Self-regulated cooperative EFL reading tasks: students’ strategy use and teachers’ support

The ADEQUA research project has gained empirical evidence on how the situationally adequate use of learning strategies can be facilitated during cooperative reading tasks in the EFL (English as a foreign language) classroom. Two video studies were conducted with ninth-grade EFL learners in German schools: the first (laboratory) study investigated the students’ use of strategies while working in dyads and without teacher support on a given task. The second study, a field study, focused on teachers’ actions to support their students while working on a series of tasks in their regular classrooms. In this paper, we present the findings from a specific subsample of students (n = 30 from the first study and n = 228 from the second one), focusing on (1) the extent to which the students employed specific strategies adequately and successfully, and (2) the types of support actions taken by the teachers and to what extent these actions facilitated the students’ strategy use. The microanalytic approach adopted here allows us to identify those strategies which especially appear to require a teacher's support in order to be employed more adequately and successfully. Furthermore, by distinguishing between teachers’ support actions which are more versus less conducive to self-regulation and facilitating students’ strategy use, we are able to provide recommendations on how to fine-tune teachers’ assistance.

In second language (L2) and foreign language classrooms across the world, reading in another language plays a crucial role. Providing plenty of linguistic and cultural learning opportunities is key to learners' L2 development. Learners need to be exposed to a rich lexical and structural environment which triggers their active construction of meaning (Finkbeiner, 2005). Depending on whether L2 texts are simply used for talk on the text or also for talk on thinking about the text, learners' language and cultural awareness might be enhanced. Furthermore, reading can serve as a springboard for oral interaction and writing (Aebersold & Field, 1997;Carrell, 1998;Finkbeiner, 2005;Nuttall, 1996;Rasinski, 2004).
Reading authentic texts -which have been written for a native speaker audiencepresents a challenge to both L2 learners and L2 teachers alike. While L2 learners need to cope with unfamiliar vocabulary, syntax, and/or cultural scripts and schemata, L2 teachers need to devise means of effectively supporting their students in this endeavour. The latter becomes even more crucial when students are supposed to tackle an authentic text as autonomous learners. At this point, the notions of self-regulated learning and learning strategies come into play, both of which aim at equipping learners with the skills necessary to plan, execute, monitor, and evaluate their learning processes (Boekaerts, 1997;Cohen, 1998;Cohen & Macaro, 2007;Finkbeiner, 1995Finkbeiner, , 2005Helmke et al., 2008;Nold, 2009;Nold, Haudeck, & Schnaitmann, 1997;O'Malley & Chamot, 1990;Oxford, 1990). These skills -whose acquisition is typically a long-lasting, non-linear process, during which the teacher's support can play a crucial role -are considered essential for sustained, lifelong learning in today's information and knowledge society.
At first glance, we seem to be looking at a pedagogical paradox here, namely teacher support and intervention in autonomous learning. However, recent research on self-regulated learning has acknowledged the importance of providing appropriate support to learners in self-regulated settings. This is also in line with sociocultural theories of (language) learning (e.g. Lantolf & Poehner, 2008;Lantolf & Thorne, 2007) and, in particular, the notion of scaffolding within learners' zone of proximal development (ZPD) (Finkbeiner, 2011;van Lier, 2004;Vygotsky, 1978).
This paper reports on selected findings from the ADEQUA research project 1 (Finkbeiner, Ludwig, Wilden, & Knierim, 2006;Finkbeiner, Knierim, Ludwig, & Wilden, 2008), which builds on the premises described above. The project has gathered empirical evidence of how autonomous learning in the English as a foreign language (EFL) classroom can effectively be supported during cooperative, text-based classroom activities by means of teacher support actions. The intended goal is to facilitate the learners' situationally adequate as well as individually adequate use of strategies.
The project's findings will be integrated into a 'support action model (SAM) for cooperative, text-based classroom activities in the EFL classroom'. This model will guide teachers in diagnosing reading comprehension difficulties and subsequently employing support actions so as to maximally maintain the students' existing autonomy or support them in the process towards becoming autonomous learners.
In the following, we will elaborate on the theoretical underpinnings of the ADEQUA research project, the distinction between adequate and successful strategy use, and the distinction between adequate and successful teacher support actions. Boekaerts (1999) sees the ability to regulate one's own learning process as the key for successful learning in school and beyond. Students who regulate their own learning process successfully engage in analysing the task at hand, setting specific learning goals and using different strategies to plan, execute, monitor, and evaluate their learning process (Butler, 2002;Zimmerman, 2002). Moreover, self-regulated learners are aware of their inner resources: They know about their motivation towards the domain to study, they know about their expertise, their skills, and adequate strategies to fulfill the task successfully. This relates back to the concept of the 'good language learner', as coined by Rubin (1975) and re-visited by Griffiths (2008), Ushioda (2008), Anderson (2008), Finkbeiner (2008), and others.

