1.5 The research methodology
1.5.5 Data collection
In qualitative research, the researcher approaches the participant with the aim of finding out about his/her experiences (Holloway & Wheeler 2002). The researcher immerses him/herself into the context of the participant to collect rich, in-depth data (Terre Blanche et al. 2006). Different data collection methods, most commonly
interviews, observations and focus groups, are used to produce qualitative data. More recently the use of photo voice (Steyn 2009), online narratives (Ardoin, DiGiano, O'Connor & Holthuis 2016), participant journals and social media (Lourens 2013), amongst others, have also been used effectively. The important consideration in the choice of a data collection method would be to choose the method that would provide the most insightful data and is feasible at the same time (Burnard 2004).
My aim was to provide an analysis of the transitions into first-time employment of graduates who meet a wide range of criteria (cf. 1.5.4 The participants). I therefore needed a relatively large sample. I further realised that the participants were employed full-time and had specific, busy schedules. It would have been challenging for them if their participation was too time-consuming. It would also be challenging for me if I had to work around more than one participant’s schedule to stage a focus group, for example. The method I chose therefore had to be as time effective as possible, and had to enable me to easily arrange an appointment around the participant’s schedule, but still have the potential to provide sufficient information while at the same time ensure a high response rate. I therefore opted for semi- structured interviews.
Semi-structured interviews involve an interview schedule with pre-determined questions, but the order or wording can be modified, questions can be omitted and/or additional questions can be added (Barriball & While 1994). Apart from the fact that I was well trained in conducting semi-structured interviews, I was aware of specific advantages of the method that the study could benefit from. Firstly, through semi- structured interviews I was able to uncover information which I probably would not have been able to access through observations or questionnaires. The semi- structured interview provided a natural way of interaction with a participant during which mutual understanding could be ensured. I was able to rephrase or simplify questions that were not understood by the participant and I could clarify what the participant meant when I did not understand (Alshenqeeti 2014; Polkinghorne 2005). Secondly, since a semi-structured interview involves an interview schedule, it ensured consistency amongst the 46 interviews (Barriball & While 1994). Thirdly, the interview schedule ensured that I covered all the topics I wanted to collect data on, but simultaneously provided flexibility since it was not as fixed as a structured
interview (Alshenqeeti 2014). I was able to follow the participant’s lead during the interview (Leech 2002) and allowed the participant to freely recount his/her experiences. The last two advantages pertain not only to semi-structured interviews, but to personal interviews in general. Fourthly, while it was not easy to arrange appointments with 46 participants, only one participant did not show up for the appointment. The widely acknowledged high response rate of a personal interview made the method particularly appealing (Alshenqeeti 2014). Lastly, the individual interview affords non-verbal indicators such as body language and facial expressions that are useful to evaluate the validity of the participant’s answer and the intensity of the experience described (Barriball & While 1994; Denscombe 2014).
Individual interviews, like any research tool, have some drawbacks. Firstly, conducting 46 interviews across two provinces was time-consuming with regard to firstly collecting, and thereafter transcribing and analysing the data (Alshenqeeti 2014). Secondly, Hammersley and Gomm (2008) point out that what people say in an interview is determined by the questions they are asked or what they think the interviewer wants. I therefore needed to be very careful not to ask leading questions or put words in the participant’s mouth. Thirdly, related to the previous drawback, this method has the potential of subconscious bias from both the researcher and the participant (Van Teijlingen 2014). Ideally another form of data collection method should thus accompany semi-structured interviews to ensure the credibility of the data. I will return to this in section 1.5.8. Despite potential drawbacks, the advantages of individual, semi-structured interviews as data collection method outweighed the disadvantages, and the interviews produced a host of rich in-depth data.
As mentioned above, an interview schedule (Appendix A) was developed to guide the interviews. The development of the interview schedule is a key phase in the qualitative research process. During my literature review I identified areas of interest and relevance that should be covered in the interview. The broad areas were categorised into more manageable groupings (Barriball & While 1994), and this resulted in four groupings namely 1) HE, 2) securing employment, 3) professional success and 4) HE’s relation to work. As the review of literature continued, I added questions to each grouping. The final stage of the interview schedule development
entailed the consolidation of questions per grouping in order to simplify the schedule and to ensure that replications did not occur. I tested the schedule during discussions with graduates and colleagues to ensure that the meanings of the questions were clear.
The semi-structured interviews of approximately forty minutes each were conducted in and around Cape Town and Johannesburg from March to November 2015. Most of the interviews were conducted during office hours at the participants’ workplaces, whereas a few were conducted in public places like coffee shops. A few workplaces did not have private spaces to conduct the interviews and communal areas had to be used. Some participants were sensitive not to use worktime for non-work related matters and often the only time available was participants’ lunch hours. For participants working shifts, appointments had to be arranged on their off days. The interviews were recorded and transcribed by a transcriber where after I checked the transcriptions for accuracy.