In this study, I did a thematic analysis on the data collected through the interviews.
According to Braun and Clarke (2006), thematic analysis is a method for identifying, analysing and reporting patterns, or themes, within data. It is widely used for analysing qualitative data, with the goal of finding themes and patterns across the entire dataset.
The authors argue that the themes can be found in one of two primary ways: inductively or deductively. In the inductive way, the data is coded without trying to fit it into a pre-existing theoretical framework. Conversely, finding the themes deductively means that the analysis is driven by the researcher’s interest in a particular area. This way, the thematic analysis gives a less rich description of the data overall, but a more detailed description of some parts of the data.
I chose to conduct a thematic analysis on the data, since I was interested in discovering what types of experiences the research participants had had participating in design teamwork remotely. The nature of my study is inductive, and therefore I wanted to see what themes would arise from the data while doing the analysis. In order to do achieve
this, I searched for themes in an inductive way without trying to fit them into an existing framework, as described by Braun and Clarke (2006).
In my thematic analysis, I used the process presented by Braun and Clarke (2006). Their process for thematic analysis includes the following steps:
1. Familiarising with the data 2. Generating initial codes 3. Searching for themes 4. Reviewing themes
5. Defining and naming themes 6. Producing the report
I will present these six steps in further detail and describe what each step looked like in practice for me when conducting the analysis.
3.4.1 Familiarising with the data
Braun and Clarke (2006) strongly recommended for the researcher to become familiar with the data prior to the analysis. This makes it easier for the researcher to find appropriate codes. The authors also argued that verbal data (e.g., from interviews) needs to be transcribed before a thematic analysis can be conducted. Braun and Clarke (2006) considered the transcription to be a key part for the researcher to become familiar with the data.
As I had conducted all the interviews myself, I was already quite familiar with the data.
To save some time, I used a software called Otter.ai for transcribing the data. The software produced decent quality transcriptions, but I still needed to ensure that the transcribed text was completely correct. Therefore, I listened through the recording of each interview and corrected the automatic transcriptions. This way, I further familiarized myself with the data, while saving some time from the tedious transcription work.
3.4.2 Generating initial codes
According to Braun and Clarke (2006), this phase involves finding features of the data that appear interesting to the researcher. These features are called codes, and they are descriptions of segments of the data. Braun and Clarke (2006) distinguished between data-driven codes and theory-driven codes, where in the former the themes depend on the data, while in the latter the researcher has specific questions in mind that is then coded around. Regardless of the approach, the authors recommended to code for as many potential themes as possible, since it might be hard in this stage to know what will be interesting later.
I coded the data using a software called Atlas.ti. Using this software, I read through the transcription of each interview and labelled interesting segments with describing codes.
All in all, I created 27 codes for my dataset. As I did not have any specific themes in mind when I coded the data, I had a data-driven approach to my coding. If needed, I coded some segments with more than only one code.
3.4.3 Searching for themes
This phase begins when all data has been coded, and there is a long list of codes that have been identified from the data set (Braun and Clarke, 2006). In this phase, the researcher analyses the relationship between the codes and combines them into more general themes, that may also contain sub-themes. The authors recommended to not yet discard any information that do not fit into the themes, as the themes will still be reviewed and refined.
As I had generated 27 codes so far in my analysis, it was not an easy task to find a way to group the codes neatly into separate categories. Three of my themes were formed easily, while there were two themes that I struggled to define properly. After a while, I managed to create themes that at least initially looked promising. This phase of the thematic analysis was also done using Atlas.ti, as the software helped me to keep track of my codes and code groups.
3.4.4 Reviewing themes
Once initial themes have been identified, it is time to review and refine them. Some themes might be broken down into two or more themes, while others might be merged into each other. Braun and Clarke (2006) described two levels of reviewing and refining the themes. Level one consists of making sure that the data in each theme is consistent
with the content of the theme. Level two focuses on reviewing whether the themes accurately reflect the data set as a whole. As coding is an ongoing process during the thematic analysis, there might be a need to re-code some segments of data. The authors, however, warned about redoing and reviewing the codes and themes too much, as it can be done indefinitely. Instead, they recommended the analyst to stop the review once refinements do not add any substantial additional value.
While reviewing my themes, I realized that two of my themes were overlapping each other. After I changed the definition of my themes, I managed to create groups of codes that were internally consistent and not overlapping. This was the level one review of my themes, as described by Braun and Clarke (2006). As the author suggested, I also looked at my themes in relation to the dataset as a whole and concluded that they accurately reflected the dataset.
3.4.5 Defining and naming themes
In order to define and name the themes, Braun and Clarke (2006) argued that the analyst must identify what the essence of each theme is about. It is important that the analyst after this phase knows precisely what the themes are about, and what they are not about.
The authors recommended to not make the themes too diverse or complex, but rather define them in a way that can be described by using only a couple sentences. The naming should also be concise and clearly give the reader a sense of what the theme is about. It is also important to identify whether a theme contains sub-themes, which essentially are themes within a theme.
To name my themes, I had to define the essence of each theme. While doing this, I concluded that none of my themes contained sub-themes. I tried to name the themes in a way that easily described what each one is about. After my analysis, I had generated five themes in total:
1. Personal experiences of remote working
2. Changes in how the work is organized 3. Effects on team collaboration and creativity 4. Effects on the workplace atmosphere 5. Opinions about the future of work.
The codes for each theme can be found in Appendix 2.
3.4.6 Producing the report
The last step of the thematic analysis is to produce the report. According to Braun and Clarke (2006), the aim is to present the content of the data-set in a coherent, concise, and non-repetitive way that convinces the reader of the validity of the analysis that has been done. The authors stated that in order to achieve this, the report must provide evidence that supports the relevance of the themes. This can be done through presenting data extracts that demonstrate the prevalence of the themes.
The way I have structured my reporting, is to describe the findings from each theme one at a time. To preserve the anonymity of the respondents, I chose not to reveal their genders. Therefore, all respondents are referred to as “he”. To provide evidence that supports the relevance of my themes, I have included quotations from the interviews. I chose the quotations that best describe the content of the interviews and that also relate to the theme I am presenting.