Print   |   Save

Coping With CLIL:
Dropouts from CLIL Streams in Germany

Carsten Apsel
Gymnasium Schenefeld / University of Hamburg (Germany)


Against the backdrop of EU and national policies in Germany promoting CLIL for an ever-expanding circle of learners in mainstream education, this article argues for an increased awareness of the factors conductive to learning in CLIL. Especially as CLIL in Germany has a distinct elitist tradition, such knowledge will be crucial to provide adequate CLIL environments to learners in the future who until recently have been regarded as non-eligible for CLIL streams. As one way to identify and investigate these factors, the author proposes research on so-called ‘dropout’ students who failed to succeed in the optional CLIL streams and ‘returned’ to traditional content matter teaching through German. He reports on an ongoing interview study with ‘dropouts’ and presents some preliminary results.


Keywords: CLIL (Content and Language Integrated Learning), dropouts, documentary method

1. Introduction

One of the political aims promoted in the educational policies of the EU can be described as ‘Taking CLIL further’. It has also been adopted in Germany by the Standing Conference of the Ministers of Education and Cultural Affairs (KMK). Such policies can be substantiated with reference to research findings on the positive effects of CLIL on language learning: On the European scale “research clearly shows that CLIL is suitable for students of varying levels of ability” […] and therefore “it is suggested that entry to the programme be granted on a first come first serve basis. In some countries demand outstrips availability and a lottery system is used.” (Mehisto et al., 2008: 23). Writing about the German context, Meyer is convinced that CLIL is a “success story” on the grounds that

[…] there is much evidence to suggest that CLIL students are equally, if not more successful at learning a subject than students learning content subjects in L1. This means that CLIL may be considered as an approach that is mutually beneficial for both content and language subjects (2010: 12).

While I do not intend to question the validity of the research behind Meyer’s claims, I consider that caution must be used when politicians use such claims to consider the widespread application of CLIL. My scepticism is based on what I perceive to be a distinct mismatch between the populations of CLIL students so far researched in Germany and other populations of students addressed by policies promoting “CLIL for all”. Since only little, if any attention at all has been given in German research to students who drop out1 of CLIL streams, the point I will argue in this article is that learning from these pupils’ experiences in times when the structure of the population of CLIL students is about to change may be of great importance and may indeed be one element of a “more rigorous and regular monitoring through a range of measures chosen to bring together a broader range of perspectives on CLIL programmes” (Coyle et al, 2010: 135).

To contextualise and substantiate my argument, I will outline the political aspirations for the future development of CLIL in some more detail as they are currently discussed on the European and German level before giving a short literature review of selected CLIL research in Germany (section two). In sections three and four, I will present the design and preliminary findings from my current research project on CLIL dropouts (sections three and four). My closing remarks (section five) will address the future perspectives for the further dissemination of CLIL in Germany in the light of the first tentative findings from my project.

Page 47
Page 48

2. CLIL – policies and research

In the following, I will discuss a number of political statements and research results in order to argue that while on the one hand “a gain in momentum is to be both celebrated and welcomed as the CLIL ‘explosion’ takes hold and embeds CLIL into the regular curriculum” (Coyle, 2007: 47), on the other hand a closer look is needed to prevent CLIL from raising hopes that cannot be fulfilled.

2.1 Political statements on the European and German level

The website of the European Commission on CLIL states that “CLIL is taking place and has been found to be effective in all sectors of education from primary through adult and higher education”. A list of positive effects of CLIL is given on the same page (EC, 2011):

CLIL's multi-faceted approach can offer a variety of benefits. It:

  • builds intercultural knowledge and understanding
  • develops intercultural communication skills
  • improves language competence and oral communication skills
  • develops multilingual interests and attitudes
  • provides opportunities to study content through different perspectives
  • allows learners more contact with the target language
  • does not require extra teaching hours
  • complements other subjects rather than competes with them
  • diversifies methods and forms of classroom practice
  • increases learners' motivation and confidence in both the language and the subject being taught

CLIL “can” offer the above mentioned variety of benefits, which implies that this may not always be the case. The statement does neither refer to any particular CLIL programme nor does it mention under which circumstances these benefits can be obtained. Essential information such as “the key issues […] how it has been made work, and which factors and events shaped success” are missing (Kiely, 2009: 114). Even though the above benefits are not put into context, “owing to its effectiveness and ability to motivate learners, CLIL is identified as a priority area in the Action plan for Language Learning and Linguistic diversity (Section 1 1.2)” (EC, 2011). On the European level, the political aim of “Taking CLIL further” is set and promoted by a number of measures to support the implementation of CLIL, such as EuroCLIC, a network providing “opportunities for contacting and learning from experienced CLIL practitioners.” (ibid.)

