Chapter 5 Methodology Methodology
5.6. Procedure for Main Study
5.6.2. Procedure
The process of data collection began upon receiving permission and ethical clearance from the University of KwaZulu-Natal’s Humanities and Social Sciences Research Ethics Committee (UKZN’s HSSREC). Details pertaining to the ethical clearance required in this study are discussed later in this chapter under ethical considerations, subsection independent review, paragraph 5.8.1. An ethical clearance letter is attached as Appendix E.
5.6.2.1. Gatekeepers’ permission.
Prior to collecting data, meetings were held with community gatekeepers in order to gain entry into the research sites and to obtain access to the participants. Gatekeepers usually have a say over who is permitted or denied entry or access to research participants. According to Terre Blanche and Kelly (1999), gatekeepers are usually parties who have vested interests either in the issue at hand or in the wellbeing of the potential research participants. These authors advise that it is beneficial to be on good terms with gatekeepers in order to ensure that the research study does not get sabotaged along the way (Terre Blanche &
Kelly, 1999). Different meetings were scheduled and held with the municipal managers and the person in charge of the SAPS units in both Greater Kokstad and Kwa Sani Municipalities where this research was conducted. At these meetings, the background to the study, its aims and objectives, as well as the data collection procedure were explained in some detail. Once the gatekeepers were satisfied and understood the study and its procedures, they signed and stamped a copy of the ethical clearance letter that had been obtained from UKZN’s HSSREC. Copies of these letters signed and stamped by gatekeepers from Greater Kokstad and Kwa Sani Municipalities are attached as Appendices F and G, respectively. These gatekeepers also provided the researcher with their telephone numbers so that they could easily be contacted in the event of problems being encountered in the community.
Interestingly, a ward committee member at Kokstad approached the team on the second day of data collection and enquired what the study is about. He was pleased to hear about the study and he also provided his contact number in case the team experienced difficulties in the community. At Kwa Sani, a ward councillor was selected during a routine random selection of participants, and she
agreed to take part in the study. Upon completion of the questionnaire she showed eagerness to help the research assistants to recruit participants. However, she was informed that the recruitment of participants is based on a certain random technique and that she therefore would not be able to recruit members of the community that she knew to participate in the study.
The heads of the SAPS from both research sites also assured the research team of safety and security. There was an agreement in these meetings that the research team would be identifiable by carrying specific bags with certain colours and identifiable name tags. On the days of data collection, the patrol vehicles of the SAPS were alerted to the presence of the research team in order to increase safety and patrol activities in those communities. A caution by Kelly (1999) that problems often arise when researchers are introduced into the research sites by gatekeepers because the participants associate the researcher with the vested interests of the gatekeeper was taken into account. As a result, the gatekeepers did not accompany the research team during data collection, but the SAPS did provide safety and security from a distance by visible patrols in the areas where data collection took place.
5.6.2.2. Data collection.
The researcher acknowledges that designing a strategy to recruit participants from their homes would have been much more desirable in this study.
However, given the scope of this research, the geographical nature (rural, scattered) of the research sites and cost constraints, it would have been very time- consuming and too expensive to conduct a randomised household survey for this study. Instead, a randomised public recruitment strategy was adopted.
Randomisation is a process that gives everyone in the population an equal chance of being part of the sample (Weisberg et al., 1996). Johnson and Onwuegbuzie (2004), as well as Zheng (2015), have argued, in their discussion of the quantitative approach and cross-sectional design, that random sampling increases generalisation about the whole population, especially when the sample is sufficient.
Data was collected between August and December of 2014. Participants in this study were randomly recruited from public places that usually attract a large
number of people, such as shopping complexes, taxi or bus stations, township streets, and some from their homes. To select participants, a specific random sampling technique called the systematic selection procedure was used with the aid of a computer program called Randomizer. In the systematic selection procedure, according to Weisberg et al. (1996), a random number is selected to choose the first research participant, and then some people are skipped to choose the next person based on the next random number, and so on. A Randomizer is a computer program that generates a set of random numbers. During each day of data collection, a Randomizer would be computed to generate 5 new sets with unique 50 integers in each, with each integer having a value between 1 and 200.
