• No results found

4. METHODOLOGY

4.6 The analysis

4.6.2 Categorization

When beginning the more formal, structured analysis, there are several approaches to analysis of interview texts. I have been inspired by Giorgi (1985), Smith (1995) and Kvale (1996). Giorgi (1985) offers four steps in a phenomenological method of analysis, so-called meaning condensation:

First, the researcher should read the entire transcript to attain a general sense of the whole. The researcher then reads the transcript from the beginning in order to distinguish ‘meaning units’, focusing on the phenomenon which is being studied. Going through the meaning units, the researcher should then look for the psychological perspective they contain.

Finally, the researcher should combine the meaning units into a consistent account concerning the subject’s experience (Giorgi, 1985).

Here, the interview is reduced to smaller meaning-units. As Kvale (1996) comments, looking for natural meaning units and explicating their main themes may help in analysing extensive and complex interview texts.

Meaning-categorization is another similar way of analysing interviews (Kvale, 1996; Smith, 1995). Smith (1995) describes how to code an interview-transcript: He proposes that you make a code for each theme, and every time a theme is found in the text, the code should be used to mark the location in the margin. Further, the material should be rearranged according to the codes by putting together all the occurrences of each theme. The categorization helps to structure the interviews and gives an overview of the different themes (Kvale, 1996). Kvale also uses this approach to be able to quantify the findings. This is not a goal for the present study. Still, I have chosen to use meaning categorization as an approach to analyse the data material, with inspiration from Giorgi’s meaning-condensation.

4.6.2.1 Step one: open coding

Boeije (2010) refers to the first round of breaking down the data into meaningful units – which I refer to as categorization – as open coding. In this phase, not much emphasis is put upon the relevance of the research material, since it is still too early to know what will be of value and what will be irrelevant. I followed Smith’s course of action, gathering the occurrences of a theme in one category. Some people choose to base the

98

categorization on themes that emerge from the interview-transcripts, but according to Kvale (1996), categories can also be formulated in advance of the analysis or be taken from theory. As a starting point, I used my main interview questions as codes (that is, names for the categories), namely mood regulation, management of emotions, regulating energy-levels and creating/maintaining boundaries. The categories, along with the questions, were based on previous research findings that suggest that these are some of the reasons why people listen to music (Bull, 2000, 2007; DeNora, 2000;

Juslin & Laukka, 2004; North, et al., 2004; Skånland, 2007; Sloboda, et al., 2009; Sloboda, et al., 2001; van Goethem, 2010). In order to avoid overlooking important and unanticipated aspects of the data, I did not restrict the analysis to the predetermined categories, but remained open to emerging categories. I categorized the interviews during the summer of 2009, and ended up with seven main categories and 26 sub-categories (see table 1). The seven main categories were Use of MP3 players; The music;

Boundaries; Physiological; Affective; Cognitive; and Identity. Thus, Boundaries was the only category that was remained, with new subcategories. ‘Mood regulation’ and ‘management of emotions’ were put under the main category Affective and ‘regulating energy levels’ was put under the main category Physiological, which also included ‘sleep’ and

‘stress’. The emerging category Cognitive included ‘thoughts’,

‘concentration’, ‘focus’, and ‘motivation’, while the main category Identity included the subcategories ‘identity’, ‘self-esteem’, ‘memories’, and

‘personal music’. Everything that had to do with the music and the use of the MP3 players were put under the last two main categories.

The categorization was carried out by reading each transcription while making notes in the margin. Using Word, the topics were thereafter collected within each category. Smith (1995) notes that one should be aware that while refining the categories, one removes oneself further away from the original interview-text and –context. It is therefore wise to go back to the original transcripts once in a while to remind oneself of the individual interviews and their context. To avoid overlooking important aspects in the interviews, and to remind myself of the original context of the conversation, I read each transcription at least twice.

99

The meaning of the subjects’ statements was decisive for the categorization.

For example, if a subject stated that, “I listen to music to shut out the surroundings” this was put in the category of boundaries. Equally, statements like “when I listen to music, people see that I’m not available”

and “when I listen to music I don’t have to listen to other people talking”

were put in the same category. However, all the categories are linked together. For example, being able to shut out the surroundings makes it easier to focus on one’s own state of being and thus regulate mood or energy.