Self-regulated learning
Self-regulated learning is seen as a complex process involving cognitive, metacognitive, and motivational aspects of learning (Artelt, Demmrich, & Baumert, 2001). The adequate use of learning strategies is considered crucial for successful self-regulation. Finkbeiner (2005) sheds light on the intricate multi-factorial interplay of interest, reading strategies, and self-regulated learning and its impact on deep processing while reading. She investigates 287 ninth graders with respect to their reading strategies, reading interest, and deep-process reading in EFL. Findings from structural equation modelling give evidence of the fact that self-regulated learning is a most crucial factor with respect to deep processing.
To promote self-regulated learning, teachers need to provide ample opportunities for working on cognitively challenging tasks, which make students choose, use, combine, and coordinate different learning strategies (Boekaerts, 1997). Depending on the level of their self-regulatory competence, students need some degree of external regulation while working on a learning activity (Boekaerts, 1999;Vermunt & Verloop, 1999). Butler (2002) states that the promotion of self-regulation should aim at (1) developing students' ability to analyse tasks and (2) setting task-specific goals. Moreover, teachers ought to help their students become aware of their metacognitive knowledge about learning and learning strategies and foster their monitoring and evaluation of their learning process.
For Vygotsky (1978), learning takes place in the ZPD, which is defined as the distance between students' actual developmental level and their potential level of competence. That is, students can master tasks that go beyond their current level of competence if an adult, for example, a teacher, and/or a more capable peer, provides individual support to them.
The concept of scaffolding describes best the type of individual support that a teacher can provide to students. Hammond and Gibbons describe scaffolding as 'support that is designed to provide the assistance necessary to enable learners to accomplish tasks and develop understandings that they would not quite be able to manage on their own' (2001, p. 15). Scaffolding includes teacher actions aiming at reducing students' frustration and fostering their motivation, indicating mistakes as well as modelling ways of problemsolving (Wood, Bruner, & Ross, 1976). As the students' skills improve, the teacher reduces scaffolding gradually in order to foster students' autonomy.
The type of scaffolding and the degree of external regulation depend on the students' needs. According to Meloth and Deering (1999), the type of teacher support during cooperative learning activities can vary from direct instruction during mini lessons to hardly any or no support. Moreover, they point out that the direct instruction and modelling of cognitive and metacognitive as well as cooperation strategies can benefit the regulation of students' learning process.

Strategies in L2 reading
L2 reading is a highly active cognitive and affective process which relies on the use of strategies, among other factors. It involves readers' constant construction of knowledge by means of hypothesising and testing (Finkbeiner, 2005). This process combines readers' knowledge about typical language-specific combinations of text markers on the surface level, such as letters, words, phrases, sentences, paragraphs, and whole texts, with content and semantic knowledge on the deep level (Erler & Finkbeiner, 2007). In this respect, L2 reading is both similar to and different from L1 reading. Similarities lie in the fact that meaning is not given in the text but constructed by the reader. However, differences might arise when L2 linguistic and/or cultural aspects diverge from the L1 knowledge the reader draws upon. Very often, readers are not aware of the fact that they interpret and construct meaning by drawing on existing knowledge that might not fully or even at all qualify as a basis for understanding the text. In this case, language awareness and cultural awareness (Fehling, 2005;Finkbeiner, 2009;James & Garrett, 1991) can be helpful to make readers re-think, re-evaluate, and re-structure their knowledge structures.
Strategies seem to play a decisive role in this complex interactive play of text and reader factors (Finkbeiner, 2005). Whereas text factors include components such as text type, text difficulty level, length, setting, and content, reader factors include prior knowledge, sociocultural background, verbal intelligence, reading proficiency level, cognitive ability, interest, motivation, etc. It is still not clear to which degree these factors can compensate for or even substitute each other.
Even with quite some research to be found on strategies (Cohen & Macaro, 2007;Finkbeiner, 2005;Nold, 2009;Nold et al., 1997), it is far from clear how best to investigate, diagnose, observe, and measure strategies, nor how to foster and elicit them. This is due to the fact that the strategy construct so far is still fuzzy. A lot of different approaches to describing strategies can be found. However, there seems to be an overall consensus that strategies can be defined within the framework of three continua: (1) from conscious to unconscious, (2) from observable to non-observable, and (c) from explicit to implicit (Finkbeiner, 2005).
Following O'Malley and Chamot's (1990) taxonomy, strategies can be divided into cognitive, metacognitive, and socio-affective language learning strategies. When readers apply cognitive strategies, for example, they highlight words, write down the most important words, look up words, or connect words or a passage to what they already know. These connection or elaboration strategies play an important role for textual deep processing (Finkbeiner, 2005(Finkbeiner, , 2006a. Readers applying metacognitive strategies think about planning, monitoring, and evaluating their reading comprehension. Socio-affective strategies or coping strategies include, for example, self-talk and cooperation. Although reading is a predominantly individual process (Finkbeiner, 2005), it can be vastly social and cooperative once certain pre-conditions are fulfilled. One of these is the creation of an anxiety-free learning atmosphere (Finkbeiner, 1995) that allows for peer learning, peer scaffolding, and co-construction of knowledge.

