CHAPTER 3: IMPLEMENTATION OF THE RESEARCH DESIGN AND
3.6. Data analysis
This study made use of conceptual analysis which is also known as “thematic analysis” by other social researchers. The steps outlined by Palmquist (1993, in Babbie and Mouton, 2006: 492) were followed. They outlined the eight steps for conceptual analysis: (1) deciding on the level of analysis, (2) how many concepts to code for, (3) whether to code for the existence or frequency of a concept, (4) deciding how to distinguish among the concepts, (5) developing rules for the coding of texts, (6) deciding what to do with irrelevant information, (7) coding texts, and (8) analyzing results.
The first step was to decide on the level of analysis. In this regard, multiple or key phrases were used in the questionnaire. The second step was to decide how many concepts to use and code, and in this regard two broad concepts or constructs, “grade progression and promotion” were utilized. These are the relevant key terms around which the questions were constructed. The third step was a decision to code for patterns or themes. The fourth step was to decide
whether to code, for instance, for the data, or be prepared to generalize around the content of the data. In fact the analysis commenced by incorporating both “incidence and generalizing” to reduce subjective bias given that qualitative research is about understanding meaning. In the fifth step, rules for the coding of texts from interviews were developed by charting or mapping the data on a schedule and drawing pointers or “rules” to guide the coding. In the sixth step simplified parameters were set for coding a set of data in terms of what would be included or excluded around a concept, e.g. promotion. In the seventh step, irrelevant information was carefully reviewed and decisions as to whether this needed to be considered or not were made. Finally,, the actual coding and analysis of the results was completed.
The conceptualization of data is a way of organizing and making sense of it, which is the first step in the analysis. The data is organized into categories on the basis of themes, concepts or similar features. Concepts were linked to each other in a sequence or sets of similar categories, which were then woven into theoretical statements. The process was guided by the research question and generated new sub-questions, motivating the researcher to higher-level thinking. It also led to theory formation. Miles and Huberman (in David and Sutton, 2004: 195) suggested a list of basic coding prompts: (a) themes, (b) cause or explanations, (c) relations among people, (d) emerging constructs, etc. Coding was thought of as reducing data into conceptual frameworks and as instruments, cases and questions that had to be refined. Data was then summarized, coded, and broken down into themes, clusters and categories ready for interpretation and analysis. Miles and Huberman (1994: 56) described coding as “tags or labels for assigning units of meaning to the descriptive or inferential information compiled during a study.” Codes were usually attached to “chunks” of varying size; “words, phrases, sentences, or whole paragraphs, connected or unconnected to a specific setting,” similar to what was mentioned in Punch (1998: 204). Miles and Huberman (1994: 56) noted that “They took the form of a straightforward
category label or more complex one.” This was done to reduce data (field notes, texts, cards, observations etc) to convenient and manageable proportions and the main goal was to facilitate the retrieval of data segments categorized under the same codes.
The following strategy, as described by Seidel and Kelle (1995: 55-56 in Coffey and Atkinson, 1996: 29), was used: (a) notice relevant phenomena, (b) collect examples of these, (c) analyse the phenomena so as to find commonalities, differences, patterns, and structures (teasing and expanding the data) which are heuristic devices for further discovery. There were three types of material: (a) unstructured questionnaire, (b) field notes (unstructured), (c) documents and circulars, etc.
The former was coded to a certain degree; hence what was required was a close reading to identify aspects that were significant within them. Sapsford and Jupp (1998: 290) stated that the focus of inquiry is clarified over the course of data collection and analysis. In order to make sense of the data, analytical categories were used first. Thereafter, categories to which the data related and had relevance to the research were written down. Any recurrences were noted which illuminated patterns of an individual educator’s perspectives, opinions, feelings and understandings of OBE, grade progression and promotion. The third step entailed the gathering together or pulling of segments of data from different parts of the interview record and field notes that were relevant to the same category. Categories, which emerged from the data, had a direct bearing and effect on answering the research questions posed in the structured questionnaire.
These categories were drawn from various strands or sources. At this stage the data confirmed expectations to a certain extent, but more analysing was required to ensure that the research question was fully challenged. A number of categories were generated so that the information or data held could be incorporated in the
content of the presentation of categories, patterns and structure. These categories informed the thesis used in chapter four.
There is an understanding that the results of a qualitative investigation might be checked against a quantitative study. The aim is to enhance the reliability, validity and credibility of the study. This is done through triangulation (Babbie and Mouton 2006: 275-276) and by writing extensive field notes, member checks, peer reviews, reasoned consensus, audit trails (to let the respondent speak freely without distorting what they say while they are interviewed), etc.