KEY COMMUNICATION AND DEVELOPMENT WEBSITES AND PROJECTS
Kenya Survey Methodology
The AudienceScapes quantitative data for Kenya are based on a national survey of 2000 adults aged 15 and older, interviewed face to face from 20 July to 11 August, 2009. The target population for the survey is 21 million adults aged 15 and older. A total of 5417 contacts were attempted to achieve the 2000 interviews. The response rate for the survey was 84 percent of eligible contacts. The margin of error for the sample is ±2.2 percent with a 95 percent confidence interval. Percentages for subgroups have larger margins of error, depending on the number of respondents in the group.
In order to achieve a nationally representative sample of Kenyan adults (age 15 and over), InterMedia chose a probability proportional to size sampling plan based on the Kenya 1999 Population and Housing census, conducted by the Kenya National Bureau of Statistics. In the first stage of the stratified sampling design, the country was divided into eight provinces; these provinces were then subdivided into districts, and further stratified into urban and rural settlements. Sampling points were randomly selected within each district, with no more than 10 interviews conducted in each sampling point. The target of 2000 interviews was achieved according to the established sample distribution:
Table 1: Kenya Survey Sampling Distribution
At each sampling point, households were selected by a random walk procedure, with interviewers contacting every fourth house on the left-hand side. Each sampling point was assigned a starting landmark, from which the first household to be contacted was determined by the summation method (the date or the sum of the two digits of the date). In the event of an unsuccessful interview, the interviewer knocked on the immediate next household until a successful interview was achieved. Respondents within households were selected randomly by a Kish grid method from a list of adults in the household.
A total of 31 interviewers and seven supervisors took part in the survey. InterMedia staff based in Washington, D.C. and the United Kingdom worked closely with the Kenyan research partner throughout the fieldwork, and were present in Kenya for training and supervision. The median interview time was 55 minutes. A minimum of two call-backs were attempted if the selected survey respondent was not at home at the first call.
A standard AudienceScapes national survey questionnaire was adapted to reflect the Kenyan environment, in collaboration with InterMedia’s local subcontractor. The questionnaire was then translated from English into Kiswahili, Luo and Kikuyu. The questionnaire covered the following topics:
Data Processing and Quality Control
Responses were recorded by hand in the field, and checked by field supervisors. Completed questionnaires were then coded into a database and subjected to rigorous quality control procedures. Coding supervisors ensured consistency by checking the first 10 to 20 questionnaires coded by each coder, with random checks thereafter of 10 percent of the sample. Additional quality control and consistency checks were performed by InterMedia’s Research and Analysis Unit prior to analysis.
The data were weighted to ensure they would represent the sampled population in terms of age, sex and region of the country.
Problems Encountered/Solutions Applied
The fieldwork contractor reported challenges reaching particularly remote rural sampling points (for example, in Turkana and some parts of Eastern province), but was able to reach all selected sampling points. Some older respondents were reportedly embarrassed by questions about birth control methods and had difficulty answering.
This research has been carried out in conformity with the standards of the ICC/ESOMAR International Code of Marketing and Social Research Practice.