Adequate strategy use -successful strategy use
Recent research on learning strategies has pointed at the necessity of going beyond merely measuring frequencies of strategy use (Nold, 2009), since 'the total number or variety of strategies employed and the frequency with which any given strategy is used are not necessarily indicators of how successful [learners] will be on a language task' (Cohen, 1998, pp. 8-9). For example, some learners might employ several strategies in order to solve a specific comprehension problem but without success either (1) because they lack the knowledge or skills necessary to actually apply the chosen strategies or (2) because the chosen strategies are not applicable to the resolution of the problem at hand. Moreover, even if learners possess the abilities to employ strategies which can aid them in successfully tackling a specific comprehension problem, they may fail to resolve it nonetheless.
It is evident, therefore, that the analysis of students' strategy use warrants a more finegrained approach which takes into account the adequacy and the success of the strategies employed (in addition to their frequencies). As for adequate strategy use, we suggest considering two dimensions of adequacy, namely situation-adequate and learner-adequate strategy use. For clear and strong exemplary demonstration, we quote examples that do not fulfill the standard of adequacy: • It is not situationally adequate to infer the meaning of an unknown word if there are several unknown words in the same clause, which takes away the base for logical inferencing. • It is not situationally adequate to use a between-parts elaboration strategy if only a few, isolated fragments of the text have been comprehended or only parts are presented and the missing parts are not reliably communicated. • It is not situationally adequate to look up many words in a dictionary if the time is very limited.
• Looking up words in a dictionary may not be learner-adequate if the learner lacks (1) basic knowledge of word formation or (2) basic knowledge of the alphabet. • Asking another learner (repeatedly) for clarification may not be learner-adequate if that learner is a less knowledgeable peer. • Making use of elaboration strategies may not be learner-adequate if the learner cannot draw on prior knowledge that is actually relevant to the comprehension problem at hand. In such a case, elaboration might result in, for example, making up a (misleading) story line based on the comprehension of only isolated fragments of the text.
Fine-grained analyses of learners' adequate versus inadequate as well as successful versus unsuccessful strategy use are of immediate relevance for the development of L2 teachers' abilities to diagnose their students' situation-and learner-specific strategy use and, subsequently, to take appropriate actions in order to optimise their students' strategy use. These issues are pivotal to the efforts of the ADEQUA research project (Finkbeiner, Ludwig et al., 2006;, which will culminate in the development of the SAM mentioned above.
The notions discussed here are not completely new in language learning strategy research. Oxford, for instance, stated: A strategy is useful if the following conditions are present: (1) the strategy relates well to the L2 task at hand; (2) the student employs the strategy effectively and links it with other relevant strategies for doing the task; and (3) the strategy coordinates with the student's general learning style preferences to one degree or another. (2003, p. 274) The overlap with our terminology is apparent: A strategy which 'relates well to the L2 task at hand' might be labelled situation-adequate, and a strategy which is employed 'effectively' might be labelled successful. However, the conditions proposed by Oxford (2003) appear to be rather general and refer to the task as a whole, whereas the ADEQUA research project analyses the students' strategy use at a micro-analytic level, that is, with respect to specific comprehension problems as they arise during task performance. Similarly, a strategy which 'coordinates with the student's general learning style preferences' (Oxford 2003) might be labelled learner-adequate. In the ADEQUA research project, however, learner adequacy is determined against the specific knowledge and skills needed to resolve a specific comprehension problem rather than against learners' more general learning styles.
To summarise, the ADEQUA research project draws on two key insights from previous learning strategy research, namely the situation-as well as learner-specific nature of strategy use (e.g. Anderson, 1991;Finkbeiner, 2005;Garner, 1990;Helmke et al., 2008;Nold et al., 1997). Against this background, the ADEQUA research project adopts a microanalytic approach which seeks to address three criteria relating to the quality of strategy use: whether it is successful, situation-adequate, and learner-adequate.

Intervening in the use of strategies
Programmes for strategy instruction describe how teachers can introduce and practice language learning strategies with their students (Farrell, 2001). These programmes also provide information on how to design appropriate tasks that make students use different learning strategies. Current models of strategy instruction aim at raising students' awareness of their strategy use and emphasise the need for the introduction and modelling of new learning strategies by the teacher (Rubin, Chamot, Harris, & Anderson, 2007).
Strategy instruction as described above is primarily done in advance of or in addition to learning activities. In contrast to this, the ADEQUA research project focuses on how L2 teachers can promote their L2 students' awareness and use of learning strategies while they are working on a reading task. As described earlier in this paper, successful learning takes place in the ZPD (Finkbeiner, 2011;Vygotsky, 1978). Accordingly, it is the teacher's task to assist students in their use of learning strategies by means of scaffolding.

Criteria for adequate teacher support actions
In analogy to distinguishing between adequate and successful students' strategy use, the ADEQUA research project is also concerned with the adequacy and success of teacher support actions. Within the context of this project, teacher support actions refer to any kind of assistance provided by teachers to their students during specific cooperative, text-based classroom activities as designed for the study; further specifics as to the setting (tasks, texts) are mentioned next.
Given the primary goal of fostering and maintaining, as far as possible, the students' autonomy during task performance, a teacher support action is considered adequate if it meets the following criteria: • It is situation-adequate if it contributes to fostering self-regulation by addressing the use of or reflection on learning strategies and/or eliciting the use of learning strategies and/or addressing possible ways of solving a comprehension problem and/or aiming at motivating the learners. • It is learner-adequate if it is appropriate to the learners' L2 and general cognitive abilities.

