The Sample Strategy for an UNICEF Survey in Croatia

The paper describes the sample strategy (sample design and estimation method) applied in the large scale Multiindicator Survey in Health and Nutrition Sectors which was carried out in Croatia in 1996. The sampling methodology was proposed and agreed by representatives of the UNICEF (Monitoring Progress, UNICEF, 1995 ; Household Survey Manual, WHO, 1994), PULS - Croatian research agency (Dumicic et al., 1996) and the project leader from Clinic for Infant Diseases - Zagreb. The target population was defined as with children under age The respondents had to be mothers (or caretakers). A stratified random sample of clusters was proposed. A is defined as a randomly selected population group of a size likely to include a specific number of children under age 5. In this survey, cluster sampling is a two-stage process involving: first, selecting a kind of area units, here called segments, in which clusters should be located and second, identifying groups of households where interviews should be conducted. A household was defined as a group of persons who live and eat together. Since the post-war situation effects different areas of the country to varying degrees, stratification was based on areas of Croatia, to express these differences both before and during the survey period.