Data analysis

The data of the present study was analyzed by content analysis. Content analysis, according to Tuomi and Sarajärvi (2009: 91), is a basic method of analysis that can be used in all traditions of qualitative research. Content analysis can be regarded either as a single method or a broad theoretical framework that can be used in various ways to do research. Qualitative content analysis, based on classification by Eskola (2001: 136), falls into three main categories: data-based (aineistolähtöinen), theory-based (teorialähtöinen) or theory-bound (teoriasidonnainen; or teoriaohjaava (theory-guided), an analogous term used by Tuomi and Sarajärvi (2009: 96); Tuomi and Sarajärvi 2009: 95–100, 107–120). These forms of analysis can be distinguished according to their connection to the logic of reasoning – inductive, deductive or abductive – used in the analysis: data-based analysis is commonly connected to inductive reasoning, theory-based analysis to deductive reasoning, and theory-bound analysis often to abductive reasoning, or a combination of inductive and deductive reasoning (Tuomi and Sarajärvi 2009: 95–100). In data-based analysis a theory is constructed from a given data, whereas in theory-based analysis, which is the classical model of analysis, a theory, model or idea proposed by an authority is the basis on which the analysis of the data is built; in theory-bound analysis the analysis has theoretical links, but the analysis does not, however, directly arise from or is not directly based on a theory (Eskola 2001: 136–137). Summing up, the differences between these three forms of analysis relate to the role of theory in directing the collection, analysis and reporting of the data; for example, both in data-based and theory-bound analysis data collection is free in relation to theoretical background, while in theory-based analysis how data collection is organized and how the phenomenon to be studied is

determined are dictated by what is already known about the phenomenon to be studied (Tuomi and Sarajärvi 2009: 98–99).

The aim of content analysis is to create a condensed, explicit verbal description of the phenomenon to be studied by organizing the fragmented data so as not to lose information, but rather to increase the informational value of the data (Tuomi and Sarajärvi 2009: 108). Qualitative data-based content analysis, according to Miles and Huberman (1994, cited in Tuomi and Sarajärvi 2009: 108), can be roughly described as a process with three main phases: 1) reducing the data, 2) clustering or grouping the data, 3) abstraction or creating theoretical concepts. According to Tuomi and Sarajärvi (2009: 109–113), the data analysis starts with reading the data and familiarizing oneself with its contents. Reducing the data means eliminating information irrelevant for the study from the data and either condensing the information or dividing it into parts.

Guided by the questions of the research task, essential expressions are sought, coded and listed. Original expressions are recorded and named with reduced expressions.

Before starting the analysis, the unit of analysis – a word, a clause or an entity of thought consisting of many clauses – needs to be determined, which is directed by the research task and the quality of the data. In clustering the original expressions coded from the data are gone through carefully and concepts describing similarities and/or differences are sought. The concepts denoting the same thing, the reduced expressions, are grouped or clustered into a class or (sub-)category which is named with a term describing the contents of the category. In the clustering phase, the data becomes condensed, as individual factors are included into more general concepts. After clustering, abstraction or conceptualization of the data follows. In the abstraction phase information essential for the study is selected and, based on it, theoretical concepts are formed; in other words, abstraction refers to proceeding from the expressions used in the original information to theoretical concepts and conclusions. Clustering is considered to be a part of the abstraction process. In this process higher / broader categories are formed from lower / narrower ones for as long as it is possible considering the contents of the data; in other words, sub-categories are grouped to form super-categories that form main categories that form a final connective category.

In this study, theory-bound / theory-guided analysis was used. As was stated above, theory-bound analysis has theoretical links. According to Tuomi and Sarajärvi (2009:

96–97, 117), theory-bound analysis of data proceeds first like data-based analysis, the units of analysis are taken from the data; the difference between the two forms of analysis arises in how empirical data is linked to theoretical concepts in abstraction. In theory-bound analysis former knowledge or a theory, i.e. ideas from the theoretical framework of the study, is introduced to guide or help in the analysis towards the end of analysis. In theory-bound analysis theoretical concepts are introduced as something that is already known concerning the phenomenon to be studied, whereas in data-based analysis theoretical concepts are formed on the basis of the data. As to at what phase a theory is to be introduced in theory-bound analysis, no rule exists, according to Tuomi and Sarajärvi (2009: 100); the decision depends on the data and the researcher. If theory-bound analysis is considered as regards inductive and deductive thinking, the earlier on a theory is introduced in the analysis to guide reasoning, the closer the analysis is to deductive reasoning; conversely, the closer the end of reasoning a theory is introduced, the closer the analysis is to inductive reasoning.

In practice, I started the data analysis by organizing the data of each participant, i.e. four (or three) documents (pre-course questionnaire, two logs during the course and post-course log), into one file, separating the data concerning learning of VS from the data concerning learning of English. All data, except for one pre-course questionnaire was in electronic form which greatly facilitated the data management. After this general grouping and condensing, it was easier to read through each set of data and form a general idea of their contents. Then, I focused on specific parts of the data guided by research questions (e.g. what VS content a participant had learned), and used a color or underlining to highlight all such relevant data. This was useful, because occasionally students’ answers included data relevant in more than one area. If there was plenty of such data, I put this information in a separate file and organized the data using different colours to highlight different areas or themes. These were mostly quite easily distinguishable, as the two main themes (VS learning and English learning) and the questions within them in the four documents facilitated the analysis process. Then, when needed, I condensed the information in the original expressions into reduced ones.

Next, I further combined and condensed those related to a theme into one reduced expression. After this, I grouped similar reduced expressions into a subcategory.

Combining the subcategories, main categories were formed, and finally these fall under the final category (for examples of the analysis, see Appendix 10).

5 FINDINGS

In this chapter, I will report the findings based on the learning logs kept by the students.

The Logs 1–3 and instructions for them can be found in Appendices 3–5. The findings are reported as individual cases and illustrated with original extracts from the student logs to voice each student’s experience as authentically as possible. The students could choose to use either Finnish or English in the logs. The students’ written language, neither their Finnish nor English, has been corrected or edited. For the extracts in Finnish translations in English can be found in Appendix 9. The extracts are numbered and whether a quote is from the pre-course questionnaire (PCQ), the first learning log (Log 1), the second learning log (Log 2), or the post-course questionnaire or the third learning log (Log 3) is indicated. First, the student experiences of content learning are described and discussed in 5.1, after which the student experiences of content learning are looked at in 5.2.

5.1 Student experiences of content learning

Before presenting the cases, I will remind the reader of the group context of the students. The first case, Piia, was in the same group with Jukka and Igor. The second one, Albin, and the third one, Olli, were in the same group, with the third student in this group being Jaana. The fourth and the fifth case, Sauli and Toni, were in the same group; the third student in their group was Henri. I will refer to Jukka and Henri only cursorily in discussion and conclusion, as they did not provide enough data for a thorough coverage (for more details see 4.4.2 Participants).

Each case starts by presenting the student’s background information (e.g. personal details and experience of playing the piano and piano VS) and expectations of content learning during the VS course before the beginning of the course (see Appendix 1).

Then, to answer the first main research question (i.e. What kind of experiences do JAMK music students have of learning the VS content in the CLIL VS course?), first, what was learned and why (i.e. the first sub-question) is discussed; and second, the factors influencing content learning and how they were viewed by the students (i.e. the second sub-question) is discussed.