Criteria for successful teacher support actions
Teachers' support actions are considered successful if they entail the resolution of the comprehension problem at hand (or a problem related to the task procedures) or if they provide sufficient input and stimuli for the students to resume their work on the task. However, this does not imply that teachers simply convey the meaning of a word, for example. Instead, they should support their students in finding the solution themselves.

Research questions
Based on the notions of adequacy and success with regard to learning strategies and teachers' support actions introduced above, we will address the following research questions: (1) Which strategies do students use in the given setting? To what extent are these strategies (a) adequate and (b) successful? (2) Which actions do teachers take to support their students in the given setting? To what extent are these actions (a) adequate and (b) successful?
Furthermore, we want to find out to what extent the teachers' support actions meet the students' needs in effectively (= adequately and successfully) employing learning strategies.

Participants
Overall, the ADEQUA research project comprises three major phases, all of which involve ninth-grade EFL learners from German secondary schools. Initially, pilot studies for all instruments and procedures used subsequently were conducted. Main Study 1 is a video study of learner dyads performing given L2 reading tasks in a laboratory setting without teacher support (n = 164 students, drawn -based on criterion and quota sampling -from a larger sample of n = 352 students). Finally, Main Study 2 is a video study of learner dyads performing a given series of L2 reading tasks in a regular classroom setting, allowing teachers to support their students while working on the task (n = 273 students).
Within the scope of this paper, we focus on two subsamples from Main Study 1 and Main Study 2. The students from these subsamples (see Table 1) worked on the same task based on the same text (see the Appendix), which allows us to relate the two data sets to one another.

Setting
Several task types and texts were used in various combinations at different stages of the ADEQUA research project . Due to space constraints, we will limit our subsequent discussion of learners' strategies and teachers' support actions to one particular setting, that is, the text Snow within the framework provided by the Click & Clunk task type (Bremer, Vaughn, Clapper, & Kim, 2002).
The text Snow is based on a chapter with the same title from Julia Alvarez's (1991) novel How the García Girls Lost Their Accents. It has been used in a L2 reading study with Table 1. Subsamples of ninth-grade EFL students in the ADEQUA project reading the text 'Snow' using the task type 'Click & Clunk'.

Subsample from Main
Study 1 (lab study) Subsample from Main Study 2 (field study) n = 10 n = 112 (c) Mean age 14.7 years 14.8 years (d) Mean number of years of EFL instruction 6.4 years 6.4 years Students' teachers -n = 11 Classes from which the sample was drawn n = 6 n = 12 Schools from which the sample was drawn n = 6 n = 8 * In most parts of Germany, secondary education is situated in a three-track system: The Gymnasium (Classes 5-13) leads students to a diploma that allows them to study at universities. Students graduating from Realschule (Classes 5-10) may, depending on their academic achievements, either start a job training including parttime enrolment in a vocational school or continue their education at a Gymnasium. Students graduating from Hauptschule (classes 5-9) may start a job training including part-time enrolment in a vocational school or continue their education at a Realschule. a similar sample before (Finkbeiner, 2005). For the ADEQUA study, it has been further abbreviated and simplified to render it more accessible -yet still challenging -for the study's target population. The narrative is set in the United States during the Cuban missile crisis and describes the events of the time -particularly, the threat of a nuclear war and the routine of air-raid drills -from the perspective of Yolanda, a young immigrant girl, who is struggling with the culture and language of her new country. One day in December, Yolanda is sitting at her desk in the classroom and looking out of the window when she sees the first snowflakes of the year falling. She mistakes them for nuclear fallout, since she has never seen snow in her life before, and panics. Only after her teacher explains to her that the white crystals falling from the sky are in fact snow, is Yolanda able to calm down again.
(The text, as it was used in the present study, is presented in the Appendix.) The text poses several challenges to the students: • Some unfamiliar lexis: for example, 'bonnet', 'peculiar', 'weapon', 'air-raid drill', 'to file' (in the sense of marching in single file), 'fallout', 'mushroom' (referring to a mushroom cloud after a nuclear explosion). • Some less familiar structures: subjunctive clauses ('we might have war', 'the radioactive material would kill us'). • Unfamiliar content: historical background of the Cuban missile crisis, characteristics of nuclear warfare. • Implied information: the students have to infer that Yolanda immigrated from a country in which there is no snow.
It should be noted that the students' familiarity with the greater amount of lexis, structures, and content of the text was ascertained based on the applicable curricula as well as textbooks used.
There is a strategic dichotomy: The challenges described above warrant the use of strategies as a means to overcome these potential comprehension problems. At the same time, however, the characteristics of these challenges constrain the use of certain strategies. For instance, it may prove difficult or risky for students to employ elaboration strategies in order to compensate for the lack of historical background knowledge; in such a case, the use of elaboration strategies may not be learner-adequate due to the student's lack of relevant prior knowledge on which he or she can draw (Finkbeiner, 2006b). This means that teachers first have to make sure that sufficient background knowledge is given, and whenever this is not the case, they should provide it before the task is assigned to the students.
The task types, which were developed for the ADEQUA research project, have four main characteristics in common: First, based on the principle of positive interdependence (Johnson & Johnson, 1994), they are designed for student pairs working collaboratively on a given text. Second, the tasks are supposed to elicit students' use of strategies which are deemed important in the given setting (e.g. summarisation, problem identification, inferencing, monitoring, questioning for clarification). Third, they are intended to provide a framework to structure and scaffold the students' reading processes; in this respect, the task types are generic and can be used with a large variety of texts. Fourth, each task contains a problem-oriented subtask which requires the students to use the main ideas of the text (instead of merely summarising them); in this respect, the task types include a text-specific component.
This paper focuses on a task type called Click & Clunk, which draws on the concept of collaborative strategic reading (CSR) (Bremer et al., 2002;Klingner & Vaughn, 1999). CSR rests on four principal reading comprehension strategies, namely previewing, 'click and clunk', getting the gist, and wrapping up (i.e. reflecting and evaluating) the reading process (Bremer et al., 2002).
As for 'click and clunk', a click is 'something that you really get . . .
[y]ou know it just clicks', whereas a clunk is 'like when you run into a brick wall. You just really don't understand a word the author is using' (Bremer et al., 2002, p. 3). Another feature of CSR is the use of roles (e.g. leader, clunk expert, gist expert, and announcer) which are assigned to the members of a group. Moreover, all group members receive cue cards to aid them in meeting their respective role responsibilities.
As can be seen from this brief description, CSR represents a comprehensive approach to cooperative reading, which necessitates a certain amount of teacher and learner training (Bremer et al., 2002, pp. 2-4) in order to take full effect. Since the ADEQUA research project aims at the development and investigation of task types that require as little formal a priori training of teachers and learners as possible, we decided to extract and simplify the core components of CSR, particularly the Click & Clunk procedure; also, our adaptation of CSR should be geared towards learner dyads (instead of small groups). The results of our efforts to 'boil down' the elaborate framework of CSR according to our needs and purposes are evident from the instructions and worksheets the students received (see Worksheets 1-3 in the Appendix). Furthermore, we produced a mini video in which two students modelled the procedure to fully standardise the instruction.

