Claudia Finkbeiner (Kassel)
1. Research design
The first stage of the main study involved quantitative analyses comprising a total of four investigations carried out during the academic year (cf. Finkbeiner 1997a, 1997b). One of these studies was aimed at investigating pupils naïve or latent knowledge which was then raised to the declarative level in this study or, in other words, the subjects were asked about their approach to their own learning by using LEFT. Prototypical similarities could be established and learning patterns formed. In addition, subjects were selected and were observed whilst performing a concrete task. Finally both the quantitative and the qualitative aspects of the process were collated and compared. Over three hundred pupils from a total of 14 different year nine classes in middle ability (Realschulen) and grammar schools in Baden-Württenberg took part in the first stage of study. Based on the results of this stage, subjects were selected for the second stage of the main study (from the total sample). The following conditions had to be fulfilled for the selection:
As a result of these preconditions, the total sample was reduced to n = 287 i.e. a total of 287 pupils from the original 350+ had participated in three of the four investigations. A selection of extreme cases in the area of reader interests and strategies was taken from this sample (n = 287) in a theoretically systematic way.
This selection was based on the following methodological steps:
The selection proceeded via theoretical sampling methods based on statistical sampling methods dealing with typologies.5
2. The qualitative investigation of the second stage of the main study
The second stage of this investigation (cf. Lamnek 1989; Mayring 1993) was concerned with qualitative, problem centred (i.e. focused on specific problem areas) interviews. These were characterised not only by being oriented towards the processes and relevant contents but also by their openness. The process orientation resulted from creating reflexive relationships amongst the individual utterances and from carrying out a flexible analysis of the individual utterances within the context of the whole group. The interviews were all one-to-one and orally based; their style ranged from neutral to informal and the questions from semi-structured to open. The communication form between subject and investigator was face-to-face. The interview was based on partially structured "key" questions. These key questions provided the basic framework whilst leaving enough leeway for openness. The questions comprised of the following points (cf. Mayring 1993).
Each interviewee had to read one narrative and one expository text and immediately afterwards, they were questioned retrospectively on their thought processes whilst reading. The two most important questions were; "What was going on inside your head when you were reading this text?" and "What was going on inside your head when you tried to understand this text?"
3. Selection of results
In this section, some of the investigation results on the interaction between strategies and interests will be presented as typical examples.
3.1 Strategies and classes of interests: selected results from the first stage of the main study
Based on the data gleaned from the Learning Strategies English as a Foreign Language - Textcomprehension (LEFT) questionnaire, factorial and reliability analyses (Crombach's alpha) were conducted for the total sample (n = 333). The analysis divided into six factors resulted in a meaningful distribution of the strategies into their respective category areas; these areas were taken from the strategy classes as in O'Malley and Chamot (1990): the questionnaire was firstly designed in accordance with their theoretical principles and then extended to apply to the specific context. The "elaboration category" (Factor I, alpha = .8512) proved to have the highest internal consistency within the reliability study. After this category came "metacognition" (Factor II, alpha = .6694) followed by "problem" (Factor IV, alpha = .8097) as "non-awareness of strategies" (Factor V, alpha = .6660.). All these categories broke down into factors which were both unambiguous and consistent. However the categories for "cognition" (Factor III, alpha = .5584) and for "metacognition/social strategies"(Factor VI, alpha = .6445) did not prove to be unambiguous: they resulted in a combination of cognitive and metacognitive strategies (Factor III, alpha = .5584) and metacognitive and social strategies (Factor VI, alpha = .6445).
The data from the Questionnaire on Interests in Texts in English as a Foreign Language (ITEF) were subjected to basically the same statistical analysis as had been the case with the data from the learning-strategy questionnaire. The factorial analysis using nine factors resulted in a stastically significant distribution of prototypical reader interest groupings. The factors comprised the following: interest in factual and practical texts (Factor I, alpha = .9435) interest in literary texts (Factor II, alpha = .8977) interest in independent activities (Factor III, alpha = .8742) interest in regional studies (Factor IV, alpha = .9090) non-interest/indifference (Factor V, alpha = .8870), interest in foreign culture (Factor VI, alpha = .8620) intrinsic interest (Factor VII, alpha = .8854) interest in competence/high achievement (Factor VIII, alpha = .9064) interest in social forms/team work (Factor IX, alpha = .7928). The reliability test produced alpha values for all factors with values between 0.7928 and 0.9435, thus confirming the hypothetically based constructs because of the internal consistencies ranging from high to very high i.e. the individual items delineated each relevant interest area in an unambiguous fashion.