Against the backdrop of the European Commission’s aim, it is worth noting the German KMK’s view on CLIL. In Germany, CLIL has been seen as a ‘treat’ for learners perceived to be ‘more talented’ (KMK, 1999: 7). But as Breidbach (2007: 31) states, so far, there has been no research on how such students are channelled into CLIL streams in Germany. Breidbach finds evidence of the alleged selectivity, for example in Helfrich’s account of admission criteria to CLIL streams in a final report of a pilot project on CLIL in Rhineland-Palatinate (cf. Helfrich, 1998: 116f., 156f.; Breidbach, 2007: 31). As early as 1999, the KMK in an interim-report states that German CLIL streams have a reputation for being selective. However, the Standing Conference argues that due to the development of CLIL-related pedagogy and more open and flexible modes of learning, such as modules, more students with diverging abilities could be attracted to and catered for in CLIL in the future (KMK, 1999: 8). The question what specific competences learners should acquire and at what levels is raised in the paper but remains unanswered (ibid.). However, the question itself shows that a pending problem had been detected but seemingly was not considered suitable for public discussion, as Breidbach remarks (2007: 31). In the same vein, Wolff argues that an elitist bias has always been a part of the heritage of CLIL in Germany: In the 1990’s, when European Schools (Europaschulen)– catering mainly for the children of highly mobile EU-staff in teaching the curriculum through various official EU languages – served as a blueprint for CLIL streams, so that the expansion of CLIL could be seen as a democratisation of what was considered a quality language learning programme (Wolff, 1997: 45).

Despite these uncertainties, the KMK labels CLIL a success story (KMK, 2006: 25f.) on the grounds of practical experience: once introduced, CLIL streams are very rarely cancelled and attrition rates are said to be low, which means that only a small number of students drop out of CLIL courses in order to join the non-CLIL course of the same subject. However, no exact dropout rate is given and to my knowledge, no statistics about this phenomenon are officially available in Germany.

Page 48
Page 49

Based on their largely episodic evaluation, the KMK suggests that CLIL streams be further developed, that they be made more flexible and allow for more subjects to be taught through CLIL, and finally to have other school types in addition to grammar schools offer CLIL (KMK, 2006: 25f.). Clearly, these recommendations mirror a similar political aim as the EU-Commission’s to foster the expansion of CLIL and to provide students with greater access to CLIL.

Even though all sixteen German federal states formally endorse CLIL on the basis of the KMK’s 2006 recommendations (KMK, 2006: 93-95), specific steps on how to expanded CLIL programming so as to include more schools and students are lacking in the KMK papers. Individual federal states, school authorities or schools are left to find their own way forward. Three states pinpoint what could be CLIL’s prospective future: Lower Saxony aims to offer – within the existing budgetary constraints – CLIL at all schools, while Schleswig-Holstein and North-Rhine-Westphalia aim to offer CLIL to all pupils as a long-term goal (KMK, 2006: 93-95). Accordingly, North-Rhine-Westphalia (NRW, 2011) launched a programme “Bilingual für alle” (“CLIL for all”) in 2007.

2.2 CLIL research in Germany and its silent agenda of selectivity

In this section, I will discuss specific research results from the perspective of the German CLIL-context against the backdrop of policies to provide more students with access to CLIL. Based on this I will then argue that in Germany there is an imminent risk which Coyle describes in more general terms as follows: “As CLIL rapidly expands there is a danger that quality might be overtaken by quantity.” (Coyle, 2007: 56) And therefore, as Coyle suggests, “some essential conditions and requirements […] need to be in place […] such as a bottom-up approach […] and a learner involvement in the process: student ‘voice’.” (ibid.)

According to a number of studies, CLIL supports students in developing their verbal skills in their CLIL working language in absolute terms (e.g. Wode, 1994; Bredenböker, 2000; Schrandt, 2003; DESI, 2006; Zydatiß, 2007), and at no cost to content learning (e.g. Bonnet, 2004; Zydatiß, 2007; Heine, 2008; Badertscher, 2009). Occasionally, CLIL students even outperformed their non-CLIL peers (e.g. Koch, 2005). In the following, I will briefly outline and discuss the results of the studies mentioned and argue that the characteristics of the population of students participating in CLIL streams may account to a very large extent for the positive results found.

Wode’s study in the early 1990’s, conducted with students after two years of intensified English instruction and only seven months of CLIL in Grade 7, focused on the development of competencies in English in comparison to non-CLIL students. The study found CLIL students used the target language more frequently and in a more creative way than non-CLIL students (Wode, 1994: 76).
Bredenböker’s longitudinal study on CLIL and non-CLIL students, covering Grades 7 and 8, showed through C-Tests that CLIL students improved in particular their reading skills and their global CLIL working language competence. However, CLIL-learners only slightly improved their overall grammatical competence (Bredenböker, 2000: 93f.). The largest difference occurred when learners were asked to complete dialogues. Here, the difference between CLIL and non-CLIL students became most obvious, supporting Wode’s findings according to which CLIL contributes to learners’ communicative competence.

