Tanzania (2007): HIV/AIDS TRaC Study Evaluating Condom Use with Non-Marital Sexual Partners among Youth. First Round

This serves as the baseline survey and it tracks several indicators related to the health of 15-24 year-olds with respect to HIV. This survey also includes segmentation analysis, which identifies the drivers / inhibitors of behavior by dividing the at-risk population into those that perform the desired behaviors, and those who do not, in order to focus on those factors that make these two groups different. The sample size for the survey was 2,056 males & females aged 15-24. A nationally representative sample was attained distributed across regions through a multistage cluster random sampling. A total of 65 wards (clusters) were randomly selected using PPS. In each selected ward, the data collectors would visit the ward center and used the ward village roster to randomly select villages to conduct interviews. In the selected villages, the data collector would spin a bottle to select the first household to be visited. If a household member q ualified for an interview, and is present, the interview was conducted. If the qualifying member of household is not present at the time, the data collector will note the location of the household and make three call backs. A systematic skip pattern of three households was used to select the next household. The data collector will then proceed to the next closest household, provided it has not been visited by another data collector. If a visited household does not qualify for either interview, the data collector will mark its door with an ‘X’. If an interview is completed, the door will be marked with an ‘√’. If the respondent is missing and the house is to be revisited, it will be marked with an ‘O’. Data was collected through face-to-face interviews, using PDAs to eliminate enumerator errors, lengthy data-entry and data cleaning, and ultimately ensuring a high-quality dataset. Data collectors received two weeks of training on research ethics, methodology, sample selection and the use of PDAs. The survey instrument was a Kiswahili questionnaire, which was pre-tested with 200 respondents. The data was analyzed in SPSS 11.5.