Uganda - Demographic and Health Survey 1995

A sample of 303 primary sampling units (PSU) consisting of enumeration areas (EAs) was selected from a sampling frame of the 1991 Population and Housing Census. For the purpose of the 1995 UDHS, the following domains were utilised: Uganda as a whole; urban and rural areas separately; each of the four regions: Central, Eastern, Northern, and Western; areas in the USAID-funded DISH project to permit calculation of contraceptive prevalence rates.

Districts in the DISH project area were grouped by proximity into the following five reporting domains:
- Kasese and Mbarara Districts
- Masaka and Rakai Districts
- Luwero and Masindi Districts
- Jinja and Kamuli Districts
- Kampala District

The sample for the 1995 UDHS was selected in two stages. In the first stage, 303 EAs were selected with probability proportional to size. Then, within each selected EA, a complete household listing and mapping exercise was conducted in December 1994 forming the basis for the second-stage sampling. For the listing exercise, 11 listers from the Statistics Department were trained. Institutional populations (army barracks, hospitals, police camps, etc.) were not listed.

From these household lists, households to be included in the UDHS were selected with probability inversely proportional to size based on the household listing results. All women age 15-49 years in these households were eligible to be interviewed in the UDHS. In one-third of these selected households, all men age 15-54 years were eligible for individual interview as well. The overall target sample was 6,000 women and 2,000 men. Because of insecurity, eight EAs could not be surveyed (six in Kitgum District, one in Apac District, and one in Moyo District). An additional two EAs (one in Arua and one in Moroto) could not be surveyed, but substitute EAs were selected in their place.

Since one objective of the survey was to produce estimates of specific demographic and health indicators for the areas included in the DISH project, the sample design allowed for oversampling of households in these districts relative to their actual proportion in the population. Thus, the 1995 UDHS sample is not self-weighting at the national level; weights are required to estimate national-level indicators. Due to the weighting factor and rounding of estimates, figures may not add to totals. In addition, the percent total may not add to 100.0 due to rounding.


Out of 8,093 households selected, 7,671were occupied, the shortfall being a result mostly of vacant houses. Of the existing households, 7,550 were interviewed, for a response rate of 98 percent. The main reason for non-response was the interviewer's failure to find a respondent at home after at least three visits.

In the interviewed households, 7,377 eligible women were identified and of these, 7,070 were interviewed, yielding a response rate of 96 percent. In the subsample of households selected for the man's interview, 2,224 eligible men were identified, of which 1,996 were successfully interviewed (90 percent response). The principal reason for non-response among both eligible men and visits to the household. The lower response rate among men than women was due to the more frequent and longer absences of men.

The response rates are lower in urban areas due to long absence of respondents. One-member households are more common in urban areas and are more difficult to interview as they keep their houses locked up most of the time. In urban settings, neighbours often do not know the whereabouts of such people.