Both studies mentioned above were conducted at German grammar schools. Schrandt confirms these findings in her study, conducted at a German middle school, thus implying that CLIL could also be applied in other types of schools (Schrandt, 2003: 193).

A large-scale study conducted in Germany on 15-year-old teenagers showed CLIL-learners taught through English to be significantly more competent in the four basic skills than their 15-year-old non-CLIL peers, who only received traditional English language instruction and studied content-matter subjects through German (DESI, 2006). These effects are statistically robust even if meta-variables such as the socio-economic background of the learners’ parents are controlled for.

Zydatiß presents comparable findings in DEZIBEL (2007), comparing CLIL- and non-CLIL learners in Grade 10 at grammar schools in Berlin. Focusing on both language and content, Zydatiß finds that not only are CLIL students more competent than their non-CLIL counterparts in using the target language but their academic discourse competence is on par with students studying content through German  (Zydatiß, 2007: 377).

The same phenomenon is also revealed in Bonnet’s qualitative case study focusing on the acquisition of subject specific competencies (i.e. Chemistry, Grade 10, grammar school) in a CLIL environment. Bonnet finds that the working language of CLIL does not negatively affect the process of negotiation of meaning among students and even provides more opportunities to reflect on scientific terms (Bonnet, 2004: 288). Heine takes Bonnet’s considerations one step further in her study on the role of linguistic processing in problem-solving activities. She uses think-aloud-protocols “to reconstruct subject- and language-specific cognitive processes of German geography learners working on typical geography tasks in their L2 English.” (Heine, 2008: 213). She suggests that a lack of linguistic knowledge triggers additional problem-solving activity which “in turn leads to a deeper elaboration of content.” (Heine, 2008: 222; cf. Wolff, 1997: 48).

Page 49
Page 50

Badertscher also reports from his study, conducted with six classes of Grades 7 to 9 in Switzerland, with either German or French as CLIL working language, that no difference in content knowledge had been observed compared to learners not taught through CLIL thus raising the question how the indisputable language handicap of CLIL learners is overcome and what role language plays in learning in general (2009: 103). He suggests that contextual factors (initial euphoria of students and teachers and the video recordings) need to be taken into account, although their influence on the study remains unclear (2009: 183).

In her case study, in which lower secondary students of a science club where taught through CLIL, Koch also finds no indications of obstacles to subject matter learning through CLIL. Rather the increase in knowledge was found to be greater in the CLIL group than in the non-CLIL control group (Koch, 2005). Whether this was on the grounds of a deeper elaboration of content due to language deficiencies is left unanswered.

Yet, however successful CLIL appears to be, it should be noted that the populations tested may account for the positive results as these students can be deemed on some level to be part of an ‘elite’ group. As CLIL in Germany is non-compulsory and offered at selected schools only, students are subjected to a selective recruitment process, which in most cases includes the average grade in English as a central criterion (see also section 4). German schools traditionally offer CLIL streams in the form of one CLIL class/course2 per Grade and pupils are taught between one and three content matter subjects in the CLIL target language. It is commonly understood that only highly motivated students with good grades are enrolled in those courses (e.g. Bredenbröker, 2000: 50). According to Bredenböker (ibid.), CLIL students performed better in all tested parameters prior to beginning their CLIL courses. Thirty-three (33) of these highly motivated students of a total CLIL population of 113 dropped out of their CLIL courses during this two-year study (ibid.: 45). This equals an attrition rate of around 30%, each course having an average of 23 students at the beginning of year 7 and an average of 16 to17 at the end of year 8.

Hence, the studies mentioned above investigated the “beneficiaries” of CLIL, i.e. learners who at the point of testing were participating in CLIL. This is a noteworthy observation because in Germany, students have the statutory right to leave the CLIL stream at the beginning of each school year in order to attend subject matter teaching in German. CLIL streams in Germany are in fact doubly-selective: Not only are pupils selected on entry but there are exit doors for them to leave in between their CLIL careers. This means that, by and large, all students tested were taken from the group of CLIL-learners, rather than from a randomised sample of all potential CLIL students in Germany. It is, however, this much larger and probably far more heterogeneous group of learners which is taken into view by the EU- and KMK- policies to expand CLIL.

In a similar vein to the KMK documents discussed above, the issue of the sample used for CLIL-related empirical research and its implications for advocating CLIL for all is usually not a topic for discussion within the individual studies. There are, however, exceptions. In the early 1990s, when research on CLIL in Germany began, Butzkamm (1992: 25) found that the learners he tested represented a group of particularly high-achieving students since low-performing students were usually strongly advised not to take the CLIL class/course. In a later publication he raises the issue again by pointing out that methodological and pedagogical deficiencies are compensated for by high-achieving pupils and therefore a development in CLIL’s methods and didactics will be needed if CLIL is to be offered to a more diverse student population (Butzkamm, 1993: 159).

