The aims of the survey were to provide information on the social and economic conditions in Lubombo Region which are likely to affect the development of homestead aquaculture, and to provide MoAC personnel with experience in the design and carrying out of such surveys. It is envisaged that this experience can be applied to similar information gathering exercises in other regions of the country and that the data obtained can act as a baseline for assessing the impact of future development effort in the particular region surveyed.
The survey field work was carried out between mid-August and early October 1990 after work on design and pretesting was completed. Coding, compilation of data and preliminary analysis were carried out in October 1990, and the main findings from this are presented below. Further work and detailed analysis is still desirable but sufficient information has been obtained to establish the main constraints to development and to identify where further investigation is necessary. Valuable experience, both in the general management of surveys and in particular aspects of the survey, has been gained.
The sampling of homesteads was set on the basis of interviewing (a) all known pond owners in the region, (b) an equal number of neighbours of pond owners and (c) an equal number of “non-neighbours”. Neighbours were defined as homesteaders located within 1 km of a pond owner and “non-neighbours” as homesteaders located at least 10 km away from pond owners. At the outset of the survey, the number of pond owners was understood to be 43 and this number of interviews was scheduled for each of the three categories. However, during the survey field work, a total of 52 pond owners was identified and the sample sizes of the other categories were adjusted accordingly. Sampled neighbours and non-neighbours were chosen by random selection. The actual spread of the sample covers the northern, central and southern areas of the region.
156 interviews were conducted, but the effective sample size was subsequently reduced to account for incomplete questionnaires and ponds owned by individuals or groups other than homesteaders. Ponds attached to schools or clinics were eliminated from the analysis. The final selection of questionnaires included in the analysis was as follows:
| Category | No. | |
| - pond owners | : | 46 |
| - neighbours | : | 50 |
| - non-neighbours | : | 52 |
| - total | 148 | |
The survey was based on a questionnaire which was designed by the project group and administered through interviews conducted by trained interviewers. These were in turn supervised by the project group. Fisheries specialists did not participate in the interviews.
The questionnaire (see Appendix 1) included sections designed to elicit information about the following aspects of the homesteads' make-up and operation:
Coded responses were included in the questionnaire before the survey field work was started. Some questions were left open-ended and the responses coded after completion of the survey.
The questionnaire was finalised in Siswati after pretesting of an English version with 12 homesteaders in the Mbabane area. These interviews were conducted by the four interviewers and served both as a part of their training for the survey and as a pretest. Interviewers and supervisors travelled from Mbabane to the Lubombo region each day to carry out the interviews, which were conducted in the Siswati language.
Supervision of the interviewers took the form of direct observation of interviews during the early stages of the survey and regular monitoring of completed questionnaires throughout the survey. Several homesteads were revisited to obtain information not available on the first attempt.
The completed questionnaires were categorized into pond owners, neighbours and non-neighbours and the responses simultaneously translated into English and coded, using a database programme on a personal computer. The coded responses were then printed out, reviewed and analyzed by the project group.