Procedure
The primary goal of Main Study 1 was to find out which strategies learners would use 'naturally' (i.e. without any particular training) in the given setting (see above). To this end, learner dyads were videotaped while working (without teacher support) on the given task in a laboratory setting outside of the classroom. After the task performance, the students participated in a stimulated recall procedure (Gass & Mackey, 2000). They were asked to view selected episodes from the video recording of their own task performance and retrospectively verbalise what they were thinking or doing during those episodes. In particular, the stimulated recall focused on episodes which the two observing members of the research team identified as opaque, for example, with regard to the students' resolution of a comprehension problem. Overall, the stimulated recall procedure provided valuable further insights into the students' thought processes.
Main Study 2 aimed at exploring the support actions teachers would provide 'naturally' (i.e. without any particular training) to their students while they were working on the given task. For this purpose, the teacher-student interaction in regular classroom settings was videotaped. Following the videotaped class period, stimulated recalls were conducted with the teacher and selected students to further elucidate the rationale of the teacher's support actions as well as the students' perceptions thereof. For both studies, the video recordings were transcribed and coded using the video analysis software Transana (Woods & Fassnacht, 2008). The coding scheme for the first main study was based on O'Malley and Chamot's (1990) taxonomy of language learning strategies, which had been used in earlier strategy studies (e.g. Finkbeiner, 2005;Nold et al., 1997). We refined the taxonomy in a data-driven manner according to the principles of qualitative content analysis (Ludwig, 2009;Mayring, 2008). The development process of the coding scheme for the second main study was primarily data-driven. Upon completion of the coding process using Transana, the data (i.e. code frequencies) were exported for further analysis with a statistical software package (SPSS). The analyses presented in this paper focus on descriptive statistics, that is, the absolute frequencies of the strategies (Main Study 1) and teacher support actions (Main Study 2) employed.
In the context of our transcribed protocols as well as student task sheets, we would like to mention the language issue. The students participating in the study were allowed to use the L1 or L2 according to their liking and convenience in completing the tasks as well as in conducting the learner-learner and learner-teacher talk. The reasons for this were that (1) we were interested in students' application of strategic knowledge in the given task and not in the way they could describe or articulate this strategic knowledge in the L2, (2) we did not want to confound the L2 reading comprehension results with L2 writing results, and (3) we wanted to avoid that a probably poor English language proficiency level would be an obstacle to meta-cognitive talk and strategy elicitation. This would not bias our findings as our focus was on finding out how students used specific strategies in order to understand the L2 texts and not in order to produce texts in the L2 (Finkbeiner, 2005).
It should be noted that in addition to the video data, data from student questionnaires and tests were collected, including demographic information, English language proficiency, interest in reading English texts, self-regulatory competence, task-based reading comprehension tests, etc. Based on these data, we were also able to ascertain the criterion validity of the study's core constructs (i.e. adequacy and success with regard to students' strategy use and teachers' support actions; Ludwig, Finkbeiner, & Knierim, 2009). However, a discussion of the findings emanating from these data sources is beyond the scope of this paper.