The total sample population was divided into three hierarchical categories (top, middle and bottom third) in accordance with the total values aimed at within these factors following the procedures as shown in the preceding sections (cf. 8.1). The subjects were selected for interview only from the top and bottom groups.
3.2 The role of elaboration in foreign language reading.
As the reliability study showed the highest internal consistency for the "elaboration" category (factor I) and in addition revealed valuable indicators concerning role of elaboration in foreign-language reading as a result of the findings from the qualitative data, the results were extracted and described in the schema below. They are of particular relevance for the context outlined in this study because they could give a possible answer to the question posed in the title "Promotion of implicit and explicit learning strategies in English instruction: a necessary aim ?"
The "elaboration" category in the questionnaire on learning strategies (English als Fremdsprache: Textverstehen (L-E-F-T)) was covered by the following items (cf. O'Malley & Chamot 1990). When I read an English text,
Elaborations in this context thus imply the connections a reader can make between the text and his/her previous knowledge or experience as well as those which can be made within the text itself.
3.3 Intercorrelation analyses for both the separate questionnaires and between questionnaires
In order to establish possible connections and causal relationships between the elaboration scale and the other learning strategy scales, intercorrelation analyses were carried out for the questionnaires. With the case of the learning strategy scales, the intercorrelations between the main strategy classes ELABORATION and METACOGNITION were (r=.3067**).6 The category NON-STRATEGY AWARENESS displayed negative relationships to METACOGNITION (r = -.3757**) and COGNITION (-.3788**), but were slight with regard to ELABORATION being (r = -.1608*). The factor NON-STRATEGY AWARENESS included all those items which indicated the fact that the subjects gave little thought to conscious strategies, such as ordering in order of priority or making notes.
This last point is, in my opinion, very important because it implies that elaboration strategies are more concerned with strategies that have become so highly developed and automated that they have already developed into routines (in the case of the subjects who claim to be using elaborations). These have become so firmly entrenched in their procedural knowledge that they are mostly subconscious routines. There is possibly an implication to be seen in this situation that there really are strategies which differ in their level of consciousness and exist along a continuum starting more from implicit knowledge becoming then explicit.
Basically, the correlative interdependence between ELABORATION and METACOGNITION would seem to indicate that readers who have a strategic approach to reading do not choose a single category of strategies but prefer instead various relevant strategy combinations. It was also established in the interviews that some strategies were not merely added onto each other but were simultaneously implemented. The input and the combination of strategies in these cases varied according to the different nature of the level and type of text, of the themes and contents, of the context and the degree of emotional involvement or, in other words, the level of commitment and interest for each individual subject.
An intercorrelation analysis between the categories from the learning strategies' and the interests' questionnaire was carried out to establish this latter connection i.e. the interaction between strategies and interests in foreign language reading: The highest correlative connections were basically established between the strategy scale ELABORATION and the following interest scales: ELABORATION and INTEREST in LITERATURE (r=.5323**), ELABORATION and INTEREST in INDEPENDENT ACTIVITY (r=.4085**), ELABORATION and INTEREST in FOREIGN CULTURE (r=.3854**) whereas the interrelationship between ELABORATION and SUBJECT INTEREST proved, however, to be relatively weak in comparison (r=.2149**). The relationship between the interest categories and the strategy scale ELABORATION proved to be very noticeable so that the attribution theory seemed to be corroborated (c.f. Meyer & Försterling 1993). Pupils who regarded themselves as having a real interest in English were led, in a more or less conscious fashion, by this implicit attitude with regard to themselves and in their actual behaviour that they were convinced by all this that they really should carry out elaborations when reading foreign language texts. These results also show that particularly for advanced foreign language pupils, "independent activity" plays a very important role with regard to a deepening of the leaning process. This independent activity can be fostered and carried outwith the widest possible range of levels of abstraction.
3.4 Results from the interviews
The interviews produced some very detailed clues as to how elaborations are made. They showed that these connections are created at very different levels of abstractions. The point is confirmed by Zimmerman's results which were obtained from his investigation of the quantity and quality of elaborations when reading school grammar texts (cf. Zimmerman 1996).
