Data analysis is an essential stage in all research because it helps researchers to test their hypotheses or answer the research question (Walsh, 2001). The process of analysing qualitative data involves primarily ‗examining people‘s words and actions‘ which in essence means that ‗qualitative research findings are inductively derived from the data‘ (Maykut and Morehouse, 1994: 121). The data analysis is an ongoing research activity which allows the research design to emerge over time.
At this stage, the researcher became aware of ‗epoche‘, a process that the researcher engages in to remove, or at least become aware of ‗prejudices, viewpoints, or assumptions regarding the phenomenon under investigation‘ (Katz, 1987: 36, cited in Maykut and Morehouse, 1994; 123). The researcher dealt with this by recording all her assumptions prior to analysing the research.
Maykut and Morehouse (1994), citing Katz (1987), believe that the setting aside of one‘s assumptions is crucial in phenomenological research in order to be able to assess the data without prejudgement or imposing meaning too soon. The researcher has to be both an ‗insider‘ (aware of one‘s thought processes) and an ‗outsider‘ in order to attempt to judge the material as objectively as possible. Wax (1971) believes that it is
only by achieving this that one can ‗assure a mental position peripheral to both; a position from which they will be able to perceive and, hopefully, describe the relationships, systems, and patterns of which an inextricably involved insider is not likely to be consciously aware‘ (p.3). Lincoln and Guba (1985) refer to this dichotomy between the insider, outsider (or subjective/objective perspective) as ‗Perspectival‘, and argue that the road to greater objectivity can be shortened by subjecting qualitative research to rigorous and disciplined analysis.
The inductive approach adopted for this study means that the data collected relates to a focus of enquiry, and hypotheses are not developed a priori. Therefore, there are no predetermined categories for the data. These emerge from the data through the process of inductive analysis. Inductive reasoning is concerned with moving from the particular to the general. It is important to note that inductive logic is not so much concerned with valid inferences but rather which inferences are probable given the evidence and the data on which those inferences are made. The constant comparative method was used to conduct an inductive analysis of qualitative data (Lincoln and Guba, 1985: Glaser and Strauss, 1967).
Analysing data has been likened to walking through a maze; with many routes available depending on the approach one takes. However, analysing data is more comparable to ‗chaos theory‘ in physics, which is actually a very highly organized and sophisticated way of understanding phenomena, and therefore the label ‗chaos theory‘ as in physics, is a misnomer (Griffin, 2001). This researcher would rather relate the analysis of data to that of ‗organising chaos‘. Data analysis is said to consist of four main elements: interpretation; coding and organising; application of counteracting theories; and the testing of alternative explanations using a ‗knock out‘ method (Anderson and Arsenault, 1998).
There are two main approaches to data analysis: analytical and thematic. The former takes the literature and theoretical background and uses them as an organisational framework, while the latter organises the data into descriptive themes. It is also possible to use both these approaches, by organising the analysis (Phase I) according to emergent themes (Phase II), and then extending the analysis to ‗examine the findings in consideration of existing literature and theory (Phase III) (Bereday, 1964: 158). This is the approach that was taken in this research. Bereday‘s approach, as described above, was used as the macro analytical tool with which to compare and analyse the data. The conceptual framework illustrated in chapter 1 was used to organise the data, while Glaser and Strauss‘s (1967) ‗Constant Comparative Method‘ was applied when analysing the data.
3.7.1 Preparation for Data Analysis
This section describes the general procedures undertaken prior to and during the analysis of data. The researcher went through three main phases when preparing for the analysis of the interviews. The documentary evidence gathered was also subjected to a similar process of analysis and will be described later in this chapter. The preparation for the analysis of the interviews may be described as follows:
1) Listening to the taped interviews: initial familiarisation. Initial familiarisation involved listening to the tapes and noting points of interest which occurred during the interview process. This process was repeated some time later, before transcribing, in order to hear the interview ‗afresh‘.
2) Transcribing the interviews: through familiarisation. Transcribing the interviews was a long and demanding process. The time varied depending on the quality of the taped interview; the speed at which the interviewee spoke; the background noise; the accent of the individual; the clarity of the interviewee‘s voice; and the volume at which it was recorded. However, this process allowed
the researcher to become thoroughly acquainted with the data. As Payne & Payne (2004:132) mention the transcription of recording is probably ‗the most tedious and time consuming aspect of these interviewing methods‘.
3) Highlighting points of interest: preliminary stage in the identification of
themes. Highlighting the points of interest is theoretically the preliminary stage
of analysing the interview transcripts. This process was achieved by firstly underlining words stressed by the interviewees themselves, and secondly, other points of interest noted by the researcher were highlighted in bold.
The preliminary focus was on the ‗sub sample‘ of the data, i.e. the data which looked most promising or the most convenient to categories and which was relevant to the research focus (Boulton and Hammersley, 1996: 290). They pointed out that:
„with a small amount of data, it is often difficult to go beyond the description of a few key themes. A larger amount of data may allow greater development of understanding of the perspectives and behaviour of the people being studied, especially in terms of looking for relationships among categories‟. (P.290)
The data was then coded into categories which showed where the data had come from. This procedure was carried out for all respondents, colleges, schools, offices and locations to aid the constant comparative method. Thus, when the answers of different responses to the same question were collected, it was possible readily to identify where each response came from by referring to the above coding system. The second stage to analysing the data is to ‗unitise the data‘.
3.7.2 Unitising the Data
At the beginning of every transcript, documentary evidence and field-notes were unitised. The phrase ‗unitising the data‘ was first coined by Lincoln and Guba (1985). This step involves identifying chunks or ‗units of meaning‘ in the data (Maykut and
and actions of the participants in the study. This is framed by the researcher‘s own focus of enquiry and is achieved by ‗first identifying the smaller units of meaning in the data, which will later serve as the basis for defining larger categories‘ (p.128). Each unit of meaning identifying in the data must be understandable, independent of explanations, and although smaller units of meaning, the main unit must be able to stand alone (Lincoln and Guba, 1985).
To arrive at ‗units of meaning‘ the researcher must first read all the transcripts, documentation and field notes thoroughly. Each unit is labelled according to its origin. Cutting and pasting together the answers to questions achieved the unitising the data. So, for example, three main folders were used. The first folder represented everyone‘s answers to a particular question from one particular institution. The second folder consisted of every answer to that question from all locations of that country, and finally, the third folder consisted of everyone‘s answer to that particular question from both countries, and locations therein. From this stage, boxes were drawn around the main unit of meaning in each response to that particular question.
The next stage is to isolate one word that conveys the essence of what was being said. These are then stored in a separate file with the appropriate references attached, so as to be easily referred to as necessary. This is done so that every piece of information is unitised. Units of meaning may vary from a word, to a sentence, to a paragraph, depending on both the informant and the point being made or expounded upon. Lincoln and Guba (1985) recommend that alongside each unit of meaning a clear account of respondent‘s background be noted, or in the case of documentary evidence, the source of the report and how it was commissioned, if at all, so as to assess the information in the most meaningful light. This will include auxiliary information such as the gender of the respondent and any other personal details which may affect the nature of the response.