Zydatiß’ study in Berlin (DEZIBEL) reports rather stable numbers of students in CLIL classes, however specific dropout rates are not given (Zydatiß, 2007: 38). Zydatiß also emphasises that in those CLIL classes tested, most students were high achievers (ibid.). Students were admitted to the CLIL streams on the basis of their grades at the end of Grade 6 in German, Maths and English (the English grade was given double weight). Furthermore, Zydatiß states that the high scores of his sample, especially with respect to the use of academic language in content matter subjects, significantly correlate with the elevated socio-economic background of the learners (Zydatiß, 2007: 477). The same phenomenon had previously been discovered within the DESI-large-scale study (DESI, 2006: 59). If CLIL is opened up to more students, it is likely that the socio-economic background of students will be more diverse (Zydatiß, 2007: 477) and that this may affect the development of academic discourse competencies. Zydatiß clearly states that the results found in his study can only be generalised to a limited number of scenarios under very specified conditions.

Page 50
Page 51

3. “Coping with CLIL” - questions, theoretical framework, methodology and method

3.1 Taking a critical view behind the scenes of my own role in teaching CLIL

In 2004, I had the opportunity to teach a CLIL module at a grammar school in Hamburg, Germany, that offered a compulsory CLIL module for all Grade 10 students in Biology, using English as a working language. The evaluation of the module through action research methods revealed that some students’ subject-related performance in Biology had declined (Apsel, 2004). It turned out that student achievement in this CLIL-Biology module depended mainly on the students’ English language proficiency. Students with low language proficiency levels were at the end of the module more likely to underachieve in terms of subject matter competence. This is a plausible outcome with respect to Zydatiß’ identification of a linguistic threshold in CLIL (Zydatiß and Vockrodt-Scholz 2007 and Zydatiß in this volume), and it made me critically reflect on my own role in teaching CLIL. I had certainly underestimated the degree of language scaffolding necessary to encourage them to respond more frequently and thus to foster students’ language production, especially when working with and applying conceptual knowledge.

The module I taught in 2004 included all learners of one class without any previous selection. In hindsight, this experience provides an example of the possible consequences of offering CLIL to all students. A questionnaire survey in this class revealed that students facing difficulties and failing to succeed in the module would have preferred to opt out of CLIL in order to retain their previous grade in the subject. From their perspective, CLIL – even if only a module – did not prove to be a success story.

I became aware of the phenomenon of dropouts next when in 2009, a number of learners from the CLIL stream of the school I worked for at that time requested to be allowed into the non-CLIL stream. I did not expect learners who had received intensified English input via one additional weekly hour in Grades 5 and 6 and had gone through a selection process3 to wish to voluntarily opt out of CLIL. For these learners, it seemed implausible to me that they claimed ‘demotion’ on the basis of a low proficiency level in English alone.

So, why do learners cope with or drop out of CLIL? Obviously, a variety of factors has to be considered for an answer including selection processes and language proficiency and, last but not least, the quality of teaching. These considerations also made me critically reflect on my role in CLIL teaching in my early days as a teacher.

3.2 Research Questions

Seeing the political aims and the institutional constraints as the given base on which CLIL operates in Germany, it appears timely to investigate the necessary provisions that will allow for CLIL to be suitable for all school learners. The research study which I will present aims at reconstructing some of the preconditions for CLIL through the perspective of those learners who dropped out of CLIL streams. I investigated factors which caused pupils to drop out of CLIL streams and whether there was a mismatch between student needs and their CLIL environment. On the basis of this evidence, I will explore in a tentative move measures on the structural level in school, and more specifically on the practical level in CLIL, which may help to decrease the dropout rates from CLIL streams in German grammar schools.

Within the field of CLIL, two major stakeholder perspectives can be distinguished: the perspective taken by providers of CLIL, such as educational authorities or schools, and the perspective of CLIL practitioners taken by teachers or pupils. Within this spectrum, my study focuses on students as stakeholders and seeks to explore their experiences and perspectives on CLIL. The sample is limited to students who dropped out of CLIL programmes from one of the 46 grammar schools running CLIL streams in the two German federal states of Schleswig-Holstein and Hamburg.

Page 51
Page 52

The questions pursued in my project may be categorised as survey and research questions.

Survey questions

a) What is the CLIL dropout rate from grammar schools (Gymnasium) in Schleswig-Holstein and Hamburg?

b) Which concepts of CLIL and modes of selection are in place at the schools identified?

