In the present study a form of qualitative content analysis was employed. In particular, the conventions of the method known as the Grounded Theory were used as the basis for data analysis. The Grounded Theory is an analysis method which can be used in both qualitative and quantitative research. When analyzing data according to the Grounded Theory, the focus is not on linguistic features, as it is in studies based on discourse analysis, but instead on the content. Since the purpose of the present study concentrates on what the participants say instead of focusing on how they say it, the Grounded Theory was seen as a suitable basis for conducting the analysis. Dey (2004:80) describes one of the characteristics of the Grounded Theory. It is said that the
frame for analysis in the Grounded Theory is data-oriented. Dey clarifies this by saying that no preselected criteria or categories are used in the analysis, nor are any hypotheses being tested. There are no preconceptions; instead the categories arise or are discovered from the data itself. The name of the analysis method is quite apt, since the purpose is to produce a theory which is grounded in data (Valanides 2010:62). Furthermore, this kind of approach is said to produce data which is closer to the reality than would be the case with filtering the findings through previous experience and speculation (Strauss and Corbin 1998:12). In line with this, in the present study no predetermined categories were used but the data was analyzed keeping the research questions in mind.
Analyzing the data by means of the Grounded Theory involves coding the data in order to form categories through constant comparison (Dey 2004: 80). The procedure consists of three stages, as Valanides (2010:60-62) demonstrates. In the first coding stage the data is dissected into pieces for a close examination in order to find the similarities and differences of incidents. The data is then assembled and core categories formed. In the third and final stage a theory explaining the phenomenon in question is produced based on the concepts identified. The present study employed this analysis framework. In other words, the interview data was coded and categorized. Before elaborating on this issue, it should be noted that the analysis is divided into two sections according to the research questions and the type of data. Chapter 5 focuses on the data derived from the semi-structured interviews (i.e. on the student teacher’s perceptions), while chapter 6 concentrates on the data produced in the stimulated recall interviews (on the motives given for language choices).
The first step of coding involved identifying whether the participants were referring to TL or L1 use in their comments. Comments made on L1 use were separated from those referring to the target language. This was done because the purpose of the present study is to reveal the roles of Finnish and English. This procedure concerned the interview data as a whole. In addition to sorting the data by language, general comments on teacher training and on planning one’s language use were analyzed separately from the language-specific comments (see section 5.1 of the analysis). As for the data derived from the semi-structured interviews, it became clear that most of the interview questions had elicited information about the perceived uses of the two languages. In other words, the participants had depicted the possible uses which the two languages may have in the classroom. Categories of uses were formed on the basis of the comments. In the same manner the data obtained in the stimulated recall sessions was first coded according to the language and then according to the types of motives given for that certain language.
Three main categories of language use were identified in both types of interview data. These included teacher-related-, learner-related and discourse-related uses. The first category relates to personal principles and to the participant’s relationship with the given language. Edmonson’s view on psycholinguistic code-switching (see discussion on code-switching in section 3.4) was broadened and applied to this category. Language choices which were seen to be prompted by the learner’s actions or which were seen to occur for the sake of the learner (i.e. for pedagogical purposes) fell into the second category. As for the third category, reported connections between a certain language and a particular type of a speech situation, textual coherence, marking the boundaries of speech situations and signaling the teacher’s stance towards what is said were categorized as discourse-related. This category shares some characteristics with Auer’s definition of discourse-related code-switching (see section 3.4 on code-switching). In addition to these three categories, one category relates to the influence of teacher training on the participants’ language use. Furthermore, comments about the perceived disadvantages of employing each language were identified from the data produced in the semi-structured interviews.
Before moving on to the analysis, a few things should be pointed out. First of all, in the analysis the focus is not on different student teacher profiles but on the profiles of the two languages. It follows that the participants are not identified in the analysis but all three are referred to simply with the word participant. When necessary for the interpretation of the data, connections between comments made by one and the same participant will be drawn. Secondly, some of the reported types of language use may overlap. That is to say, the category boundaries are flexible: some learner-related uses may exhibit discourse-related features, for instance. Furthermore, in relation to the recorded lessons and stimulated recall interviews it cannot be emphasized enough that the focus was on how the participants perceived their actions, not on classroom interaction. The researcher did not analyze the actions (i.e. language choices) that took place in the classroom per se. Instead the researcher was interested in how the participants themselves viewed the language choices they had made. In other words, the present study assumed a second-order perspective. Next the focus will shift to the analysis itself.
5 PERCEPTIONS OF LANGUAGE CHOICE
In this chapter the data derived from the semi-structured interviews is analyzed. During the interviews the participants were asked a series of questions regarding the teacher’s language choice in the foreign language classroom. These questions can be found as an appendix in the end of the present study. First general comments made on teacher training and planning one’s language choices are given attention. After this the analysis proceeds to the language-specific comments. Comments regarding the target language are considered first. This is followed by the analysis on the comments concerning the use of the L1.