At this point in your research journey, you will hopefully have completed your data collection. The focus of this blog post is on what to do with all of this rich data. It goes through the stages of data analysis, using an example of a research article to demonstrate the process.
Stage 1: Data Coding
It is probable that you have ‘too much’ data. This is a typical outcome of qualitative research, which if methods are used well, can produce large amounts of data, not all of which can be included or are relevant to your specific research question/topic. The first stage of your data analysis is to begin to organise all these data into a manageable form. Imagine your research as a story; in order to create an engaging tale, you will include the most significant elements in a structured way, and you may even have to omit some interesting digressions along the way to maintain a coherent narrative.
In terms of your research, you achieve this through first ‘coding’ your data into meaningful units. Coding is applied to textual data to identify and label snippets of text that relate to certain meanings or ideas. These ‘codes’ can be allocated using short, descriptive names or tags. As you code the data, try to keep in mind ideas that are central to your research question.
The codes that you apply may come through the literature you have previously analysed (deductive coding), using a ‘top-down’ approach.
However, you may also allow new ideas to emerge through the data (inductive coding), using a ‘bottom-up’ approach’.
These approaches feel different, but can be complementary rather than conflicting. An interesting reflection by Amy Blackstone gives examples of the uses of these approaches.
For example, in our data analysis work group, I coded the interview transcript from my previous research as follows:
Stage 2: Categorisation
Once you have finished coding the data, you need to put it back together again. Some have described this process using an omelette analogy; to make an omelette, you have to first break the eggs (your data) before you can combine them again in a new way, into a new dish. Others describe the data analysis process as akin to making a jigsaw; your data are the little pieces that you begin to group together to create the big picture (This group makes the sky, this group makes the trees, this group makes the buildings, etc.).
In our group session, we used the ‘Quality Street method’ to visual categorisation. Incidentally, I was first shown this method by a researcher called Mary Kellett, who is founder of the Children’s Research Centre in the Open University. She has undertaken some fascinating work with children as researchers, which you can hear about from the young researchers’ own perspectives.
Categorisation involves grouping your coded text together into bigger concepts or themes. You can use these categories to complete a set, and to use each set as the basis of a chapter. For example, the orange-coloured sweets might represent one set; pinks, reds, and purples might be a second set, and; greens and blues a third set. As we discovered in the Quality Street exercise, there are many possible ways to categorise the codes, and these will be driven by your various research questions, interests, and values, as well as by the data. You should look out for the frequency with which certain ideas occur, similarities between ideas, and corroboration, where numerous data confirm an idea as an important category. You might also find some surprises along the way. Categorisation, then, is an iterative process.
See, for example, how Tania de la Croix describes the coding process in her fascinating research on emotional labour and youth work (there is a link to the full article below), and how this facilitated her development of certain themes, which she reflects on in her methodology section:
Analysis started with detailed line-by-line coding to enable me to become deeply familiar with each interview, moving towards a more open and instinctive analytical approach to draw out common themes. I wrote up tentative theories as I went along, comparing these ideas within and between interviews, with relevant literature and with my ethnographic youth work practice journal. I had not expected or intended to write about love and passion and only read about emotional labour after analysing the interviews. If I had preconceptions about my findings, it was that these youth workers might be demoralised at a time of increasing control and bureaucracy combined with impending cuts and redundancies. While these aspects of workplace reform were indeed key issues for the interviewees, this made it all the more striking that love, passion and enjoyment also arose as strong themes. (de la Croix, 2013: 35)
Note that she mixes inductive and deductive coding, allowing herself to be guided by both the data and her theoretical knowledge. Note also that she lets herself open to new ideas; she points out that she explored literature on emotional labour after she had collected and analysed the data. In the next section, I again take de la Croix’s research article as an example of how to explore these themes for presentation.
Stage 3: Presenting Findings and Analysis
Having organised the data into codes (Stage 1) and the codes into categories (Stage 2), you are now ready to begin presenting the data and reflecting on the data. In de la Croix’s article, she presents and reflects on the data in three subsections, including:
- Love and passion in youth work
- Exploited emotions?
- Passion and resistance
In the first of these sections (pp.35-40), she eases the reader into the data. This section primarily focuses on presenting the ‘feel’ of the data to the reader, and is effectively engaging. Note how she presents findings in a generalised way:
Some focused primarily on building relationships with peer groups, talking with them and engaging them in activities such as quizzes, sports, games and even circus skills, while others worked mainly with individual young people who wanted support. Some did most of their work on the streets, whereas others used mobile youth buses and local community buildings. Although their approaches differed, they all emphasised the importance and satisfaction of getting to know an area and its people. (ibid: 35-6)
Thereafter, she peppers the section with illustrative quotes from her data. She is not too heavy-handed with the citations, but gives just enough to evidence her arguments. This can be difficult to get right. If you have developed good rapport with well-chosen research participants using solid data collection instruments, then it is likely that you will have many beautiful quotes that you would like to include, but you will need to be ruthless in choosing only a few, which serve a purpose in supporting an argument or idea. Having presented some direct quotations from her data, de la Croix makes it explicit why she has chosen to include these:
The section above includes a relatively large amount of interview data in order to build up a picture of the passion with which these youth workers discussed and engaged with their work. (ibid: 39)
In her second section (pp.40-44), she discusses the data outlined in section one with reference to theory, and in particular, to the concept of emotional labour. Unpacking this section, she presents:
- An overview of the concept and the research(er) that is associated with establishing this concept.
- The relevance of this concept to public sector work and caring professions.
- Its relevance to contemporary youth work.
- A critique of the concept and its further development in relevant research.
- Reflections on the importance of the concept for understanding youth work.
- A vignette from her primary research that encapsulates her arguments.
- Final reflections to hammer home the argument.
A journal article is slightly differently structured to a Masters dissertation, so you may find it more appropriate to include parts 1-3 in your literature review chapters. Then, when it comes to interrogating the concept with reference to your data and research findings, you will have already mapped out its meaning and significance.
In the third section (pp.44-47), de la Croix considers the ‘So what?!’ of her contribution in further depth. She achieves this by again drawing in some illustrative quotations, by reflecting on the contemporary context, and by pointing out explicitly the implications of her research findings for youth work practice.
Across these three sections, then, de la Croix skilfully balances the presentation of the findings with theoretical development and with reflections on youth work policy and practice through a tightly structured discussion that flows beautifully and maintains the reader’s interest throughout. Her structure follows the Results-Analysis-Discussion format that is typical of qualitative research dissertations and empirical articles. You don’t have to stick rigidly to this approach, but it is an effective one. Bear in mind that a journal article is approximately 5-7,000 words. Given that your dissertation is either 15,000 or 20,000 words (depending on the route you’ve taken into the Masters), you will have space to expand your themes and your analysis and discussion accordingly.
Over to you!
de la Croix, T. (2013) ‘I just love youth work!’ Emotional labour, passion and resistance, Youth and Policy, No. 110, May 2013: 33-51.
Qualitative data analysis: Classic texts
Glaser, B. and Strauss, A. (1967) The Discovery of Grounded Theory, Chicago: Aldine
Miles, M. and Huberman, A. (1984) Qualitative Data Analysis, London: Sage
Silverman, D. (1993) Interpreting Qualitative Data, London: Sage
Strauss, A. and Corbin, J. (1990) ‘Grounded theory methodology: an overview’ in N. Denzin and Y. Lincoln (eds.) Handbook of Qualitative Research, Sage: Thousand Oaks, CA