SECTION 3 RESEARCH DESIGN AND METHODOLOGY
3.6 Data Analysis
The data analysis was based on the examination of raw data collected during the interview process and the interpretation of it into useful information that informed the research topic. Data analysis can be described as “a systematic search for meaning” (Hatch, 2002: 148 as cited in Leech & Onwuegbuzie, 2007) and Basit (2003) describes it as an attempt by researchers to develop a better understanding of what they have studied and to refine their interpretations. The overall intent of my data collection was to develop an in-depth understanding of the learning orientations of part-time adult students, their satisfaction with the college experience in terms of realising their own aims and the contribution of college learning to their career development in a globalising world. I organised the data according to the themes that occurred in the literature review, research questions and interview schedule.
I took an iterative approach when I collected the data which according to Bryman (2001: 389) implies that data analysis will commence after some data have been collected and that the results of the analysis will then shape further data collection. In this way, I was able to adjust the interview schedule if necessary and I was able to determine if I needed to add or adjust any questions. By using this method I wanted to avoid discovering after collecting the raw data that I should have asked more or different questions.
The object of analysing qualitative data is to develop categories and relationships that inform the topic of research (Basit, 2003). My analytic strategy incorporated features of grounded theory which I used to analyse the data I collected. According to Bryman (2001: 397) grounded theory represents the most influential strategy for analysing qualitative data but the extent to which the tools associated with this strategy are used varies between studies. One of the tools of the grounded theory approach identified by Bryman (2001: 391) is coding, which leads to categorisation and constant comparison analysis (Bailey & Jackson, 2003). Coding entails breaking down the raw data into component parts which are given names and used to interpret the data. It is a process that allows the researcher to ask questions, compare data, create and change categories with the aim to find commonalities, differences, patterns and relations between various pieces of data (Basit, 2003; Seidel & Kelle, 1995 as cited in Basit, 2003; Thorne, 2000). The process of comparing data continues until the data collected from each interview have been compared with all the other sets of data (Thorne, 2000).
Coding is central in data analysis and codes can be described as links between the raw data and sets of concepts or ideas, or categories, which will enable the researcher to interpret the data (Coffey & Atkinson, 1996 as cited in Basit, 2003). Miles & Huberman (1994 as cited in Basit, 2003) and Leech & Onwuegbuzie (2007) argue that categories used in data analysis can be formulated before the commencement of fieldwork, a method which corresponds with a deductive approach. These categories could be based on the conceptual/theoretical framework that informed the research as well as the research question. In my analysis I favoured a deductive approach and I used the literature and my research questions to develop my interview questions. I commenced the coding process with a provisional list of categories which I searched for in the data. However, in line with a more inductive approach to data analysis, I remained open to the possibility of new codes and categories emerging during the coding process.
Bryman (2001: 398) suggests coding data as soon as possible after collection, effectively coding as you collect the data. The reason for this is that it will sharpen the researcher’s understanding of the data and also avoid the researcher becoming overwhelmed by the magnitude of the data collected. Bryman (2001: 398) also suggests that if data collection involves audio recording interviews, the researcher starts transcription as soon as possible. Critique of the coding approach cautions that it could present a possible problem of losing context as it takes text out of the context in which it appears and it could also result in the fragmentation of data (Bryman, 2001: 400). However, Bryman (2001: 402) notes that coding is a prominent method for qualitative data analysis
because it is widely accepted in the research community and it is associated with grounded theory which is a very influential framework.
After I collected the data by audio recording personal interviews with participants, I transcribed the interviews and if necessary translated it to English. I manually coded the data in order to identify concepts and categories and indicated what each participant said by writing their name next to their response. If another participant had a similar response I added their name to the statement. This allowed me to establish how many times a particular response was repeated and by whom. In addition, similarities and differences in the opinions of the participants were highlighted by comparing the categories. After identifying the trends and patterns that emerged from the data I interpreted my findings by comparing the results to the conceptual/theoretical framework I discussed in my literature review as well as my research questions. The data analysis process discussed above enabled me to analyse the data and develop new insights relating to the FET sector.