Data Analysis

asked to give their consent to be recorded. No notes were taken as this was considered as a possible distraction for us as interviewers, instead, we were able to listen actively and carefully and respond to what the interviewees told us. As all interviews were conducted in quiet rooms without background noise, a good recording quality was secured.

All interviews were conducted in Swedish since this was the mother tongue of all the interviewees. Interviews in mother tongue enable the interviewee to be more comfortable, speak more freely and for the researcher to capture more nuances of what is brought up (Andrews 1995; Tsang, 1998). To increase the comfort and relaxation even more, the interviewees got to decide time and place for the interview.

When the interviews had been conducted, the recordings were used to transcribe the interviews.

This was done in order to allow the obtained information to be coded and analysed. The transcription process started directly after the first interview had been performed and continued in parallel with further interviews. We agreed on a transcribing style of verbatim and word by word and signs for pauses and expressions like laughter. Some interview parts were transcribed by both of us, in order to investigate if there were any discrepancies and to align the style of transcribing. This strengthens the transcription reliability and thus its quality (Kvale, 1996). At the point of time when all interviews had been conducted and fully transcribed, a respondent validation was composed. This means that a summary of the interview was sent out to respective interviewee in order for him or her to control the accuracy. A respondent validation enables the interviewees to control and confirm the data interpretation and thereby decreases the risk that the researcher’s own subjective interpretations and views mislead the data, thus increasing the credibility and authenticity of the data (Bryman & Bell, 2011). Provided discrepancies between the perception of the interviewees and the transcribed interview material, corresponding changes were made.

actual analysis. The analysis process can thereafter be divided into three parts; sorting, reducing and arguing. Sorting included coding of data, which means that the collected material was sorted into different categories (Saunders et al., 2009). The categories were developed based on the literature review and the collected data by searching for repetitions, similarities, differences and theory related material (Ryan & Bernard, 2003). To sort the material, the categories sustainable fashion consumption, renting and swapping were used, which are the ones we based the interview questions upon. The theme sustainability was removed before sorting since we found the collected data regarding this to serve better as a complement to the other sections rather than as a separate theme providing findings relevant for answering the research question. Within each of the categories, we constructed subcategories. For instance, under the category of clothes consumption, subcategories such as ‘quality’, ‘self-image’ and

‘price premium’ were created.

The next activity in the analytical process was to unitise the data. This means to attach relevant

“units” of data to the appropriate category. Sentences, a number of sentences or complete paragraphs from the interview transcripts were added to the relevant category accordingly. This selective process, guided by the research purpose allowed reducing and rearranging of the data, resulting in a more comprehensive and manageable form (Saunders et al., 2009). At this stage, the categories were revised, and data were rearranged continuously as we searched for key themes, relationships and patterns in the data. Categories were further subdivided or integrated as some categories attracted a great number of units and data and needed to be subdivided in order to enable further analysis. Also, new insights were gained within existing categories and thus suggested new ones. The reducing part also involved selecting parts of the subcategories that were found most interesting, not necessarily most frequently mentioned, and excluding the less interesting components (Styhre, 2013). For example, in the category renting, the subcategories ‘uncertainty’, ‘scepticism’ and ‘trust’ formed one untied subcategory named

‘uncertainty’. The reducing process resulted in initial findings on how the collaborative consumption models swapping and renting are perceived as well as the underlying motivations and barriers.

In order to explicate the meaning of the interviewees’ statements, three contexts of interpretation were applied. In a self-understanding context, we attempted to formulate what

the interviewee him- or herself believed to be the meaning of his or her statement in an abbreviated form. The interpretation is a rephrased condensation of the meaning of the interviewees’ statements from their own point of view as these were understood by us. The sorting and categorization of data happened within the context of the interviewees’ self-understanding. In the next context, critical common-sense understanding, our interpretations went beyond rephrasing the interviewees’ self-understanding. The interpretations included a broader frame of understanding than that of the interviewees themselves. We critically viewed their statements and focused on the content of what was said by including general knowledge about it. By doing so, we could enrich and amplify our interpretations of the statements (Kvale, 1996). Lastly, in a third context, we applied a theoretical frame for interpreting the statements’

meaning. In this context, the interpretations went beyond both the interviewees’ self-understanding as well as a common-sense self-understanding and we applied the theories presented in the literature review in order to interpret the meaning of the statements. Thus, by doing so, the findings were argued by comparing the reduced empirical material to the existing literature.

3.4.1. Translation

The transcribed material was in a final step translated from Swedish to English, resulting in some methodological considerations. We performed the translations ourselves and as such the process of translating the transcriptions was not entirely neutral or unbiased since we choose and decided on the most suitable words and construction of sentences for expressing the interviewees’ original statements. Hence, the original statements were transformed from Swedish to English, and the interviewees´ initial wordings were altered. Consequently, the interviewees´ statements are being expressed and portrayed via our perceptions and understandings of their viewpoint. It is important to address and reflect upon these implications since they result in that some of the interviewees´ expressions were altered or in some cases even to some extent lost in translation.

Thus, finding and selecting subcategories in Swedish that represented the accurate meaning from the literature resulted in some challenges. In some cases, translating the words directly would have resulted in implications of the words’ meaning, since some of the words cannot be translated directly and some words do not have an accurate counterpart in the two languages.

In these situations, we considered the difference in the actual meaning of the words, their

definition from a dictionary and considered what meaning the interviewees’ expressed when discussing these keywords. Furthermore, it was challenging to in a sufficient manner portray some of the interviewees’ statements. This became apparent when the interviewee used phrases and words that are deeply embedded in the Swedish language and therefore lack a direct translation that reflects the meaning in English. In order to minimise these aforesaid challenges, we have during the process been aware of the possible misinterpretations and implications that may occur. Therefore, when translating the statements and words we discussed and questioned each other in order to find the most suitable words or phrases that represented the intended and original meaning instead of solely translating the words and statements directly.