Data analysis


 

 
I thought this excerpt on data analysis from my action research report on graphic novels might be helpful (Di).

 

 

'One of the reassuring aspects for teachers engaging in action research using qualitative data collection methods is the potential flexibility in the analysis of data. There are ‘no strict rules which have to be followed in qualitative analysis’ (Williamson & Bow in Williamson 2002, p.293). Instead, in an approach described by Tesch (in Shenton 2004, p.144) as ‘interpretational analysis’, researchers are able to select methods of analysis that are most appropriate to organise and bring meaning to the collected data. Interpretational analysis is a bottom-up approach, therefore, where researchers immerse themselves in their pool of data and, by identifying patterns and themes, gain understanding and create meaning.

 
In accordance with the majority of qualitative analysis, the analysis of data for this project was both formative and summative (Shenton 2004, p.159). In regard to the former, data was already being examined in the early stages of collection. For example, during my interviews with participants I was continuously analysing responses, if only at a superficial level, and using this analysis to either guide the discussion along avenues that I felt warranted further exploration, or to steer it away from matters that I considered off-topic. Similarly, formative analysis was taking place during and after interviews and classroom observations as I made field notes in which I noted my impressions or noted quotations that I thought might be useful. Finally, as I made written transcripts of my key-informant and focus group interviews I was already noting recurrent patterns that might later be used to determine themes.
 
The first step in my summative analysis was to read, in its entirety, the data I had collected in the form of transcripts, journal entries, observations, field notes, photos and surveys. I felt it was essential to develop an overall understanding of what I had gathered and be familiar with it before beginning its deconstruction. During a subsequent rereading of the data, Lincoln and Guba’s (in Shenton 2004, p.148) feels right/ looks right procedure was used to identify categories for like-minded data. These categories were created on the basis of both my understanding arising from asking myself questions such as: ‘What is this person saying and what are the issues being raised?’ (Williamson & Bow in Williamson 2002, p.296), and on the recurrence of particular words or phrases in participants’ responses. As I identified categories I noted these in my research journal and, to ensure consistency in the identification process, also noted a definition. I attached a colour code to each category and then colour coded the relevant passage in the data. As I moved through the data, passages were either coded according to existing categories, or if there were no such similar units of meaning, a new category was created. This linear method of categorising data is termed the ‘constant comparative method’ (Glaser & Strauss in Shenton 2004, p. 147). As increasing amounts of data were processed, it became apparent that some passages of data fitted the definition ascribed to several categories and so these were multi-coded. After categorising the data, categories were then grouped into themes which were based upon the three general areas of the research – reading attitudes and habits; response to the traditional text of Macbeth; and response to the graphic novel version of Macbeth. These themes then provided the structure around which my research results were analysed and discussed.
 
In the analysis of results there is a heavy reliance on written and spoken quotations from the participants. Not only does this use of the participant voice transport the reader into the reality of the research, but it also enhances the transparency of the research by providing an audit trail, i.e., making visible the raw data from which I created meaning and drew conclusions. Participant quotes used within the body of the text are italicised to give emphasis.'
 

References

 

Shenton, A 2004, ‘The analysis of qualitative data in LIS research projects: A possible approach’, Education for Information, vol. 22, iss. 3/4, pp.143-162, accessed 2 April 2006 via EBSCOhost Research Databases.
 
Williamson, K 2002, Research methods for students, academics and professionals, 2nd edn, Centre for Information Studies, Charles Sturt University, Wagga Wagga, NSW.