Research questions

c) Which factors are responsible for pupils dropping out of CLIL streams/programmes?

d) What are the underlying concepts and frames of reference for dropouts to conceptualise CLIL and the context of their retreat from CLIL?

3.3 Theoretical framework and methodology

The survey questions a) and b) are intended to identify the population from which to generate the sample of potential cases for the main study. This data is quantitative in nature. Research questions c) and d) require a qualitative approach with a flexible, explorative design.

In order to be able to describe the complexity and development of motivational processes leading learners to drop out of CLIL streams, I consider the emphasis on the participants’ individual perspectives on teaching and learning as it can be found in Bildungsgangforschung4 as the appropriate approach to my research topic. I also hope to be able to discover motivational patterns which studies such as DESI and DEZIBEL through taking a test-based approach have neither been able to detect nor to account for.

3.4 Methods and procedures

For an overview of existing CLIL grammar schools, I requested an official list from the ministries of education in Schleswig-Holstein and Hamburg. These lists were cross-checked in a web-search through which I could identify further CLIL grammar schools. Information about the different CLIL concepts (modules, streams or whole school policy5, modes of student selection and school CLIL coordinators6 had to be gathered through the schools' websites, personal phone-calls and/or e-mail. Since no official lists of dropouts are kept at the schools in Schleswig-Holstein and Hamburg, the respondents had to rely on their memory. My main focus was on the academic years 2009/2010 and 2010/2011.

For the research questions c) and d) I conducted episodic/narrative interviews with former CLIL learners. This format allows for the students to elaborate freely on what is relevant to them (Trautmann, 2010; Helfferich, 2009). Each interview is supplemented with a background questionnaire. The interviews were open at the beginning and then continued with topic cards that were successively put onto the table in front of the students (Trautmann, 2010: 177). Topics on these cards are aspects of teaching and learning in general (Grob et al., 2003: 94) and CLIL in particular. They gave the students the chance to decide for themselves how to shape their interview and helped them to keep their narration going. Interventions from my part were reduced to a minimum, when things were not understandable or when I was interested in obtaining further information. All interviews were videotaped and will be transcribed for further analysis.

Page 52
Page 53

All interviews will be analysed using the documentary method (Bohnsack et al., 2007; Bonnet in this issue). I consider the documentary method as suitable for two reasons: first, it is open in respect to content and relevance (Bohnsack, 2010: 32); second, it focuses on both the content and the way the content is formulated and arranged (Bohnsack, 2005: 73). This seems particularly appropriate for my interviews because I expect the stakeholders’ perspectives on CLIL to be elaborated explicitly by the interviewees but also through narrated experiences based on certain underlying concepts which in turn need to be methodologically reconstructed (Nohl, 2009: 8). The documentary method seeks to identify the underlying concepts (frames of reference – cf. Bonnet in this issue) through iterative analyses of interview transcripts (Nohl, 2009: 45).

The selection of cases from the sample for such an analysis will be made through theoretical sampling on the basis of the similarities or differences that are found in cross-case analyses, leading to a qualitative sampling design (Kelle et al., 2010: 50, 51; Flick, 2008: 297). A pilot study has been completed for one grammar school and interviews at other schools are currently being carried out.

4. Preliminary results from the survey and pilot study

4.1 The survey

The survey established an overview of all grammar schools with a CLIL stream in Schleswig-Holstein (SH) and Hamburg (HH). At present a total of 26 grammar schools in SH and 20 in HH offer CLIL, which is about a quarter of all grammar schools in the two federal states. However, there are considerable differences in the CLIL format offered. Most schools offer CLIL streams in the form of one CLIL class per year (usually out of three or four parallel classes per Grade). Whole school policies are found only occasionally: one school in SH and two schools in HH offer the CLIL programme to all students beginning in Grade 7 with an increasing number of CLIL subjects every year. Two schools in SH teach all non-language subjects for all students in all Grades through Danish. In Hamburg, one school teaches all non-language subjects in English for all students in all Grades and another one teaches all such subjects in French for all students in all Grades. The concept of modules is only found at one school in SH and at three schools in HH. The survey includes only schools which mentioned the concept of CLIL-modules in the school’s profile published on the internet.

In most cases, English is the language used for CLIL, but other languages such as French are also used. In SH, there are 23 schools using English, one school using French and two schools using Danish. In HH, there are 14 schools using English, three using French, one each using Chinese, Spanish, and Italian. History and Geography are the most common CLIL subjects, followed by far lower numbers for subjects such as Drama, Economics, Biology, Art, Physical Education, and Chemistry. In general, all schools provide one additional hour per week of English as a foreign language (EFL) in Grades 5 and 6. While schools offering a whole school policy or modules do not have regulations for admission to CLIL, schools offering CLIL streams usually do. However, these regulations vary from school to school.