Research question 1: students' strategy use
To determine which strategies students use in the given setting and to what extent these strategies are (1) adequate and (2) successful, we examined the frequencies of strategies used obtained from our content analytic coding of the video material. Table 2 shows the mean values, the standard deviation (SD) as well as the frequencies of those strategies which the students, on average, used at least once during the task. In addition to the frequencies, the proportions of adequate and successful use are given for each strategy. The focus here is on situational adequacy and not on learner adequacy. A number of interesting observations can be made: Considering the challenges posed by the text (see above), the number of strategies used to tackle comprehension problems can be considered quite low. Even though we found 69 strategies altogether, only a few of them were used more than once in each session. Among the nine most frequently used strategies, there are six cognitive strategies (summarisation, resourcing, language-related world elaboration, inferencing, translation, content-related world elaboration), three socioaffective strategies (three types of questioning for clarification), but no metacognitive strategies (such as monitoring). On the other hand, two types of elaboration strategies, which are considered especially useful in promoting deep-level processing of a text (Finkbeiner, 2005), are among the top nine strategies used by the students. However, despite the fact that the elaboration strategies belong to the top 10 of the most frequently used strategies, the mean average shows us that the average use of each strategy is not really high: each one of the personal elaboration strategies (language-related and content-related) was only used once on average per session. Most of the strategies were used situationally adequately with proportions well above 90%. (It should be noted that this does not imply that the strategies were always learner adequate.) There is one noteworthy exception: Inferencing was used in a situationally adequate way 'only' in 87% of all cases. It is important to note that the proportions of successful strategy use are considerably lower than those of adequate strategy use. The success rates for content-related world elaboration and vocabulary-related questioning for clarification are particularly low. Here, we can see a discrepancy between frequency, success, and adequacy of strategy use. A mere look at frequencies does not help explain the success of the reading process.

Research question 2: teachers' support actions
The findings here focus on the actions teachers take to support their students in the given setting as well as to what extent these actions are (1) adequate and (2) successful. In the following, selected results and transcript excerpts of the second phase of ADEQUA are presented. In particular, we look at those teacher support actions that were identified most frequently and discuss two transcript excerpts. Table 3 lists those teacher support actions that were identified most frequently (mean per lesson) during a 45-minute English lesson. This list contains teacher support actions that, according to our definition, are considered adequate if they foster self-regulated learning (marked with an asterisk) and teacher support actions which are considered non-adequate if they do not foster self-regulated learning (no asterisk).
Without any specific prior training or coaching, teachers predominantly support their students by carrying out actions for them instead of supporting the students in carrying out these actions themselves, as the table shows. In particular, teachers summarise (parts of) the text and paraphrase or translate unknown words. Moreover, teachers use information from the text (text-immanent elaboration) to explain unknown words to the students and give explicit positive feedback extensively to the students.
In contrast, the number of teacher actions that aim at promoting students' autonomy and self-regulatory competencies is relatively low. Eliciting learning strategies barely takes place, and teachers hardly ever review the use of learning strategies and ways of problemsolving with their students.
We now turn to two examples of teacher-student conversations which illustrate two different approaches to supporting students' comprehension of the expression 'air-raid drills' in the second paragraph of the text Snow (see the Appendix). Transcript excerpts dealing with this compound have been chosen because (1) this compound is central for the understanding of important passages of the text Snow and (2) it was incomprehensible for many students.
The first excerpt (Table 4) shows a teacher's student-centered support action aiming at fostering students' self-regulatory competencies. In contrast, the second excerpt (Table 5) is an example of a more teacher-centered approach, exemplifying and underlining the findings discussed above.
During this student-teacher conversation, different teacher actions can be identified (see Table 4). First, the teacher asks the students whether they know parts of the compound 'air-raid drills'. This support action belongs to the category 'eliciting the cognitive strategy "transfer of morphological/lexical L2 knowledge"': When students apply this strategy, they use their foreign language (L2) knowledge to recognise the morphological structure of a compound. Asking the students the meaning of the individual parts of the compound might be seen as the teacher's effort to elicit the cognitive strategy translation. Finally, the teacher asks the students how they came up with their answers. The students answer by summarising parts of the text. As a result, the teacher's inquiry into the students' problemsolving procedure elicits the cognitive strategy discourse-level summarisation. To conclude, in this conversation, the teacher functions as a facilitator of students' mental actions. It is the students who find the answer themselves. The teacher encourages them to use and rely on their language knowledge and makes the students share their thinking processes, which may have a positive influence on students' awareness of their learning processes and autonomy.
In the second example (see Table 5), the teacher carries out most of the actions for the students. First, the teacher explains that a bilingual dictionary may contain different translations of one word. This advice might seem necessary, since the students have chosen a Table 4. Teacher support actions to disentangle the meaning of 'air-raid drills' -transcript excerpt 1.
Transcript Trigger for teacher's support action Description of teacher's support action Potential effect of teacher's support action on students' actions Student 1: 'What's the meaning for air-raid drills?' Student is uncertain about the meaning of 'air-raid drill'. The teacher asks S1 whether he knows the meaning of parts of the word.
-Student recognises the unknown word as a compound.
Teacher: 'Ok, do you have any idea? Are there any words you know alreadyany parts?' -Student identifies known lexical items, uses his foreign language (L2) knowledge, and suggests a translation. translation from the dictionary that does not fit into the context of the text. This information may have a positive influence on students' future use of the bilingual dictionary. Second, the teacher states that there is a word in German (i.e. the students' L1) which is similar to the English word drill. This advice can be interpreted as the teacher's attempt to elicit the cognitive strategy transfer of students' L1 morphological/lexical knowledge. Following this, the teacher paraphrases and translates the word drill and summarises parts of the text. Finally, the teacher asks the students whether they have understood her explanations. The students indicate that they know the meaning of air-raid drill. Without verifying this, the teacher leaves the students. This type of example is characteristic of most of the teacher support actions identified. Most teacher support actions are teacher-centered and instructional. The majority of teachers support their students by carrying out actions for them instead of eliciting strategies which might enable students to carry out the actions themselves and solve the problems more autonomously.
In the subsample examined here, 69% of all support actions were rated as successful and 5% as unsuccessful, whereas the success of 26% of all support actions could not be clearly ascertained. The latter figure attests to the fact that in a considerable number of instances, teachers did not verify the success of their assistance: If they had concluded their support actions by explicitly checking the students' comprehension, they would have received a clear indication of whether their help was fruitful. This, in turn, would have been observable in the video recordings analysed here.
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. So, we hardly ever observed metacognitive reflection phases in order to raise the students' awareness of how they tackled the comprehension problems, why or why not their approach was adequate, or which alternative approaches they might want to consider when reading other texts in the future. Yet, this is what counts and is crucial when it comes to the quality of teachers' assistance and students' success in learning. Overall, only a few teacher support actions could be identified as actions that foster students' self-regulated learning by means of supporting students' use of cognitive, metacognitive, and socio-affective learning strategies and by means of raising students' awareness of their learning processes.

