CHAPTER 3: CONTEXT AND METHODOLOGY
3.4 Methodology
3.4.3 Household Survey
During the second phase of the study a household survey was conducted. The aim was to conduct interviews in 200 households in each site. A total of 200 men and 200 women were to be selected for the study. It was decided that 400 households would be able to provide sufficient information for this stage of the study. While the number of households was rigorously reached as planned, the number of subjects was exceeded for nine subjects in the male sample and seventeen subjects in the female sample. In total, 209 men and 217 women were interviewed. The primary purpose of the survey was to collect detailed information on the sexual behaviour of men and women including number of sexual partners, type of sexual partners (such as regular, non-regular, commercial), frequency of specific behaviours (such as vaginal, oral and anal sex), as well as the conditions under which sexual activity takes place (such as having intercourse under duress or due to economic need). The survey also collected information on awareness and perception of risk of HIV infection and strategies to protect themselves against the risk of HIV infection. In addition, the survey obtained detailed socio-economic and demographic information that may impact on risky sexual behaviour and ability to adopt strategies to protect themselves.
Sampling Design: A multistage probability sampling method was applied. For the study only residents in each household were randomly selected for the interviews. A previously recorded enumerator area list in each study site was used (National Institute of Statistics, 1997). Within each urban and rural site a number of sub-districts (geographic enumerator areas) were chosen and then eight smaller areas selected within these sub-districts. In total, 16 enumerator areas (8 in the urban and 8 in the rural areas) were chosen for survey data collection. In each enumerator area 25 households were randomly selected from a list of 80 to 100 households which had been previously made with the assistance of local community leaders (National Institute of Statistics, 1997). The household questionnaire was used to identify all eligible men and women in the household. In order to be eligible for the interview, the household member had to be a resident of the household and had to be aged 20 to 49 years. Those men and women selected for the interview were asked a series of introductory questions to determine if they were sexually active.
The study only included men and women aged 20 and above as it was felt that they were more likely to have ever had sexual intercourse. In addition, in Mozambique a person is eligible to marry without parents consent by the age of 18. Unmarried men and women were only asked to complete the interview if they ever had sexual intercourse. Those men and women who fell in the acceptable age range and reported themselves to be married were considered eligible for the interview. In each household, one index adult respondent was randomly selected for the interview using the Kish grid (Németh, 2003). The Kish grid gives a procedure of selecting the respondent which will answer the questionnaire among eligible members in the selected household. The Kish grid ensures that all eligible members in a given household have an equal chance of being selected to complete the questionnaire (Németh, 2003). When using the Kish grid, the interviewer uses a simple procedure for ordering members of the household on a cover sheet assigned to each sample household. The interviewer lists each adult on one of the lines on the form. Each member of the household is identified in the first column by their relationship to the head of the household. In the next two columns, the interviewer records the sex, age, education and marital status of each adult. Then the interviewer assigns a serial number to each adult. Males are first numbered in order of decreasing age, followed by females in the same order. Then the interviewer consults the selection table. This table provides the identification number of the adult to be interviewed (Németh, 2003).
Field Procedures: The questionnaire for men and women were relatively similar, with some minor differences. The questionnaires were pre-tested to identify any unanticipated problems with specific items. The pre-testing aimed to assist in refining of the questionnaire. The pre- testing phase took place in Marracuene district, 15 kilometres North of Maputo city. After this was done, the research team had to improve the questionnaire in the following week. During the survey, respondents in the household were briefed about the purpose of the study and then were asked if they were willing to participate in the study. At that stage, respondents were presented with an informed consent form which indicated their willingness to participate in the study. Some of the respondents were not comfortable in signing the consent form, but they gave their verbal consent to participate in the study. The team was the same for both rural and urban sites.
Respondents were interviewed by an interviewer of the same sex. The principal investigator of the study assumed overall responsibility for the supervision of field work by ensuring that households were correctly identified and only eligible members of the household were interviewed. At the end of the second phase of the study, both men and women were asked if they would be willing to be interviewed at a later stage.
Table 3.1 presents the general sample characteristics of study participants. A total of 426 respondents participated in this study, 209 men and 217 women. Among men, 49 percent were interviewed in the rural area and 51 percent in the urban area. Among women, 49 percent were interviewed in the rural area while 51 percent were interviewed in the urban area.
Table 3.1: Percentage distribution of respondents by selected socio-demographic characteristics
Sample characteristics
Men Women
% % Place of residence
Urban 50.7 51.2 Rural 49.3 48.8
Age
20-29 51.7 56.3 30-39 26.3 27.9 40-49 22.0 15.8 Level of Education
None 13.4 23.5 Primary 47.4 48.8 Secondary or higher 39.2 27.6
Marital Status
Married/cohabiting 66.0 55.3 No 34.0 44.7
API
Low 21.1 30.9
Lower medium 28.2 23.0 Upper medium/high 50.7 46.1
Religion
None 16.7 4.6 Roman Catholic 19.1 27.2 Protestant/Evangelical 41.1 45.6 Muslim 1.0 1.4 Siao/Zion 18.2 20.3 Other 3.9 0.9
Ethnic group/Mother tongue
Chope 16.3 13.4 Ndau 1.9 1.4
Ronga 28.7 27.6 Changana 39.2 46.1 Bitonga 5.3 3.3 Macua 2.4 1.4 Xitsua 2.9 2.8 Other 3.3 4.0
Employment status
Employed 38.8 12.9 Not employed 61.2 87.1
Total 209 217
The majority of respondents of both sexes were aged between 20 and 29 years, comprising 52 percent of the male sample and the 56 percent of female sample, respectively. Among men, 13 percent of men never attended a school. Forty-seven percent had primary education and 39 percent had secondary education or more. Level of education was higher among men than women. Among women, about 24 percent never attended a school, 49 percent reported having primary education and only about 28 percent had secondary education or more. The majority of men and women reported that they were married or cohabiting. Sixty-six percent of men and 55 percent of women were married. The majority of respondents were Christians. Among those, 41 percent of men and 46 percent of women were Protestants/Evangelical. Nineteen percent of men and 27 percent of women were Catholic. Thirty-nine percent of men were currently employed at the time of the survey compared with only 13 percent of women. The socio-economic characteristics of participants were measured by the Amenities Possession Index (API). Table 3.1 shows that the 56 percent of men and 46 percent of women had a medium-high socio-economic status. The majority of respondents were originally Changana or Ronga speakers.