98
99 the characteristics or parameters of the whole population (Neuman 2014). It should be representative of the population to ensure that the results can be generalised to the population as a whole. There are two types of sampling, namely probability and non-probability sampling. It is common for mixed research methods to use more than one kind of sample and also use samples of different sizes, scopes, and types within the same piece of research (Neuman 2014). As such, this study utilised both probability and non-probability sampling strategies.
The current study followed online sampling methods because at the time of data collection the country and world were going through the Covid-19 pandemic which limited face-to-face contact either through social distancing or through people having to work from home. Table 4.2 below shows the different online sampling methods for studies that are conducted online. The current study used list-based random sampling as a probability sampling technique. List-based sampling was chosen instead of intercept, pre-recruited panel survey or non-list based random sampling because it is straight forward to implement as only the contact information of the respondents is required. Unlike non-list random sampling, a sample can be selected without actually enumerating on the sampling frame. List based sampling is preferred in the current study because the sample frame can be determined, making it suitable for the scholarly set up.
Table 4.2: Online sampling typology
Probability sampling Non-probability sampling
Surveys using a list-based sampling frame Entertainment polls
Surveys using non-list-based random sampling Unrestricted self-selected surveys
Intercept (pop-up) surveys Surveys using “harvested” email lists (and data)
Mixed-mode surveys with internet-based option
Surveys using volunteer (opt-in) panels
Pre-recruited panel surveys River sampling Source: Fricker (2016)
Probability sampling, is a “method of selecting a sample wherein each element in the population has a known, non-zero chance of being included in the sample” (Neuman 2014:57). Probability sampling was used for the quantitative phase of the study, that is, the survey. It uses a random selection of units from the sampling frame to be included in the sample. The procedures in probability sampling are clearly defined (Hair, Wolfinbarger, Money, Samuel and Page 2015). One of the advantages of probability sampling is that sampling error can be
100 calculated. The use of probability sampling in online research has been met with scepticism (Vehovar, Toepoel and Steinmetz 2016). If the survey concerns the general population, significant biases can result from under coverage and non-response. Probability-based sampling methods/techniques begin with knowledge of a sampling frame. Although sampling for an internet-based survey can be difficult, most organisations have a fixed number of employees and maintain lists of their employees, thus making it feasible to draw a probability sample. In choosing the sample for the current study, Krejcie and Morgan’s (1970) table for determining sample sizes was used (Table 4.3 below). The table indicates that for a population of 139 the sample size is 97. According to Fricker (2016), a list-based random sampling frame can be conducted just as one would for a traditional survey using a sampling frame.
To implement list-based random sampling requires contact information on each unit in the sampling frame. The researcher requested the 139 names of the target population. Each name was assigned a random number between 1-139 as an identifiers. Then researcher used an online random number generator available at https://www.calculator.net/random-number-generator.html to generate numbers. This was carried out to eliminate the researcher’s bias. Every number that popped out, its corresponding name was set aside. After, the 97 names that comprise the research sample were reached, the researcher requested the emails of the selected employees from the human resource office of the studied parastatals. The survey link was sent via email to the randomly selected participants.
Table 4.3: Krejcie and Morgan’s (1970) table for determining sample sizes
N-n N-n N-n N-n N-n
10-10 100-80 280-162 800-260 2800-338 15-14 110-86 290-165 850-265 3000-341 20-19 120-92 300-169 900-269 3500-346 25-24 130-97 320-175 950-274 4000-351 30-28 140-103 340-181 1000-278 4500-354 35-32 150-108 360-186 1100-285 5000-357 40-36 160-113 380-191 1200-291 6000-361 45-40 170-118 400-196 1300-297 7000-364 50-44 180-123 420-201 1400-302 8000-367 55-48 190-127 440-205 1500-306 9000-368 Source: Krejcie and Morgan (1970)
101 According to Bless, Higoson-Smith and Kagee (2006), non-probability sampling is defined as a sampling technique in which the researcher selects samples based on his or her subjective judgment rather than random selection. Thus, in non-probability sampling, samples are selected in a non-random manner, unlike probability sampling but the degree to which the sample differs from the population is unknown. Also, unlike with probability sampling, non-response error does not arise in non-probability sampling. Non-probability sampling is sometimes preferred because the procedures used to select units for the sample are much easier, quicker, and cheaper than probability sampling. However, Neuman (2014) noted some disadvantages of non-probability sampling. One of its disadvantages is that an unknown proportion of the entire population may not be included in the sample group or there may be a lack of representation of the entire population. Non-probability findings thus have lower levels of generalisation of research findings compared to probability sampling. Babbie (2013) notes that it may be difficult to estimate sampling variability and identify possible bias.
The type of non-probability sampling adopted for the qualitative phase of the study (the interviews) was purposive sampling. Bless, Higson-Smith and Kagee (2006) assert that the method is based on the judgment of the researcher regarding the characteristics of a representative sample. The sample is chosen based on what the researcher considers are typical units that best fit the study’s criteria. The decision to include individuals who are to be part of the sample is carried out by the researcher. The decision is based on their particular specialist knowledge of the research issue as well as the individuals’ capacity and willingness to participate in the research. Purposive sampling picks a small number of cases that yield the most information (Bless, Higson-Smith and Kagee 2006).
The sample that is investigated is usually smaller when compared to those selected via probability sampling.
Thus, in terms of the study, respondents from each parastatal were selected based on their knowledge of electronic records. The study targeted six purposively selected respondents, two from each organisation. The first of the two was the records managers in each organisation. They were the custodians of the records and responsible for their management. The other two respondents were appointed by the CEOs to represent them. Five of the six selected participants were interviewed. One CEO representative kept promising to grant the interview but failed to honour appointments.