Hatch (2002) differentiates between what he calls typological and inductive analysis. In the inductive analysis, the categories emerge from the data, while the typological analysis is based on predetermined categories.25 This implies that my analysis has been based on what Hatch labels a typological approach, which has its strengths and weaknesses:

The primary strength of typological analysis is its efficiency. Starting with predetermined typologies takes much less time than

“discovering” categories inductively. The potential weakness is that applying predetermined categories will blind the researcher to other important dimensions in the data (Hatch, 2002, p. 161).

Hatch stresses that all qualitative research is based on “[…] inductive rather than deductive information processing” (2002, p. 161). However, he claims that what he labels inductive analysis works less well for research that is based entirely on interviews or studies where the focus of the research is narrow. This implies that a strict inductive method would work less well for the present study, which is based on interviews. However, as new categories emerged from the material as additions to the categories that were formulated in advance of the analysis, the current research must be said to be based on a typological-inductive analysis.

25 This is perhaps more commonly referred to as deductive analysis (e.g. Boeije, 2010).

100

Table 1: Categorization of the interviews, step 1.

1. Use of MP3 players o Use

o Loudspeakers vs headsets o Home vs outside

o Much music o Radio

o The importance of the MP3 player 2. The music

o Music choice o Music experience o Music’s effects 3. Boundaries

o Boundaries o Environment o Private time 4. Physiological

o Energy o Sleep o Stress 5. Affective

o Emotions o Mood o Alone/lonely 6. Cognitive

o Thoughts o Concentration o Focus

o Motivation 7. Identity

o Identity o Self-esteem o Memories o Personal music

101 4.6.2.2 Step two: axial coding

The analysis continued as I wrote about the findings. Here, theory was merged with the data, which brought out new understandings of the phenomenon. This led to some adjustments of the categories, and I chose to divide the data in three main categories: The use of the MP3 player; The MP3 player as a technology of self-regulation; and The MP3 player as a technology of coping. Hence, the categories I started out with were modified as new theory was incorporated in the analysis. Boejie (2010) refers to this modification of categories as axial coding: “The relationships between salient categories (axes!) and subcategories can be generated, modified, refined, elaborated or even rejected throughout axial coding” (Boeije, 2010, p. 108).

I came to understand that much of what was talked about in the interviews was related to self-regulation, and so the different categories cognitive, affective, and physiological were merged into the overarching category self-regulation. When it came to the creation of boundaries and private space, it became clear that this was part of a coping strategy, and I chose to place it within the category coping. The subcategory ‘stress’ was also moved into this category. The two main categories use of MP3 players and the music were merged into one category, as these two aspects were linked together.

The question of identity has for the most part been left out of the final text, as the issue is somewhat on the side of the research focus.

As I kept writing, the categories were further refined (see table 2). For example, I found it interesting that several of the informants compared music listening to meditation, and chose to include this as a subcategory to cognitive. It also became clear that an important function of the MP3 player was its ability to facilitate unwinding, particularly after work. ‘Unwind’ was thus also added as a subcategory to cognitive. The aspect of control appeared to be an important side of coping, and was therefore added to this category. The hierarchy of categories was also refined, leading to four levels of categories.

102

Table 2: Modified and refined categorization of the interviews, step 2.

1 Use of MP3 players o Use

! Routines o Music experience

! Loudspeakers vs headsets

! Home vs outside o Much music

! Music choice

! Music’s effects

! Radio

o The importance of the MP3 player 2 Self-regulation

o Cognitive

! Thoughts

! Unwind

! Meditation

! Concentration/focus o Affective

! Mood

! Emotions o Physiological

! Energy

! Working out

! Sleep 3 Coping

o Boundaries

! Environment

! Private time

! Personal music o Control

o Stress and coping

103

Notably, the twelve informants are quite unanimous in their experiences. In other words, there is little dissonance to be found in the interviews. Of course, the informants are twelve different individuals, and their use of - and experience with their MP3 players are not identical from one person to another. Still, in general, I find that their narratives are roughly in agreement with each other.