This is surprising in that elaborations, on the one hand, seem to relate to the topic area and context or, in other words, seem to vary depending on the type of text and the contents whereas certain general elaboration characteristics which go beyond the topic field and context can, on the other hand, also be identified. These are particularly concerned with the various levels of abstraction with regard to the elaborations themselves.
In addition, the interviews showed that elaborations can be represented in very different ways. Examples from the two main classes of connections are cited below. The concrete statements based on this division were all investigated during a retrospection stage following the reading process:
Basically, it emerged that the nature, the quantity and the specific detail of the subjects' previous knowledge played a very big role in the understanding of foreign language texts and in the formation of elaborations. On the whole, the literary texts were more inclined to arouse elaborations coloured by personal factors than the factual text. Most of the connections had a basically visual orientation. It was very often the case that pupils connected the texts with images they had seen either on television or at the cinema.
On the whole, more intertextual connections than intratextual ones could be established. Intratextual connections were made less frequently which would have been obvious, particularly in narrative texts, between the name of the author and the text. In this context, the school text book most commonly used in years 9 and 10 had had an influence in its function as hidden curriculum. In the text books author's names are not given and there are often no explicit titles. Possibly as a result of this, a large proportion of the subjects simply had not noticed, but had simply ignored the author's name given in the narrative text. Hardly any conclusions were drawn about the contents of the text from the author's name.
It also emerged that generally most of the subjects did not understand the role of elaboration for understanding texts and some were not even aware of that forming elaborations. During the course of the interviews, they revealed that they had never been so explicitly questioned about their thought processes and mental connections made whilst reading and they had never taken themselves seriously in their "active" role. In addition, the following extremes as well as "limits" established themselves for the formation of elaborations:
One subject produced extremely personal and highly creative elaborations. However, whilst reading the elaborations became so personal that they finally resulted in completely solving from the text.
The subsequent interview revealed that this formation of elaborations motivated in a highly personal way constantly took place whenever the subject was reading so that it had already become a real problem for learning and concentration. Up till that time, however, this problem had neither been recognised nor diagnosed.
One female subject almost failed to conduct personal elaborations. This was all the more surprising in view of the fact that, after the official interview was over, it emerged that the subject had come to Germany from the Philippines two years prior to the interview and had thus gone through the same experiences of being a foreigner as was the case with the main character in the text. She succeeded in making the connection only between her own experiences of estrangement and those described in the text after the tape recorder had been switched off. The previous situation had been so tense that the formation of elaborations had been severely hindered by the subjective emotional situation which she had felt to be stressful. She reported that she had always felt the modern language situation in Germany to be artificial and thus, both tense and stressful. She maintained that as a result of this, that she avoided using English in a school context although in the Philippines she had had a native-like command of the language.7
In several investigation sessions another phenomenon that could be observed during the reading process stages was the active lip movement when reading in silence. A few subjects had obviously not learnt to read without the motor support provided by lip movements. In two cases this motor support was further supplemented by finger movements i.e. the finger moved forward with every word that was read. The "finger" readers in particular showed a strong tendency to translate word for word and could hardly go beyond the surface of the text to carry out elaborations.
The connection between emotional involvement in the text and in-depth analysis of the text proved to be absolutely clear and unambiguous. At the end of the interview, the subjects were asked to compare the two texts they had read. They had been given no criteria of comparison. In almost every case the subjects spoke in terms of the interest level of the text itself and their own interest in the text as their own emotional participation in the text and thus their willingness to become cognitively involved in the text. The following text extract would seem to support this contention.
Example:
Please compare both texts and tell me what you think the difference is between them.
Well, all that's about space travel, its sort of without feelings and it's more a description of something, how somebody experiences something but this one here is perhaps more a tale or a story about somebody.
Was there also a difference in the way you read the texts?
No, not really. Except that the well (...) that the snow text is perhaps more interesting to me personally! And I'd 've continued reading it.
Did that mean you read in a different way?
Mm, it was more interesting to understand.
What do you mean by more interesting to understand?
Well, er, with the texts I don't like so much, I read them more superficially and I don't try as hard to understand.
But this time, you did take trouble to understand.
Yes.
What does that mean?
It means I connect the new unknown words with the text and try to see what they mean.
8.3.5 Synopsis of a few selected results from the first two stages of the main study:
quantitative and qualitative results
Finally some of the results from stage of the main study (quantitative) will compared with those of the second stage of the main study (qualitative). Basically, there was a high correlation between the number of elaborations encoded in the interview and the type of school, which means in this case that the grammar school population had carried out a much higher number of elaborations than the "Realschule" population had done.
