicative dimension of intercultural interaction processes. Bosse (Reference Bosse and Bauer2009) presents a way of implementing the LAC approach in the context of an intercultural program for university students in Hamburg (HOPIKOS, Cross-University Project for Fostering Intercultural Student Competence). The project involved introductory intercultural training with socio-psychological and cultural-anthropological content alongside linguistically-oriented two-day training workshops. Audio- and videographies of the classes and retrospective interviews were studied using conversation and content analysis. The study, still incomplete, provides first indications of how LAC can be fostered in the context of intercultural training over a short period of time. The results show that the training program may have a positive influence on cognitive pre-conditions for intercultural communication skills. The strength of the HOPIKOS project so far lies in its linguistic focus, which seems helpful in reflecting on one's own intercultural experiences (Bosse Reference Bosse and Bauer2009: 344). The implementation of a thoroughgoing linguistic approach to intercultural hands-on training is innovative. As a next step, research should focus on supporting the training's effectiveness through statistical measures.

Cultural research is currently in transition from interculturality to transculturality. One major question is whether the construct of culture can be measured and sub-divided along certain dimensions. Much research is still needed to obtain a better grasp of the nature of cultural competence. The importance of language awareness and cultural awareness research was highlighted by a number of researchers in Germany who presented their studies at the international conferences of the Association for Language Awareness in Kassel in 2010 and in Montreal in 2012. More research output is to be expected in the next couple of years.

5.1 Cooperative learning

Cooperative learning has become a popular topic in the debate on autonomous learning, social competences and student heterogeneity. With social, cultural and ethnic diversity increasing in German society, cooperation and communication are considered key qualifications for a successful school and employment career and are therefore set as overall goals in all official curricula (e.g. Hessisches Kultusministerium Reference Hessisches2010: 8). Cooperative behavior is seen as a complex interplay of personal characteristics that include various social, linguistic and cognitive skills as well as affective and attitudinal factors (Stanat & Kunter Reference Stanat and Kunter2001). Much of the research in this area can be traced back to Anglo-American researchers.

In Germany, there has so far been little research on cooperative learning in the FL classroom. Most investigations of ‘cooperative learning’ or ‘group work’ have been carried out by researchers in the field of educational research or educational psychology, and have addressed cooperation and interaction between students without regard to subject. Today's understanding of group work and cooperative learning emerges largely from the work of the education researchers E. Meyer & G. Meyer (Reference Meyer and Meyer1996) and Huber (Reference Huber2001). Many publications in which cooperative learning is placed in a FL learning setting give practical advice on how to implement different methods that foster cooperative learning (cf. Huber Reference Huber2001). Little is known, however, about the effects of cooperative learning on FL learning, about ways in which students can support each other's language learning process or about how small groups can best be composed to optimize FL learning. In addition, there have been many studies of students working cooperatively which have not focused on the cooperation itself. One of the advantages of collecting data in a cooperative setting is that students naturally verbalize their thought processes during interaction. This demands particular methods such as the ‘think-aloud’ method. As can be seen below, research on cooperative learning is often combined with other topics, such as computer-assisted learning, reading comprehension and task-based language learning. Due to both the lack of, and the need for, more research, only a few studies matched the selection criteria used for this article.

Spörer, Brunstein & Arbeiter (Reference Spörer, Brunstein and Arbeiter2007) and Spörer et al. (Reference Spörer, Seuring, Schünemann and Brunstein2008) investigated the reading comprehension of 74 seventh-grade students during peer-scaffolded learning by adapting the American PALS-program (Peer-assisted learning strategies, cf. Fuchs et al. Reference Fuchs, Fuchs, Kazdan and Allen1999) for implementation in a German comprehensive school. The aim was to find out if PALS leads to an increase in reading competence in German and English. Using a pre/post/follow-up design, two experimental groups worked with the PALS-program, one in English, the other in German; the control group received no treatment. Before the students worked with the texts, they were trained to use three specific techniques. Data were collected by means of protocols written by the students and the application of several tests. In the post- and the follow-up tests, the students in both experimental groups achieved higher scores in reading comprehension than the students in the control group, with the students in the German group achieving better results than those in the English group. Students experiencing both treatments improved in their reading and summarizing of texts during the training. The authors suggest that the outcomes of the training in English could be improved if the teacher's instructions, the tutor's help and the procedure were practiced in German first and then implemented in FL classes. All in all, these findings can be seen as an indication of the positive effect peer-assisted learning can have on the acquisition of language skills and reading competences.

This is one of the few representative studies on peer-assisted learning in a FL setting in Germany. The authors successfully adapted a training program which was originally developed in the USA and tested its efficacy in a very sound research study. They supply statistical evidence that positive results can be obtained through adapted cooperative training. Much more of this kind of research is necessary to facilitate effective language learning through cooperative learning.

For further research on cooperative learning and FL learning, refer to the ADEQUA project mentioned earlier (Finkbeiner et al. Reference Finkbeiner, Knierim, Ludwig and Wilden2008; Finkbeiner et al. Reference Finkbeiner, Knierim, Smasal and Ludwig2012), which focuses on the use of reading strategies during cooperative learning (see section 4.2.4 on strategies), as well as Assbeck's study (Reference Assbeck2007) on peer correction.

5.2 Computer-assisted language learning (CALL)

How new media can adequately and effectively be used in the language classroom or in language teacher education programs is one of the most pressing questions in the area of computer- or web-enhanced language learning and teaching. A considerable number of recent publications can be found on the use of new media in an educational context. Recently, more differentiated approaches have taken over, such as blended learning: a combination of face-to-face and e-learning; it is therefore also referred to as a hybrid learning environment (cf. Grünewald Reference Grünewald2009).

