Skip to main content

Sampling unit

(go to Outline)

This discussion of sampling methods generally assumes that the household is the basic sampling unit. Wherever you read "household" in the following discussion, please keep in mind that you could substitute "child" or " individual" if that were the sampling unit in your survey.

You can sample any sampling unit, including children less than 5 years of age, adult women, men, etc. Which sampling unit you use is largely determined by the target groups in your survey and the data you have about the population in which the sampling will be carried out. Regarding target groups, of course you would not use women of child-bearing age as a sampling unit if this is not one of your target groups. Regarding data on the population, some populations have lists of children less than 5 years of age or lists of members of other target groups. In such situations, if you only wish to measure indicators and outcomes in one target group, for example young children, you can use children as the sampling unit. In other populations, such as in some developed countries and in stable refugee camps, there is an up-to-date list of everyone in the population. In these populations, you could use each target group as sampling units and select separate samples for each target group. For example, you could choose separate samples of children less than 5 years of age, women of child-bearing age, adult men, adolescents, etc. if you wanted to collected data on these groups.

Nevertheless, in most populations, population data are most readily available on households, such as the number of households in each village or lists of households in villages. In addition, if you have more than one target group in your survey (for example, if you are collecting data on the household, young children, and women of child-bearing age), it may be easiest to have one sampling unit, the household, and then include everyone in each selected household who belongs to one of the target groups. For these reasons, household is probably the most common sampling unit in population surveys.