Discussion
Pertaining to the strategy use by the students in Main Study 1, we would like to highlight the following insights and their possible implications: • A major proportion of the most frequent strategies was elicited by the task type (as intended; e.g. questioning for clarification, summarisation, resourcing, inferencing). At the same time, the elicitation of other strategies which may be desirable in this context (e.g. monitoring, between-parts elaboration) probably needs to be more explicitly built into the task as a trigger. Generally speaking, this result appears to corroborate the assumption that specific strategies may indeed be triggered by task specifications.
(However, this statement should not be misread as implying some direct or universal relationship between task design and students' learning processes.) • Given the novelty of the task type, the number of questions for clarification relating to the task procedures was pretty low: this indicates that instructions and worksheets were fairly easy to use -even without any particular, lengthy training -and well designed.
• There was a considerable difference between the success rates of language-related versus content-related world elaboration; the former may be argued to be less demanding, whereas the latter is more closely dependent on text-specific content knowledge (which, in the case of the text Snow, the students might not have had at their disposal). Especially with regard to teaching English as a lingua franca, the presence or even ubiquity of the English language in the everyday lives of EFL/ESL (English as a second language) learners across the world can be harnessed for their benefit by raising their awareness of cross-linguistic similarities and differences as they relate to one's experiences and world knowledge. • Vocabulary-related questioning for clarification was used relatively unsuccessfully. A possible reason is that it might be a less demanding strategy which can be very quickly used. Some students used this strategy in a highly formulaic, ritual-like fashion. They should be made aware of a more goal-directed, sparing use (e.g. to ask the partner only for the meaning of those words which one considers most important instead of asking for the meaning of each unknown word). Furthermore, this strategy needs to be accompanied by others, such as comprehension monitoring in order to increase the likelihood of successfully resolving the comprehension problem at hand. • The students in this study seem to require additional training and practice with the use of dictionaries, especially in combination with monitoring and inferencing strategies in order to find the most appropriate meaning in the given context. This is a very important pre-condition which is often ignored. A lot of students do not possess the basic knowledge about how to use a dictionary effectively. It might be argued that in the era of electronic dictionaries, the skills needed to locate a specific lexical item in a dictionary have become less critical; on the other hand, students still need to be able to determine the most appropriate meaning of an unknown word, considering the specific context. • The use of inferencing strategies appears to require training at two levels: (1) Students should be enabled to realise under which circumstances they can or cannot rely on this strategy (i.e. determine when it is adequate) and (2) they need to combine inferencing with monitoring strategies in order to increase the probability of successfully resolving the comprehension problem at hand. • Since metacognitive strategies did not figure prominently in the subsample examined here even though they may be combined with cognitive and socio-affective strategies very effectively, amendments to the task type and/or specific learner training appear to be worth considering.
The findings from Main Study 2 indicate that most teacher support actions are teachercentered and instructional as opposed to providing scaffolding. This may be a result of teachers' former training and/or subjective theories that they have to provide their students with definitive answers. The majority of teachers support their students by carrying out actions for them instead of eliciting strategies for autonomous problem-solving. This might be due to the fact that they believe they need to be, so to speak, omniscient and always provide the right answer. Moreover, the majority of teachers do not verify the success of their assistance -but this is crucial for the quality of teachers' assistance and students' success in learning. Only a few teacher support actions could be identified that foster self-regulated learning by means of supporting students' use of cognitive, metacognitive, and socio-affective learning strategies and by means of raising students' awareness of their learning processes. Thus, training alone would not necessarily trigger change in teachers' support actions, but training together with metacognitive talk about the teachers' perceived and adopted roles as good language teachers might have an impact. Thus, an approach that emphasises teachers' and students' shared responsibilities for the learning process (Finkbeiner, 2001(Finkbeiner, , 2004 might be the answer, together with specific training for teachers targeting the elicitation of strategies.
When relating the findings from Main Study 2 to those of Main Study 1, it is conspicuous that the teachers' support actions seem to miss their students' 'weak spots'. Most obviously, the use of metacognitive strategies on the students' part is just as rare as the teachers' elicitation or facilitation of metacognitive strategies. There are several possible explanations for the lack of metacognitively oriented teacher support actions: • Teachers might not be aware of the relevance of facilitating the use of metacognitive strategies in general. • Teachers might not be aware of their students' lack of (adequate) metacognitive strategy use, and they might not possess the skills necessary to diagnose their students' potential deficits regarding the (adequate) use of metacognitive strategies. • Teachers might not be equipped with the skills needed to actually scaffold their students' use of metacognitive strategies (e.g. by modelling them in a context-based fashion which is yet transferable to new contexts or situations). • Teachers might prioritise their students' efficient completion of the task at hand over longer-term goals, such as developing their students' strategic competence. • Also, the given task type might not sufficiently emphasise or require the use of metacognitive strategies. This certainly applies to planning strategies, which the task type considered here does not target; to a lesser extent, this may also apply to evaluation strategies. In contrast though, Click & Clunk tasks lend themselves very well to using comprehension monitoring strategies, especially in combination with cognitive and/or socio-affective strategies, which has already been noted.
The low presence of metacognitive reflection complies with the findings by Finkbeiner (2005), who refers to this phenomenon as metacognitive illiteracy.
This list of possible explanations is not intended to be exhaustive. It becomes clear, however, that the issues mentioned here pertaining to the students' use and teachers' elicitation and scaffolding of metacognitive strategies are likewise applicable to the students' other 'weak spots' identified in Main Study 1 (e.g. adequate use of inferencing, questioning for clarification, resourcing, content-related world elaboration).
As a result, there is a clear need for teacher training programmes that raise teachers' awareness of how they can support and interact with their students so as to provide strategyoriented scaffolding. This includes the development of process-oriented diagnostic skills to identify students' 'weak spots' but also strengths with regard to their strategic competence. Hence, in contrast to solving problems for the students, teachers should involve them in the problem-solving process, strongly encourage them to rely on their abilities and knowledge, and guide them through the (reading) activity. To that end, teachers may tie in with students' prior world and language knowledge, ask them for possible ways to solve the problem at hand, or demonstrate and explain adequate problem-solving strategies. By doing this, the teacher takes over the role of a learning facilitator and role model who provides guidance and support and gives the students enough space for developing self-regulatory competencies (Finkbeiner, 2004).