Another shift in research can be seen in the tendency to focus on didactical or methodological questions rather than technological problems, as at the beginning of CALL research (Klippel Reference Klippel2007: 11).

Despite the mass of publications, empirical research studies remain rare, especially concerning FL learning in school. Such research as there is has been mostly in the field of blended learning environments in teacher training.

Studies on the effectiveness of blended learning call for newly designed research instruments, a challenge addressed in two publications (a pilot and the main study) by Kupetz & Ziegenmeyer (Reference Kupetz and Ziegenmeyer2005; Reference Kupetz and Ziegenmeyer2006), who investigated a blended learning environment for future teachers of English at a German university, who participated in a seminar on a general introduction into English teaching methodology (Kupetz & Ziegenmeyer Reference Kupetz and Ziegenmeyer2005: 180). The study attempted to establish whether the blended learning scenario could lend support to a teacher's knowledge of teaching methodologies and his or her observations and reflections on teaching practice gained from video recordings as well as from the students’ autonomous and active learning (Kupetz & Ziegenmeyer Reference Kupetz and Ziegenmeyer2005: 181; Kupetz & Ziegenmeyer Reference Kupetz and Ziegenmeyer2006: 69).

The pilot study involved 34 students; 53 attended the seminar as part of the main study. While the pilot study was mainly concerned with the generation of hypotheses, the main study aimed at testing them. An action research approach was favored because both researchers also taught the seminar.

Two online questionnaires on the students’ background, experiences in teaching and e-learning attitudes were used in both studies, slightly adapted after piloting. The second questionnaire evaluated the specific components of the seminar (Kupetz & Ziegenmeyer Reference Kupetz and Ziegenmeyer2005: 187; Kupetz & Ziegenmeyer Reference Kupetz and Ziegenmeyer2006: 69).

Students were instructed to work with video recordings of an English lesson in order to, firstly, ‘support an understanding of planning lessons and reflecting upon them in the context of fundamental TEFL issues, and secondly [. . .] help the students to broaden their perspectives and see lessons increasingly from a teacher's perspective as well’ (Kupetz & Ziegenmeyer Reference Kupetz and Ziegenmeyer2006: 73). Additionally, an e-interview with the teacher from the videos was designed so that students and teachers could share experiences (Kupetz & Ziegenmeyer Reference Kupetz and Ziegenmeyer2006: 71). The e-interview was rated as unhelpful by 55% of the students in the pilot study, but only 23% in the main study (Kupetz & Ziegenmeyer Reference Kupetz and Ziegenmeyer2006: 75).

Most students (92%) greatly appreciated the work with digital video recordings because they illustrated teaching practice, but fewer (65%) saw the link to the theories of teaching methodology (Kupetz & Ziegenmeyer Reference Kupetz and Ziegenmeyer2006: 78). As the results are mainly based on the students’ own reports, further effort is needed to find an additional research tool in order to find out how the developed learning environment can actually foster students’ active and autonomous learning.

Legutke & Rösler (Reference Legutke and Rösler2005) combined cooperative language teaching with computer-mediated communication in the university setting and analyzed data from two research projects on the computer-mediated training of teachers of FLs at Gießen University. In project 1, the Gießen electronic work experience (Gießener Elektronisches Praktikum), the focus was on factors contributing to successful collaboration between tutors and tutees, the incidental acquisition of knowledge about the target language during collaboration, and tutors’ perception of exercises and tasks. Teachers of German as a FL took on the role of advisors and tutors in email exchanges with learners of German from Hong Kong and Wisconsin. Emails, interviews, the tutors’ diaries and data from the accompanying seminar were analyzed.

In Project 2, future primary school teachers of English attended the master's course ‘E-Lingo – Teaching Languages to Young Learners’, which consists of twelve modules on the development, implementation and analysis of a mini action research project. The participants were assigned to work in groups of three, with additional participants located in different parts of Germany and interacting via a virtual learning environment. Within this setting, the students were offered two channels of communication: a private forum for discussions and their own chat-room; individual and cooperative work were intertwined.

Data collected in both projects showed great variation in the role of the tutor. In addition, the computer-mediated communication tools available for cooperation were to some extent avoided, due to a perceived lack of privacy. The integration of digital media into teacher education environments raises new questions for research into the cooperative dimension. Legutke & Rösler (Reference Legutke and Rösler2005: 190) conclude that ‘forms of training will have to be found which enable future teachers to reflect on individual learning processes and which make cooperative learning an integral part of university learning’.

This description of a number of smaller research projects, especially in the field of blended learning, underlines the need for more empirical studies. Research has to take account of the practical experiences of learners and teachers and take new theoretical and conceptual ideas as a basis for empirical research that is both broader and more differentiated and situated, investigating, for example, the effectiveness of hybrid learning arrangements. Adjustments in the methodology seem necessary, including:

1. a) a better understanding and integration of students’ advanced prior knowledge right from the start, including surveys on their probable multi-modal knowledge access skills (via mobile phones, iPads and tablet computers, etc.) as well as their specific hybrid language use and language change which might be situated between different modes (from written to oral)

2. b) the triangulation of research instruments to obtain a clearer picture of the effects of hybrid learning scenarios, and

3. c) a greater focus on more detailed and sophisticated data analyses.