Rwanda - Demographic and Health Survey 2014-2015

The sample for RDHS 2014-2015 was a stratified sample selected in two stages from the 2012 Population and Housing Census frame.
Stratification was achieved by separating each district into urban and rural areas. The urban and rural areas of each district form each a sampling stratum. In total, 60 sampling strata have been created. Samples were selected independently in each sampling stratum, by a two-stage selection. Implicit stratification and proportional allocation were achieved at each of the lower administrative unit levels by sorting the sampling frame within the explicit stratum according to administrative unit in different levels before sample selection and by using a probability proportional to size selection at the first stage of sampling.

In the first stage, 492 EAs were selected with probability proportional to the EA size and with independent selection in each sampling stratum with the sample allocation. A household listing operation was carried out in all of the selected EAs before the main survey. The household listing operation consisted of visiting each of the 492 selected EAs; to draw a location map and a detailed sketch map; and to record on the household listing forms all residential households found in the EA with the address and the name of the head of the households. The resulting list of households served as the sampling frame for the selection of households in the second stage. Some of the selected EAs were found large in size in the household listing operation. In order to minimize the task of household listing, the selected EAs with an estimated number of households greater than 300 were segmented. Only one segment was selected for the survey with probability proportional to the segment size. The methodology and the detailed household listing procedure are addressed in the household listing manual.
At the second stage, a fix number of 26 households was selected from each selected EA. Among the 492 EAs, 113 EAs were from urban areas and 379 EAs came from rural areas. The total number of household selected is 12792, among them, 2938 household were from urban areas, and 9854 households from rural areas.However, during data collection, one of the households was
found to actually be two households, which increased the total sample to 12,793. The urban area was slightly oversampled because of the low urbanization in most of the districts, where at least two urban clusters needed to be selected. With the request of representative results for some indicators at district level, the total sample size is tight and therefore an equal size allocation was adopted, with a slightly larger sample size for the districts in the province of Kigali City because the low fertility level in Kigali City. In fact, the equal size allocation is not far from the proportional allocation which is the best allocation, because the district sizes are quite homogeneous.

Regarding the coverage, all 492 enumeration areas selected for the sample were surveyed for the 2014-15 RDHS. A total of 12,793 households were selected, of which 12,717 were occupied at the time of the survey. Among these households, 12,699 completed the Household Questionnaire, yielding a response rate of 99.9 percent (Table 1.1). There was little variation in response rates by urban-rural residence. In the 12,699 households surveyed, 13,564 women age 15-49 were identified as being eligible for the individual interview; interviews were completed with 13,497 of these women, yielding a response rate of 99.5 percent. Male interviews were conducted in every second household. A total of 6,249 men age 15-59 were identified in this subsample of households. Of these men, 6,217 completed individual interviews, yielding a response rate of 99.5 percent. Response rates among men were slightly higher in rural areas, while rates among women were almost the same in rural and urban areas.


The household response rate was 98 percent; the women's individual response rate was 99 percent; and the men's individual response rate was 99 percent.


Due to the approximately equal sample sizes in each district, the sample is not self-weighting at the national level, and weighting factors have been added to the data file so that the results will be proportional at the national level.

Due to the non-proportional allocation of the sample to the different provinces and to their districts and the possible differences in response rates, sampling weights are required for any analysis using the 2014-15 RDHS-V data to ensure the actual representative of the survey results at national level and as well as at domain level. Since the RDHS sample is a two-stage stratified cluster sample, sampling weights were calculated based on sampling probabilities separately for each sampling stage and for each cluster.