Representativeness of Three Survey Methods in Ethnic Health Research

Abstract

Study objective: To identify the survey method which generates the most representative sample of respondents in an ethnic community Design: Three survey methods (including sampling from the telephone book and telephone surveys, cluster sampling of households from areas with known high concentrations of migrants from the target community and face-to-face interviews, and sampling of persons listed on the electoral roll and a mailed survey) were employed concurrently in a defined area. Setting: Metropolitan Sydney, Australia. Participants: Persons aged 18 years or more born in Lebanon. Main results: The telephone survey method consistently outperformed cluster sampling and sampling from the electoral roll in terms of cost, response and contact rates, ease and speed of administration and representativeness of the sample. All methods generated a similar age profile but over-represented females compared with census data for the study area. There was relatively little duplication of respondents. The proportion of respondents who rated their health as fair or poor was comparable. Conclusions: Telephone surveying of migrant populations is a recommended strategy for health research in the Lebanese migrant population in metropolitan Sydney, and is also likely to be recommended for migrant populations with ethnically identifiable surnames.