Chapter 5 Research Methodology
5.5 Sample selection
5.5.1 How the population was determined
Selection of the population is critical to the success of any research.
Selecting a suitable population in the South African context was done as follows:-
The CIPRO database of Closed Corporations and Companies includes approximately 1,200,000 legal entities, and is by far the single largest database of businesses outside of the South African Revenue Services (SARS) database. However, this universum does not allow economically inactive entities to be excluded. Nor does it
allow for the exclusion of professionals who have partnerships and legal entities, nor the differentiation between SMEs and large business as determined by the National Small Business Act. The attractiveness and suitability of the CIPRO database is further weakened by the fact that the database is only tracking activity within the legal entities, and does not track activity within sole proprietorships and partnerships. This makes the CIPRO database unsuitable for research purposes due to the large accuracy error that is already resident within this universum.
The SARS database has a larger universum to analyse, being all tax paying businesses in South Africa, but all efforts to date to gain access to statistics from their database have been refused. As access to this population is not available, it is difficult to assess whether the data could be sorted in an appropriate manner to suit this study.
The next option was to look at the various bodies representing organised business. In South Africa, there are four such bodies. The first two, FABCOS and NAFCOC, traditionally existed to service black-owned businesses. However, they do not have local offices, and therefore communicating with their membership is very difficult.
Similarly, they tend to have a greater proportion of informal sector businesses than formal sector businesses. This makes this universum unsuitable for the purposes of this study.
The next organisation is the Afrikaanse Handels Instituut (AHI), the body that traditionally served the needs of Afrikaans, mainly white, businesses. This body’s members however are often members of both the Chambers of Commerce and the AHI, and form a very small universum.
The last body to consider are the constituent members of the South African Chamber of Business (SACOB), known as the South African Chamber of Commerce & Industries (SACCI) since 2007. This body has the largest universum within organised business. The organisation has been in existence since the early 1900s, and they generally have local offices in many towns and cities. The membership tends to comprise both English and Afrikaans speaking business people, and has grown its base of black business people.
Albeit that the demographics of the body do not reflect the national population demographics, the possibility exists that they do more closely reflect the demographics of the formal business population, albeit that this cannot be measured. This universum was therefore deemed best suited to this research study. However, this universum has limitations, in that the greater percentage of members, are white- owned businesses. A further limitation is that not all businesses are members of a Chamber of Commerce or similar body.
Chambers of Commerce use the size of the business in respect of the number of employees, to assess the subscription fees, thereby allowing the selection of SME members only for the study.
Chambers of Commerce have historically tended to attract businesses in the formal sector and this trend continues. It attracts businesses in the formal sector who are recent start-ups as well as existing SMEs in various stages of the enterprise lifecycle. The current SME component of Chambers of Commerce varies around 90% of the membership, depending on the individual Chamber of Commerce’s profile. Their membership structure makes it suitable for this research study.
This study will however not consider the complete country, but rather the province of KwaZulu-Natal. With a universum of 4012 SME members specifically within Ladysmith, Pietermaritzburg, Durban and Richards Bay, our study area, it is a large enough universum to provide useful information that can be statistically analysed.
Therefore, this universum is intended to be used for this study.
This universum will allow the research of start-ups as well as existing enterprises.
The chosen Chamber of Commerce databases have their own limitations, as representivity, based on random selection of the national universum of businesses in South Africa, cannot be claimed.
Cognisance will be taken of these limitations when finalising the assessment of the primary data collected.
5.5.2 The method for selecting the sample
An initial pilot sample of five (5) SME respondents was selected from the sample frame to test the validity of the questionnaire and whether or not the questionnaire functioned as expected. These results were not included in the research control group, nor the treatment group.
It was decided to use the Pietermaritzburg Chamber to firstly select a control group. The reasons for selecting a control group, was that the author was twice previously a President of the Pietermaritzburg Chamber of Commerce, and therefore it was expected that a higher level of honesty in the responses from respondents would result. This would then allow for comparison to the broader KZN sample.
The full SME universum of SMEs who are members of the Ladysmith, Pietermaritzburg, Durban and Zululand Chambers of
Commerce, excluding those used in the control group, will be utilised.
The questionnaire will be emailed to all parties, and therefore there is no cost attached to the mailing.
5.5.3 The data collection procedures
The questionnaire will be emailed to the balance of the Pietermaritzburg Chamber of Business SME population and the SME members of the Ladysmith, Durban and Zululand Chambers of Commerce in KwaZulu-Natal. Six (6) weeks after the mail survey has been sent out, a second copy of the survey will be sent out again.
The population is 3398 SME members, adjusted for the 246 control group members.
The responses from the two groups will be assessed independently in order to ensure that there are no variances in the responses from the early responders and the later responders.
Six (6) weeks after the second mailing date non-respondents will be contacted telephonically and personally requested to respond and, if necessary, be interviewed by the author.
The reason for using an e-mail survey is to take advantage of the ease of access to contact vast numbers of potential respondents in a short space of time. It is also the preferred method for collecting data from a large universum which is also geographically dispersed (Wegner, 2001:16).