I contacted the CLIL coordinators at 26 schools in SH either through e-mail or telephone. Thirteen coordinators from different schools responded, with nine of them reporting on dropout rates. Out of these nine schools, there were six with and three without dropouts. In HH, I was able to speak to 11 CLIL coordinators out of 20 CLIL schools altogether. Only six reported on dropout rates. Out of these six schools, four reported having and two as not having dropouts. Most schools with dropouts reported having between one to two cases per year. The average annual attrition rate at schools with reported dropouts varies from 3% to 8%. Most dropout cases occur in the first and second year of CLIL (Grade 7 and 8). The annual attrition rates I found are considerably lower compared to 7% to 16% as reported by Bredenbröker, who also covered a two-year span (Bredenböker, 2000: 45).

Page 53
Page 54

On the basis of the CLIL-coordinators’ reports, the survey also cast some light on the variety of admission procedures. One of the following seven regimes can be found at grammar schools offering CLIL in SH and HH:

  • extra hours in English (EFL) in Grades 5 and 6 for all students followed by a lottery at the end of Grade 6 for students applying for the CLIL stream in Grade 7,
  • extra hours in English (EFL) in Grades 5 and 6 for all students, admission by teacher recommendation only at the end of Grade 6,
  • extra hours in English (EFL) in Grades 5 and 6 for all students, admission by teacher recommendation and grades at the end of Grade 6,
  • extra hours in English (EFL) in Grades 5 and 6 for all students, admission by selection according to grades at the end of Grade 6,
  • extra hours in English (EFL) Grades 5 and 6 for all students who then have to sign up for a mandatory CLIL stream starting in Grade 7, a system that is called “whole school policy”, admission usually requires a grammar school recommendation at the end of Grade 4; these schools strongly advise parents that pupils hold a recommendation for grammar schools issued by the primary school7,
  • extra hours in English (EFL) in Grades 5 and 6 for all students, unrestricted admission for students voluntarily signing up for the CLIL stream in Grade 7,
  • extra hours in English (EFL) in Grades 5 and 6 only for students signing up early for the CLIL programme starting in Grade 7 with no limit in numbers.

4.2 The pilot study

The pilot study was conducted at a school witha relatively high dropout rate (four out of 30 per year as compared to one or two out of 30 per year at other schools). This triggered an investigation of the contextual factors at this school. At first sight, it seemed to be the school’s mode of selection that distinguished it from the other schools. Here, a ranking system by grade only is applied in which the grade in English is given twice the weight of the grade in the prospective CLIL subject.

The pilot study was then run with seven students (dropouts). The students I interviewed had been taught history through CLIL and thus their History grade had been relevant in Grade 6 for the students’ admission to the CLIL stream.

A content analysis of the interviews shows that pupils who dropped out of CLIL report facing difficulties with texts and with the learning of vocabulary. They reported on motivational problems in general, missing a sense of closeness to the teacher in charge, anxieties about losing their previously good grades and their wish to better understand what was being taught in history. A comparative analysis of the interviews indicates that these pupils have some of the following characteristics in common: their motivation, their ability to act as autonomous learners and their capacity to structure information. These findings suggest that grades alone may not be a reliable or sufficient criterion for admitting students to CLIL programmes. The system of admission to the CLIL stream found in this school thus appears to be inadequate.

With respect to other modes of selection found in the survey, this can hardly be expected to be a singular problem. In most schools with dropouts, grades play a (central) role for admission to CLIL streams. It may therefore be advisable to research the systems of admission to CLIL and their effects on learning in more detail.

Another implication from these preliminary findings addresses the quality of teaching. Especially learners with comparatively low language proficiency and cognitive academic abilities may not be catered for sufficiently in conventional CLIL streams at grammar schools in Germany. Looking at political aspirations such as “CLIL for all”, it seems advisable for CLIL teachers to acquire diagnostic, pedagogical and didactic competences to identify and support students across the entire range of ability. The initial results indicate that a combination of factors including language, learning skills, content knowledge and commitment seem to function as a complex threshold-level for CLIL reaching beyond discourse competence alone. The interrelationship of these factors has, however, not yet been fully understood.

Page 54
Page 55

5. Outlook

Kiely, with respect to the evaluation of language learning programmes writes that the “search for the secret of effectiveness has often focussed on the nature of the programme inputs and the measurements of outcomes” (Kiely, 2009: 114). Against this, he considers it far more advisable “to examine the interactions and factors which shape input use and impact” (ibid.). By the same token, Zydatiß (2007) gives a clear warning that the positive results produced from his DEZIBEL-study with CLIL learners in Berlin may heavily depend on the existing conditions at the time of testing and that the spreading of the CLIL model in a new context of budgetary cuts may not lead to the same positive results.