Appendix H is an example of 5 sets of random numbers generated by this computer program. Research assistants would each be given a set of these random numbers to use to select participants. This means, if the integers of the first set are 1, 4, 11, 12, 15, 22, 35, the first person to be met when recruiting participants would be approached and asked to participate in the study, this would be followed by the 4th person, then the 11th person, and so on. In this way, selection bias was eliminated. 787 participants were intended to be recruited for the study.
This number was informed by the population sizes of the two municipalities (Greater Kokstad and Kwa Sani), as depicted in Table 5.1 in page 122 , as well as Watson’s (2001) guidelines for establishing a base sample size. This table assumes a 95%
confidence level, a 5% margin of error and 50% degree of variability.
A brief meeting was held with the research assistants every morning prior to them and the researcher going out to collect data. The aim of the meeting was to motivate and remind them of the importance of adhering to recruitment and data collection procedures, more specifically to adhere to the ethical procedures required in a study of this nature. These meetings also provided an opportunity for the researcher to answer or clarify questions or issues if the research assistants had any. Research assistants were then dropped off at different locations for data collection. These were mostly places that tend to attract a large number of people such as towns, shopping malls, bus or taxi ranks. Other participants were recruited from the streets in the communities.
When a potential participant was identified using random numbers generated by Randomizer, they were approached and asked to participate in the study. The research assistants first introduced themselves and explained the reason
for approaching the potential participant. The first objective would be to establish a rapport with and to ascertain whether the person was over the age of 18 (because no one below the age of 18 was allowed to take part in the study) and if they were willing to participate in the study. If the person refused to participate, the research assistant would check what the next number was on the list of random numbers and would begin counting to obtain to the next possible participant. If people were met in a group, the counting would still be conducted in the same way and only the person identified through this counting would be asked to participate. If they agreed to participate in the study, then they would be requested to move away from the group to ensure less distraction.
Once a participant had agreed to take part in the study and rapport had been established, the study was introduced to the participants. The rationale for the study and information contained in the informed consent form was explained (Appendices B and D), and instructions pertaining to filling in the questionnaire were also provided by the research assistants. Particular emphasis was placed on confidentiality and anonymity, and research assistants were specifically instructed to ensure that participants understood the role of their participation and their rights.
It was also made very clear to all to participants that they were allowed to withdraw, without any repercussion, from the study at any point should they feel the need to do so, and that their participation is absolutely voluntary. Participants were asked whether they understood the instructions or had any questions before completing the questionnaire. The research assistants were available to answer any questions or to provide clarity where necessary during the completion of the questionnaire. Participants had a choice of completing the English or isiZulu version of the questionnaire. For participants who could not read and/or write, the research assistants read the questionnaire aloud for them and recorded their responses. Each participant was randomly given a questionnaire depicting only one of the three vignettes representing depression, schizophrenia or alcohol dependency, as described earlier.
5.6.2.2.1. Challenges experienced during data collection.
Although data collection was a smooth process, there were some difficulties that were experienced and worth noting. Firstly, the time required to complete the questionnaire was experienced by some participants as too extensive. The average time taken to complete the questionnaire was
approximately 40 minutes. Given that participants were asked to take part in the study while going about their daily business, some indicated that they were in a hurry and therefore unable to participate in the study. There were also participants who began the process of completing the questionnaires but then withdrew from the study because they were in a hurry to go somewhere. However, it became apparent that if participants were informed that it would take them close to an hour to complete the questionnaire and their participation was valued, that this increased the questionnaire completion rate. This adaptation was incorporated into the standard recruitment procedure at a very early stage of data collection and it proved successful.
Another problem experienced during data collection was to do with participants who wanted to know more concrete details related to the vignettes represented in the questionnaire than the information that was provided. For example, some participants wanted to know if the people depicted in a vignette had matriculated, what type of job they were engaged in, if they had siblings and what their birth order was, as well as whether some traditional rituals had been performed, just to mention a few. To deal with this problem, the research assistants were encouraged to tell participants that not much information is known about this person in the vignette, and that they were only expected to answer on the basis of what little information was provided.
The rate at which data collection was taking place also became a concern to the researcher. On average, each research assistant would have about eight completed questionnaires per day. There were days where some had only 5 completed questionnaires. The research assistants reported that older participants tended to take more time thinking about their answers and asking clarity-seeking questions. It was also reported that some participants, after completing the questionnaire, would want to recommend certain people to be included in the study. These participants were very disappointed when they were told that there was a specific procedure to be used to select participants and that their recommendations in this regard would not be followed through.