CHAPTER 2: RESEARCH METHODOLOGY
2.5 Impact evaluation indicators of the twinning agreement
2.6.1 Sampling procedure
Sampling is the act, process, or technique of selecting a suitable sample, or a representative part of a population for determining parameters or characteristics of the whole population (Frey et al., 2000). The sampling procedure of this study will be discussed as follows:
(i) Definition of target population
For the purpose of this research, we defined two levels of the characteristics of the respondents. The first level of respondents were municipal officials involved in the twinning agreement. The study revolves largely on the Information gathered from the groups of officials based on their roles in the twinning agreement process. The second level of the target respondents was the residents of Musina and Beitbridge located in South Africa and Zimbabwe respectively. The idea was to collect data mainly from the implementers of the twinning, and to collect information on the opinions of the proposed beneficiaries of the twinning agreement. The study was both a qualitative and quantitative of study, where 14 key informants where interviewed and 385 questionnaires were administered to local residents.
The total number of questionnaires distributed was determined using the sample size calculator as explained in detail under section (iii) entitled the criteria for sampling.
(ii) Sampling frame
There were two groups of respondents; key informants (municipal officials) and the general respondents, the local residents. The exact number of key informants who participated in the twinning agreement implementation could not be obtained. The initial respondents were identified through literature review whereby the initial participants’ names were extracted from the twinning document. Some of the people which participated in the signing of the twinning document were no longer traceable while some had passed on. To identify further participants which were not enlisted on the twinning document, snow-ball sampling technique was used.
This technique was used because the initial participants identified further stakeholders whom they knew were involved although they were not enlisted in the twinning document.
Consequently, 14 key informants were interviewed. The sampling frame for the local residents in Musina was based on the 2011 population census which indicated a population of about
19 Outcomes of trans-border spatial development cooperation: Insights from Musina and Beitbridge’s Twinning Agreement.
68 358 and 20 042 households, of which 15 000 consisted of foreign migrants. (Musina IDP, 2014). Beitbridge town had a population of 42 137 and 10 545 households in 2013, (Beitbridge TC Annual Report, 2013). The total number of households for the two towns was 30 587, which was the sampling frame used in the study.
(iii) Criteria for sampling total respondents for questionnaire administration
The sample size calculator stipulates that the minimum sample size for a population which is more than 10 000 but less than or equivalent to 100 000 is 383, at a 95 % level of certainty (Thanulingom, 2000). Hence, 385 questionnaires were distributed, and 347 were returned.
The study was a qualitative and quantitative, one focusing on interviewing the implementers of the twinning agreement and the beneficiaries of the twinning implementation. Therefore, 14 key informants were interviewed and 385 questionnaires were administered to respondents.
(iv) Sampling technique
In order to draw out the desired sample of target respondents, the sampling technique was based on two groups of respondents.
Sampling of key informants
Initially the key informants were extracted from the twinning agreement document to confirm stakeholders who participated in the Musina-Beitbridge twinning agreement. Later on, the snowball sampling technique was employed, making use of the initial key informants to identify further key informants who are or were the key actors in the twinning agreement. The 14 key informants from Musina and Beitbridge included the mayors, Chief Executive Officers, technical directors, immigration officers, as well as the directors of spatial planning in both towns.
Sampling of the respondents (local residents)
A stratified random sampling method was used for the administration of questionnaires with respect to the respondents, who are the local residents of Musina and Beitbridge. The researcher used ward proportional representation to select the respondents. The ward demarcations were used as the strata in the two border towns as illustrated in table 2.2:
20 Outcomes of trans-border spatial development cooperation: Insights from Musina and Beitbridge’s Twinning Agreement.
Table 2.2 Stratified sampling using municipal ward demarcations as strata.
BEITBRIDGE
Wards Urban/Rural Population Total number of
households
NQD %
distributed
NQR % Returned
1 Urban/ Low density 5 130 1 283 16 12 13 10
2 Urban/Low density 2 276 569 7 5 4 3
3 Urban/Medium
density
4 175 1 043 13 10 10 8
4 Urban/High density 10 704 2 676 33 25 30 23
5 Urban/High density 7 555 1 888 23 18 19 15
6 Urban/High density 12 347 3 086 39 30 37 28
Total for Beitbridge 6 42 137 10 545 131 100 113 86
MUSINA
Wards Urban/Rural Population NQD %
distributed
NQR % Returned
1 Rural 13 365 4 009 51 20 48 19
2 Rural 16 747 4 811 61 24 54 21
3 Urban 12 760 3 808 48 19 42 17
4 Urban MUSINA 1 402 18 7 16 7
5 Urban Total number of
households
3 006 38 15 37 15
6 Urban 9 928 3 006 38 15 37 15
Total for Musina 6 68 358 20 042 254 100 234 92
GRAND TOTAL 12 110 495 30 587 385 100 347 89
Source: Author’s Field Data 2016, Zim stats (2012) and SA census (2011)
Table 2.2 shows that 385 questionnaires were distributed proportionally based on population proportions i.e., 131 questionnaires in Beitbridge and 254 questionnaires in Musina. Ward.
Proportional allocation was used to select number of samples in each ward. The sample from each selected ward was proportional to the household population size of that particular ward.
Table 2.2 above shows the total wards for the study areas, the total number of households in each ward and the total number of questionnaires administered per ward. The proportional sample size was derived using the following formula:
Sample size of ward = Total households per ward × Total number of questionnaires (385) Total households of all wards
For example, Ward 2 of Musina has a total of 4 811 households.
Therefore Sample size of ward 2 in Musina = 4 811 ×385 = 61 questionnaires 30 587
21 Outcomes of trans-border spatial development cooperation: Insights from Musina and Beitbridge’s Twinning Agreement.
Since the twinning agreement affects residents in all the wards of Musina and Beitbridge, stratified random sampling was used to select respondents in all wards where each member and household head had an equal chance of being selected. In Beitbridge, the selection of the household heads within the wards was based on the intervals of the street arrangement i.e.
for every two streets; the first household head was picked until the total number of respondents required was selected. In Musina urban area, the same procedure was used as in Beitbridge.
However, in rural areas, where there are no streets, the researcher used census enumeration blocks to select the households.