The study presented in the second part of this paper takes a first step towards analysing the factors which may function as the necessary preconditions for successful CLIL implementation for non-selected groups of learners the CLIL. The aim of the study is to investigate these factors through the perspective of dropout students who withdrew from CLIL after having had some experience with it. At the time of writing, results from a pilot interview study point at ways in which CLIL can be made accessible to all students. There is evidence that the linguistic threshold hypothesis can be applied successfully to CLIL but at the same time it becomes evident that the level of language competence, taken as the only determining variable, is neither able to account for the dropout phenomenon alone, nor will it be a sufficient predictor of success in CLIL.


Apsel, C.: 2004, Ein bilingualer Unterrichtsversuch zum Thema „Augen“ in einer 10. Klasse, schriftliche Hausarbeit zur pädagogischen Prüfung an Gymnasien, IQSH – Regionalseminar West, Abteilung für Gymnasien.

Badertscher, H. and Bieri, T.: 2009, Wissenserwerb im Content and Language Integrated Learning: Empirische Befunde und Interpretationen. Haupt, Bern.

Bohnsack, R.: 2005, Standards nicht-standardisierter Forschung in den Erziehungs- und Sozialwissenschaften, Zeitschrift für Erziehungswissenschaft 7(4), 65-83.

Bohnsack, R., Nentwig-Gesemann, I. and Nohl, A.-M. (eds.): 2007, Die dokumentarische Methode und ihre Forschungspraxis. Grundlagen qualitativer Sozialforschung, 2. erweiterte und aktualisierte Auflage. VS Verlag, Wiesbaden.

Bohnsack, R.: 2010, Rekonstruktive Sozialforschung – Einführung in qualitative Methoden, 8. Auflage. Barbara Budrich, Opladen & Farmington Hills, MI.

Bonnet, A.: 2004, Chemie im bilingualen Unterricht: Kompetenzerwerb durch Interaktion. Leske & Budrich, Opladen.

Bredenbröker, W.: 2000, Förderung der fremdsprachlichen Kompetenz durch den bilingualen Unterricht. Empirische Untersuchungen. Peter Lang, Frankfurt am Main.

Breidbach, S.: 2007, Bildung, Kultur, Wissenschaft. Reflexive Didaktik für den bilingualen Sachfachunterricht. Waxmann, Münster.

Butzkamm, W.: 1992, Zur Methodik des Unterrichts an bilingualen Zweigen, Zeitschrift für Fremdsprachenforschung 3/1, 8-30.

Butzkamm, W.: 1993, Bilingualer Unterricht – Fragen an die Forschung, DNS 92/1,2: 151-61
Coyle, D.: 2007, The CLIL quality challenge, in D, Marsh and D. Wolff (eds.), Diverse Contexts-Converging Goals: CLIL in Europe. Peter Lang, Frankfurt am Main

Coyle, D., Hood, P. and Marsh, D.: 2010, CLIL – Content and Language Integrating Learning. Cambridge University Press. Cambridge.

DESI: 2006, Unterricht und Kompetenzerwerb in Deutsch und Englisch. Zentrale Befunde der Studie Deutsch Englisch Schülerleistungen International (DESI). Deutsches Institut für Internationale Pädagogische Forschung [DIPF], Frankfurt am Main.

Dreher, E. and Dreher, M.: 1985, Wahrnehmung und Bewältigung von Entwicklungsaufgaben im Jugendalter: Fragen, Ergebnisse und Hypothesen zum Konzept einer Entwicklungs- und     Pädagogischen Psychologie des Jugendalters, in R. Oerter (ed.), Lebensbewältigung im Jugendalter, (30-61). Edition Psychologie VCH, Weinheim.

European Commission (EC): 2011. (

Flick, U.: 2008, Qualitative Forschung, 6. Auflage. Rowohlt Taschenbuch Verlag, Reinbek bei Hamburg.

Grob, A. and Jaschinski, U.: 2003, Erwachsen werden – Entwicklungs-Psychologie des Jungendalters. Beltz Verlag, Berlin.

Havighurst, Robert J.: 1948/1972, Developmental tasks and Education, 3. Auflage. David McKay Company, New York.

Page 55
Page 56

Heine, L.: 2008, Task-based cognition of bilingual learners in subject-specific contexts, in J. Eckerth and S. Siekmann (eds.), Research on task-based language learning and teaching. Theoretical, methodological and pedagogical perspectives, (203-226). Peter Lang, Frankfurt and New York.

Helfferich, C.: 2009, Die Qualität qualitativer Daten – Manual für die Durchführung qualitativer Interviews, 3. Auflage. VS Verlag, Wiesbaden.