Conclusion
The ADEQUA study examined students' 'natural' use of strategies and teachers' 'natural' support actions -that is, without any particular pre-study training -during cooperative EFL reading tasks. The selected findings presented in this paper represent good news and bad news. Let us start with the good news: We have seen quite a wide range of strategies in Main Study 1 as well as a variety of teacher support actions in Main Study 2. The bad news is that we have hardly observed any reflection either on the use or on the adequacy or success of these strategies and teachers' support actions, respectively. This corroborates the findings of other studies (Finkbeiner, 2005;Nold et al., 1997). Metacognition does not seem to play a major part in the participating students' and teachers' respective 'repertoires'.
So, the results are both promising and enlightening: There is a 'natural' and largely adequate use of strategies, yet in the end -depending on the individual strategy type -there is room for improving the success rate of strategy use. In particular, the use of strategies requires constant monitoring and evaluation, which would allow students to further optimise their learning process and its outcomes.
Based on the findings of our study, we have begun to design a teacher SAM that can serve as the starting point for the development of a training model as well as a base for a new research study. The findings show that it must not be understood as a general training model, but that it has to be defined and used in a domain-specific way, considering the diversity of text, task, learner, and teacher factors. So, the ideal SAM needs domain-specific adaptations.
Our support action model emphasises the importance of: • eliciting students' learning strategies as well as their reflection on applied strategies; • eliciting students' peer scaffolding to maximise the co-construction of L2 knowledge, skills, and strategies (given the limited time per student for interaction with the teacher); • using proactive feedback which is transferable to and applicable in future learning situations; • scaffolding students' situationally adequate strategy use; • considering learner adequate strategy use; and • supporting task and text motivation as well as task and text interest. 2 In a nutshell: Teachers who teach in the framework of self-regulated learning scenarios, at least as observed in our study, seem to really make an effort to make use of the new 'engine', which is strategies, yet they are not aware of the fact that their students need to test-drive this engine and, furthermore, test the suitable fuel and oil. In addition, teachers are not aware of the fact that in contrast to the car industry, they themselves as well as their students cannot rely on one single prototype but that they need to question which one is the best one in each specific situation. Ultimately, teachers need to promote their students' self-activity with respect to problem-solving and strategy use so that students 'will be in the drivers's seat and navigate smoothly through the literacy world' (Finkbeiner, 2006b, p. 49

Worksheet #3
Finally, the students work on a task which asks them to apply their understanding of the text to a specific scenario going beyond the text itself. Here, the students are supposed to imagine that they are Yolanda's classmates and write down what they would tell their friends about what happened at school that day. Note: So as not to confound L2 reading comprehension with L2 writing skills, the students were asked to respond to this task in German (being the official instructional language at German schools).