Simplistic analysis of the data was ongoing throughout its collection, beginning as early as the interview itself. Impressions were created by the children’s non-verbal responses as well as from their spoken words. Therefore, notes were made and initial judgements were made regarding the children’s confidence and attitudes to school. The enthusiasm of some of their responses would not be captured by transcription and so it was important to note non-verbal as well as verbal responses either during or immediately after the interview.
By doing the transcription as quickly as possible after the interview I was trying to prevent myself from ending up with an overwhelming volume of material to be processed. As previously mentioned in ‘Methodology,’ the data for individual classes were also given to the class teacher and a short meeting with the teacher and myself arranged. This was to enable individual class issues to be discussed privately. After reviewing each class’s collective data, I followed the recommendations of Altrichter et al (1993), by writing a summary.
The purpose of this was two-fold: firstly, it enabled me to have easy access to the data later and, secondly, it gave me an immediate overview of what the data offered relating to my research questions. Also during each stage in the research process, ideas and theories came into mind and these were jotted down. Altrichter refers to these as ‘theoretical notes’, which seems a rather grand description for, what in my case, were quick jottings, trying to capture each stage in my developing thought process.
Although the data collected can be most accurately described as being qualitative, there were also a surprising amount of quantitative measures within them. These arose from the counting and comparing of the comments of individual children in order to make certain generalised statements or from yes/no answers to closed questions. These will be reported on accordingly. I had been warned by tutors that data may feel out of control and that it was all part of the process, but I still felt unprepared for the feeling of total panic after I had printed out the full set of data.
Although I felt I had a good understanding of the implications from class and year group data, I was unsure as to how one began to theorise and generalise about so many seemingly disparate statements generated by the respondents. I looked back at my notes from one of the Foundation sessions and was reassured as I read ‘you need to lose control for free flow of thoughts’. At least I had fulfilled one of the criteria for the M. Ed. Course! I was also acutely aware of the sensitivity of the issues relating to research with children and the subjectivity of their perceptions.
Throughout the process I had also been aware that good qualitative data should be as unbiased as possible and yet the very nature of the method leaves it open to criticism. I, therefore, returned to the literature on data analysis to help resolve or justify my anxieties. Issues concerning analysing quantitative data Miles (1979, p. 591) has described qualitative data as ‘an attractive nuisance,’ commenting: ‘… But the most serious and central difficulty in the use of qualitative data is that methods of analysis are not well formulated……
The analyst faced with a bank of qualitative data has very few guidelines for protection against self-delusion, let alone the presentation of ‘unreliable’ or ‘invalid’ conclusions to scientific or policy-making audiences. ‘ The method of data collection is also open to prejudice as it is ‘collected by human beings, and because people are interested in certain things and not others, selections are made. People tend to record as data what makes sense to and intrigues them’ (Le Compte, 2000). This reassured me that my feelings were justified.
However, I was also mindful that the quality of my data might have been restricted by the possible limitations of my questions, questioning, the choice of respondents and the instrument used. Thereby, when interpreting the respondents’ views, all these variables need to be considered. Most of the data analysis undertaken has been using the intuitive, interpretivist approach but following guidance (such as Hitchcock and Hughes, 1989), which has directed me to return to the data and to be aware of my own bias having an undue influence.
I have also tried to be aware of the effects of both tacit and formative theory, as Le Compte describes these as: ‘the sources of selectivity (and bias) because they create something analogous to a filter that admits relevant data and screens out what does not seem interesting – even if, with hindsight, it could have been useful’ (p. 146). Hopefully, by being informed and aware of these difficulties, I will have been protected from the ‘danger of rampant subjectivity where one finds only what one is predisposed to look for’ (Lather, 1986 – p. 259).
The Method of Analysis I found Miles and Huberman’s (1984 – p. 23) summary of the essential elements of the analytical process, helpful in guiding me through the process. They are as follows: This first section for me could well have been entitled ‘reading and immersion in’ data. I listened to the tapes, read through the transcribed data, summaries and theoretical notes all several times. I checked again that the transcripts were an accurate account of the interviews and found the tapes to be more helpful in recalling the interview than the transcripts.
This was not because of any inaccuracy in the transcripts, but by replaying the tapes memories of the interview seemed to be reconstructed more vividly. They also gave me time to critically question and the opportunity to correct possible false interpretations. As Drever states (p. 62), a transcript cannot ‘capture the interplay of question and answer and the emphasis and added meaning construed through intonation and body language’. Through reading and immersion in the data, patterns began to emerge which led naturally into the next section – that of selecting and categorising the data.
Selecting and Categorising Data I then began to select or sort the data into categories or groups by comparing and contrasting statements (Glaser and Strauss, 1967), and by identifying patterns or ‘things that go together’ (LeCompte – p. 150). Some of the selection of data was determined by the frequency of certain responses, other by the fact that they are not mentioned at all. As mentioned earlier, the interview schedule was specifically divided into categories using Hay McBer’s nine dimensions of classroom climate measures (these are described in the questions section of Methodology).