Helfrich, H.: 1998: Der Modellversuch aus Sicht der wissenschaftlichen Begleitung für Englisch, In: Pädagogisches Zentrum Rheinland-Pfalz (ed.), Modelversuch 206 A: Entwicklung und Erprobung eines didaktischen Konzeptes für Bilingualen Unterricht an Hauptschulen und Realschulen. Abschlussbericht. Pädagogik Zeitgemäß 29. Bad Kreuznach. 106 - 178

Kelle, U. and Kluge S.: 2010, Vom Einzelfall zum TypusFallvergleich und Fallkontrastierung in der qualitativen Sozialforschung, 2. Auflage. VS Verlag, Wiesbaden.

Kiely, R.: 2009, Small answers to the big question: Learning from language programme evaluation, Language Teaching Research 13/1, 99-116.

KMK: 1999, Konzepte für den bilingualen Unterricht – Erfahrungsbericht und Vorschläge zur Weiterentwicklung. Bericht des Schulausschusses vom 04.01.1999. Bonn

KMK: 2006, Konzepte für den bilingualen Unterricht – Erfahrungsbericht und Vorschläge zur Weiterentwicklung.

Koch, A.: 2005, Bilingualer naturwissenschaftlicher Anfangsunterricht: Entwicklung – Erprobung –Evaluation. (

Mehisto, P. and Marsh, D and Frigols, Maria Jesus: 2009, Uncovering CLIL – Content and Language Integrating Learning in Bilingual and Multilingual Education, Hueber Verlag, München

Meyer, O.: 2010, Towards quality-CLIL: successful planning and teaching strategies, Pulso: Revista de Educación 33, 12.

Nohl, A.-M.: 2009, Interview und dokumentarische Methode – Anleitungen für die Forschungspraxis, 3. Auflage. VS Verlag, Wiesbaden.

Ministerium für Schule und Weiterbildung des Landes Nordrhein-Westfalen (NRW): 2011.
( ...,

Richter, R.: 2004, And now – in English, please, Unterricht Biologie 298/2004, 7.

Schrandt, C.: 2003, Evaluierung des bilingualen Unterrichts an Berliner Realschulen. Diss., Berlin.

Trautmann, T.: 2010, Interviews mit Kindern. Grundlagen, Techniken, Besonderheiten, Beispiele, 1. Auflage. VS Verlag, Wiesbaden.

Wode, H.: 1994, Bilinguale Unterrichtserprobung in Schleswig-Holstein. 1&f Verlag, Kiel.

Wolff, D.: 1997, Bilingualer Sachfachunterricht: Versuch einer lernpsychologischen und fachdidaktischen Begründung. In: Freie und Hansestadt Hamburg (ed.): Initiativen für den Fremdsprachenunterricht an Hamburger Gymnasien. Dokumentation der Fachtagung vom 23. – 26. April 1997. Hamburg: Behörde für Schule, Jugend und Berufsausbildung, Amt für Schule

Zydatiß, W.: 2007, Deutsch-Englische Züge in Berlin (DEZIBEL).Eine Evaluation des bilingualen
Sach­fachunterrichts an Gymnasien: Kontext, Kompetenzen, Konsequenzen. Peter Lang, Frankfurt am Main.

Zydatiß, W. and Vockrodt-Scholz, V.: 2007, Zur Interdependenz von Englisch- und Sachfach­kompetenzen im bilingualen Unterricht – ein empirischer Beitrag zur `Schwellen-Hypothese`, Zeitschrift für Fremdsprachenforschung (ZFF) 18/2, 209-236.


1 I use the term ‘dropout’ to denote students who leave a CLIL stream to receive further subject matter teaching in German at the same school. Students who drop out of CLIL streams for other reasons such as moving away are not considered as dropouts in this sense.
return to text

2 Cf. Richter (2004) for an overview of the types of CLIL offered in German schools.
return to text

3 Here, a ranking system by grade was applied in which the grade in English was given twice the weight of the grade in the prospective CLIL subject.
return to text

4 Bildungsgangforschung (research on educational experience and learner development)focuses on the individual’s perspectives onlearning and teaching (Bonnet, 2004: 23; Wegner in this issue). Within this branch of educational research the balancing of learners’ needs and aspirations on the one hand and societal demands on the other hand are addressed through the concept of developmental tasks (Havighurst, 1948/1972; Dreher and Dreher, 1985).
return to text

5 A Whole school policy this implies that all students at a school are taught through CLIL, starting in Grade 7 first in one, later then in two or sometimes more subjects. There is no alternative of taking a non-CLIL course in the subject/s that is/that are taught via CLIL at this school.
return to text

6 Coordinators at CLIL schools are responsible for the CLIL concept, implementation and management. Only a few schools have an official position for CLIL coordinators. In most cases, this work is done by CLIL teachers.
return to text

7 In SH and HH, a recommendation for a particular type of secondary school is a mandatory element in the final report students receive from primary schools at the end of Grade 4.
return to text