World Values Survey 2007, Wave 5

Some houses were not easily accessible In some residential areas, sampling was hard due to the way the houses have been arranged. E.g. in storey houses. Sampling in rural areas was a challenge due to the distances between houses, this lead to the number of questionnaires being done per day in rural areas becoming low. As mentioned above, WVS sampling was done at different levels and stages. Probability sampling methods namely random sampling, were used in this survey, to select the various units of the survey. All respondents in the universe (countrys population aged 16 years and above) had equal chances of being selected to form part of the sample. However institutions such as prisons, hospitals, military bases etc, were not included in the sampling process. Random sampling was also used to select the enumeration area (EAs), starting points and households /dwellings. The EA is the lowest geography division level that the country is divided into for census purposes. In each EA, there are specific numbers of dwellings houses which are known to the census bureau.

Selection of Sectors/ EAs. The selection of EAs was done through random probability sampling technique, in which secondary list of EAs was drawn (randomly) from a list of all EAs in Zambia. The randomly selected EAs were then sed as the primary sampling point for the study. The sample allocated to each EA took into account the density of the areas i.e. high, medium and low. A specific number of EAs were allocated to each province and these served as entry level for the survey. The number of EAs per region took into account the rural/urban split as well. An average of 6-8 interviews were conducted in each EA.

Selection of starting point. Selection of the starting point (point where the interviewers enter the EA and starts selecting the households) was done by the supervisors. The selection was based on easily identifiable features/land marks such as school, church, mosque, school, hospitals, roads, etc. Once the starting point had been selected, the selection of the first household within the EA was commenced.

Selection of Household: From the starting point within the EA, the interviewers walked to specific direction alternately to the residential to identify the first household for the interview. Households were selected using the fourth left hand rule, whereby the interviewers counted houses on the left upto the 4th house, then conducts interview in the 5th house. At the end of the street, the interviews turns left and continue with the counting until the 4th house, the same process is repeated after every successful interview.

Substitution of Households: Households were substituted if they were in any way inaccessible or had no member meeting the requirements of the survey e.g. in terms of age, citizenship and residency. In such a situation, the interviewer would move to the next house.

Selection of Respondents: Once a household had been selected, the interviewer had the task of selecting the person to be interviewed(respondent) from the selected house. The respondent was selected using a Kish grid (see a sample below). In selecting the respondent, the interviewer listed all members of the household aged 16 years and above and who met other survey criteria, on the Kish Grid. They then identified the respondent by a working a line on the last digit of the questionnaire serial number with the number of people in the household

Call Backs/Substitution Criteria: In situations where the selected respondent in the household was not available at the time of call, the interviewers were instructed to make up to two additional recalls on different times of the day including evenings when the selected respondents were said to be at home. However, where the selected adult was not available for interview within the day of selection, interviewers were asked to regard such a case as a non response situation or ineffective call. This was also the case when the selected respondents expressed unwillingness to participate. No substitution of respondent within the same household/dwelling structure was allowed.

Remarks about sampling:
A lot of refusals/unwillingness to participate in the survey was witnessed among those in the social class A and B especially in urban Lusaka and Copper belt regions. Although the questionnaire was translated into 5 major languages, the field team still encountered language barriers and since some respondents could to speak any of the languages.

The sample size for Zambia is N=1500 and includes national population aged 16 years and over, for both sexes.