Survey on Insecticide-Treated Nets 2004

Plan abd Procedure
The sample was composed of 1000 Ethiopian households. Respondents were women of reproductive age (15-49) who were mothers or guardians of children under five years of age.

In the interest of comparability, the same procedure was used in all countries surveyed. A multistage sampling procedure was used to select respondents, as follows.

1- Selection of primary sampling units: Purposive sampling was used to select the five primary sites: Bahir Dar, Nazret, Dire Dawa, Dessie and Awassa. Only sites in malarious areas were selected. In all other countries surveyed, the capital city was included; however, since malaria is not a significant problem in Addis Ababa, the capital was not included for the Ethiopia survey. Further criteria included geo-ethnic diversity and the potential for NetMark to be active in product distribution and/or programs to provide targeted subsidies for vulnerable groups. In each site, the target sample was 200: 80 respondents from the urban center, and 120 households from up to 200 kilometers from the urban center. Therefore, the sample has an urban-rural ratio of 40:60. The sample is proportionately more urban than that of the country as a whole. The 1994 census was 14% urban and the Demographic and Health Survey of 2000 was 26% urban.

2- Selection of sampling points: Within each of the five sites, 20 sampling points (villages or urban neighborhoods) were randomly selected from electoral lists using quota sampling: 8 from within the city (urban) and 12 from within 200 kilometer radius from the city (rural). Ten households per sampling point were selected for inclusion. This stratification scheme was designed to meet the purposes of the evaluation. Since a key objective of NetMark is to increase ownership of ITNs across the socio-economic spectrum, it was essential to include urban centers with the potential to be reached by product distribution systems, as well as include households located at varying distances from the urban center where lower socio-economic status (SES) individuals typically reside.

3- Selection of households: Ten interviews were conducted per sampling point, each in a different household. For each sampling point, a starting point (a fixed landmark or address) and the direction from which to start the data collection were chosen. Interviewers were instructed to go to the starting point and walk in the chosen direction until they located a residence with a qualified respondent. After a successful interview, interviewers were instructed to skip five residences (or less if residences were far apart) and seek another qualified respondent.

4- Selection of eligible respondents: An eligible respondent for the evaluation was a female 15-49 years old who was the parent or guardian of a child less than five years old, i.e., aged 0-4. Females aged 15-49 were selected to maximize the sample size for calculating the proportion of females of reproductive age sleeping under a net. Similarly, only those women who had a child under five were included, to maximize the sample size for calculating the proportion of children under five sleeping under a net.