Analysis of qualitative data

Data analysis is an activity that students are likely to attend to towards the latter phase of their research project. Students doing independent research projects and dissertations will often engage in proposal work with their supervisors at the start of their research journey, and, in my experience, there is a tendency to omit any clear

articulation of the proposed approach to how the analysis might be done and what type and level of analysis best fit with the remainder of the research design. A robust research plan will think forward and part of this forward planning will contemplate whether the research questions and research methodology lend themselves to a largely deductive approach or a constructed/constructivist approach that is interpre- tivist and inductive. Deductive research tends to be associated with theory or hypothesis-testing inquiries, whereas inductive research has clearer links to ‘grounded theorizing’. Grounded theory is associated originally with Glaser and Strauss (1967) and refers to an approach which starts with data collection, and, as data emerge, constant comparisons are made between data and theory (Davies, 2011).

Thus, depending on the nature of the research being undertaken, some of the pointers towards how the analysis will be affected are already built into the chosen methodological approach. You may be using documents or other people’s data such as crime and victimization data as part of your research. In criminological research, this is most likely to be original data sets – raw data – collected by agencies such as the police or the courts and which belongs to a government department, a criminal justice agency or an academic researcher (Semmens, 2011). Examples of secondary sources include census data and recorded crime statistics (see also Chapter 1). Secondary analysis can be done using data from these sources. Such data will have been collected for some other purpose than your own. You may, however, be

gathering primary data through questioning people via surveys, interviews, focus groups and listening, or through life histories, observations and ethnographic research, and each of these methods will determine, to a greater or lesser extent, the manner in which you will be analysing your data. Documents can be analysed via content and discourse analysis as can interview data from transcripts of interviews. Interviews can also be analysed by making summaries and comparisons, by coding and categorizing to identify themes, commonalities and differences in data sets (see Boxes 12. 3 and 12. 4).

In the literature, there are different definitions of analysis but most definitions capture how qualitative analysis entails disassembling or segmenting and reassem- bling or synthesizing your data. Boeije (2010) offers the following definition:

Qualitative analysis is the segmenting of data into relevant categories and the naming of these categories with codes while simultaneously generating the categories from the data. In the reassembling phase the categories are related to one another to generate theoretical understanding of the social phenomena under study in terms of the research questions. (Boeije, 2010: 76)

This definition emphasizes the emerging character of analysis where a stream of several ‘thinking and doing’ activities, such as sorting, naming and categorizing, go hand in hand. Critically for Boeije (2010), the disassembling part of the analysis enables the researcher to specify which building blocks their research contains, thereby allowing for the emergence of theoretical concepts.

Analysis allows us to discover findings. The analysis we engage in must allow for the accurate representation of the perceptions, views and experiences of our respondents. How you plan to analyse data should, therefore, not be an afterthought but a fundamental part of your initial project plan and, as noted above, within your research proposal there ought to be consideration of how you will approach the analysis. Even these early plans for analysis will ensure that the proposed approach allows for you to attend to meaning. Paying attention to social meanings ensures the research does not become divorced from social reality. Prioritizing social meaning will safeguard the producing of conclusions which are grounded in such meanings and expressed in terms which would be used by the respondents themselves (Davies, 2011: 50). Again, this reminds us of the importance of the interplay between induc- tive and deductive approaches, where one approach is often tempered by the other in effective and rigorous criminological inquiry. It is important ethically that your choice and use of analytical devices and tools is capable of capturing the essence of the original data you have collected. At the same time, it is important that your analysis is capable of shedding light on the research objectives and research ques- tions. Decisions made at each stage of your analysis should be made transparent. Analysis is therefore a critical component of research and how it is accomplished can impact on the quality of the entire research project. There are few texts that get behind the pure abstract principles of doing quantitative and qualitative data analysis

(see Bryman, 2016; Creswell, 2014; and Neuman, 2014, for example). The third section of Jennifer Mason’s (2002) book Qualitative Researching provides some use- ful guidance on organizing and indexing and making convincing arguments with qualitative data. Below, I illustrate how there are various different levels of analysis that can be achieved from qualitative research and how analysis is guided by your epistemological position (see Chapter 6 in this volume) and your perspective on the problem under consideration. These factors underlie the importance of being reflexive. First however, let us focus on my own continuing efforts to grapple with the problems of how to analyse data arising out of a qualitative research inquiry into a very sensitive area of victimization.