Although the correlations between the ELABORATION INTERVIEW factor and the strategy classes in the questionnaire turned out to be similar to the relationship of the ELABORATION QUESTIONNAIRE factor and the rest of the strategy classes in the questionnaire, they were, however, not as clearly defined: ELABORATION INTERVIEW and COGNITION QUESTIONNAIRE (r = .2150), ELABORATION INTERVIEW and METACOGNITION QUESTIONNAIRE (r = .1711), ELABORATION INTERVIEW and ELABORATION QUESTIONNAIRE (r = .1392). The last figure cited showing the low correlation between the elaboration category of the questionnaire in comparison to the one in the interview shows that there is obviously a huge gap between knowledge about declarative and processing strategies. It was, however, not always the case for all the pupils that the extremely high or extremely low placing in the ranking orders within the total values of the strategies always accorded with each other. In some cases, there was a discrepancy.
In addition, the original hypothesis concerning explicitness and implicitness together with the context-dependency factor for the input of knowledge strategies would seem to be confirmed by the data. It also emerged that right up to the interview, many pupils had little knowledge or awareness of their own processes when reading English texts.
This does not, however, mean that these "subconscious" readers did not undertake to carry out strategies when reading, but more that they did not have a positive picture of their worth and use. There had also been a group of students who, in their questionnaires had claimed not to use strategies but who, in interview, at times carried out very abstract elaborations.
For example, there was a case of one subject who had categorised himself as relatively uninterested and not particularly gifted linguistically: this pupil had felt so deeply involved in the factual text about weightlessness in space that he had carried out macro-propositional elaborations as in van Dijk & Kintsch's (1983) terminology. This had been the case because he was an "expert" in this field. and his "expert" knowledge had allowed for a completely different access into the text, thus enabling processing to take place at a deep level.
The final communicative validation showed that the discrepancy between the answers in the questionnaires and the actual observable behaviour during the interview stage for some of the pupils was based on the fact that they had answered the questionnaires while still working in a school context which obviously offers little time and space for elaborations:
Example:
You claimed in your questionnaire that you did not empathise with the characters or that you could not forget the time whilst you were reading so that this sort of activity had little effect on you.
Yes, that's right.
But now you have been able to empathise with the girl.
Yes, you see it's like this, when you're reading in the classroom and then you have to do something about some points on the text it's then discussed straight away, - well then you just simply don't have enough time.
The correlations between the category ELABORATION INTERVIEW and the interest categories of the interest questionnaire came out as very high and to a certain extent highly significant: ELABORATION INTERVIEW and INTRINSIC INTEREST QUESTIONNAIRE ( r = .6342**), ELABORATION INTERVIEW and INTEREST in LITERATURE QUESTIONNAIRE (r = .2712), ELABORATION INTERVIEW and INTEREST in INDEPENDENT ACTIVITY (r = .3514).
This correlation confirms the view that interest favours the formation of elaborations and deep processing.
5 Statistical sampling and theoretical sampling: the approach in theoretical sampling (theoretical sampling by Glaser and Strauss, 1967, 1979) is of an exploratory kind with its use of flexible theoretical concepts (Witzel, 1982 in Lamnek 1988: 177).It aims at a selected target group which emerges from the theoretical bias of the researcher as in a quantitative paradigm (cf. Lamnek 1988: 223). Thus, the researcher's theory formation is regarded as a process which can be seen to take place in the gathering, codification, analysis of the data and this leads to a new, possibly even revised selection. The formation of groupings and extreme types which arise from the sampling process should then, by means of qualitative investigatory and interpretative processes, allow for the reconstruction of both interpretation models and behavioural models which are typical for those social groupings to which the subjects belong (cf. Lamnek 1988: 175). As in object description, it can be assumed that even individuals can be identified by using specific continua referring to typicality (cf. Eckes 1991: 143). However, differences have to be expected in the object field in contrast to the formation of prototypes. In addition, the following viewpoints can be suitably differentiated (cf. Eckes 1991: 144 f.): Stability versus variability, internality versus externality, intersubjectivity, causal attribution, consequences for the self-image, behavioural contingencies, perceiver as interaction partner.
6 The double asterisk (= **) means that p< 0.001 i.e. the result is highly significant.
7 In this case, the interview was conducted in English because the subject felt more at home in this language and so,was better able to express her statements concerning her own thought processes.