Following the development of the questionnaire, data were collected from students who were currently enrolled at Unisa during the time of the study. The data were collected over a period of 5 months, from September 2014 – February 2015. The target population for this study included men and women aged 20 years old and above, registered with Unisa for any degree programme of study and residing in five regions. The exclusive criteria were students registered with Unisa for non-degree modules, for example bridging courses. Data were collected from a quota sample of 400 men and women across all ages and gender groups.
5.4.1. Sampling technique
The sampling technique used in the study was quota sampling, which is a method of stratified sampling in which the selection of research participants within strata (groups) is non-random. According to Schmidt and Brown (2014) the difference between quota sampling and stratified random sampling is that in quota sampling, research participants are conveniently selected from each stratum rather than randomly selected. The choice of quota sampling was motivated by the following reasons: Firstly, a probability random sampling technique to determine the sample for
this study was not feasible because of privacy regulations at Unisa. In this university, students’ records – names, addresses and telephone numbers – are restricted by the Protection of Personal Information Policy of the university. So, a sampling frame from which a random sample could be drawn was not accessible to the researcher. This restriction therefore invalidated the process of random sampling. Schmidt & Brown (2014) point out that quota sampling is used in “predominantly” quantitative studies where it is difficult to determine a sampling frame due to the absence of a list from which to draw a sample. Thus, the availability and unavailability of the sampling frame will determine the choice of the sampling technique – probability sampling or non-probability sampling.
Secondly, according to Creswell (2002), sample size is often a problem in a mega-university because the final size is dictated by issues such as the number of participants who volunteer to participate in the research and the number available to the researcher. Thirdly, the literature (Barbie 2012; Dodge 2003; Schmidt & Brown 2014) regards quota sampling as equivalent to stratified sampling, which is a probability sampling technique. According to Barbie (2012:192), quota sampling, like stratified sampling, addresses the issue of representativeness in research, although the two techniques approach the issue differently. The difference is that whereas stratified sampling uses random sampling to fill the groups (strata), quota sampling uses judgement/purposive sampling to assemble a representative sample. For example, in quota sampling, subjects who bear suitable characteristics that represent the population are handpicked on a volunteering basis to form a representative sample. According to Barbie (2012:192) these subjects are selected into a sample on the basis of pre-specified characteristics, so that the same sample will have the same distribution of characteristics assumed to exist in the population studied.
This study rests therefore on quota sampling’s premise that if the sample effectively represents the characteristics of the population being studied, the population will be “correctly” represented. The ability to generalise research results to the target population depends “heavily on the appropriateness of the sampling method used” (Schmidt & Brown. 2014). According to Barbour (2001) quota sampling is the most widely used technique in market research studies.
Parasuraman et al (1988) collected data for the refinement of their 97-item instrument from a quota sample of 200 male and female adults aged 25 years old and older.
In this study a combination of quota sampling and purposive sampling, both of which are non- probability sampling techniques, was used. It has to be emphasised that although quota sampling is a non-probability sampling, it is regarded by researchers as equivalent to stratified sampling, which is a probability sampling technique (Yang & Banamah, 2014). Kangai et al (2011) used stratified sampling for their research. In light of this, Unisa’s four regions and an international region were considered as strata: Gauteng, North West, Northern Cape, Orange Free State and an international regional centre (Lesotho) - (Figure 5.1). Students in all these regions have similar characteristics: men and women, across different age groups, who are registered for different courses and are therefore current users of student support services. Purposive sampling was used to sample students from different strata who volunteered to participate in the study.
5.4.2. Questionnaire administration
When the context is too diverse and segmented, as is the case with most DE institutions, it is always wise to use as many different methods of questionnaire administration as possible to ensure a good response rate. Initially, the researcher had planned to use as many methods of administering the questionnaire as possible; for example, e-mailing questionnaires to students; using a website (MyUnisa); distributing questionnaires to study centres; and going door-to-door. However, the researcher was confined to using two methods of administering the questionnaire due to policy restrictions.
Figure 5.1: Map showing Unisa regional centres
Although the researcher had applied for an ethics certificate and had obtained permission from the university’s Senate to conduct the research, students’ information such as e-mail addresses, postal addresses and telephone numbers could not be released to the researcher due to personal information protection policy of the institution. So the main method the researcher had to use was door-to-door. The emailing collection method was eventually used after obtaining email addresses through other student respondents.
The door-to-door method was a very slow, tedious and expensive process because in certain cases, the researcher had to phone to find out which students were willing to participate in the study so that the questionnaire could either be taken to them or e-mailed. Another method used that was also expensive was to look for students to distribute the questionnaire to other students who belonged to one study group. Professionals of goodwill also helped the researcher distribute
the questionnaire to students they knew – however, this did not work very well as evidenced by the low response. Although the e-mail method of questionnaire administration was cheap, it did not yield a good response either.
5.4.3. The response rate
Four hundred people were sampled and 600 questionnaires were administered to the research respondents face-to-face and on e-mail. These were self-completion questionnaires. Questionnaire administered to students face-to-face received a relatively good response rate. However, questionnaires delivered electronically, via e-mail, did not yield a satisfactory response rate. Of the 300 face-to-face questionnaires, 196 questionnaires were returned. The face-to-face response rate was 58%. On the other hand, of 100 questionnaires sent via e-mail, 17 were returned. The response rate was 17%. Despite numerous reminders and requests to respond to questionnaires, the overall number of questionnaire collected and correctly filled was 209 and the overall response rate was 50.75%. According to Babbie and Mouton (2001:261), a response rate of more than 70% is considered very good. Nevertheless, some researchers believe that the response rate between 45–50% is not uncommon. For example, Babbie (1990) argues that a response rate of 50% is adequate. The survey results of a study conducted by Liebenberg (2012) on the use of technology by Unisa students, reveal that of 282 248 online students, 22 216 managed to complete the survey, resulting in a small response rate. It has been noted that some successes in high response rates, in certain cases, are achieved if a cash prize is promised to respondents. This can happen if the research is financially well supported.