STUDY OBJECTIVE: To maximise the response rate in a community survey among ethnic minorities by combining postal questionnaires and interviews and to evaluate the validity of combining results from these different methods. DESIGN: A cross-sectional community survey of a local population using postal questionnaires with interview questionnaires for non-respondents. Postal questionnaires were in English and interview questionnaires were prepared in South Asian languages. A sub-sample completed both postal and interview questionnaires. SETTING: Two general practices in Tameside, Greater Manchester, UK. PARTICIPANTS: Questionnaires were mailed to 1,267 people. People were included if they defined their ethnicity as Indian, Pakistani, Bangladeshi or a combination of these. Fifty-five people who returned postal questionnaires were also interviewed. MAIN RESULTS: Overall response rate was 75%. Comparison of questionnaire and interview responses produced values of kappa ranging from marginally below zero to one. Equivalence was greater with a shorter time between postal completion and interview and where questions were more objective. CONCLUSIONS: It is possible to achieve a good response rate for an epidemiological study among ethnic minorities by using both postal questionnaires and interviews. Care should be taken when results from these two methods are combined, since equivalence is uncertain.
STUDY OBJECTIVE: To maximise the response rate in a community survey among ethnic minorities by combining postal questionnaires and interviews and to evaluate the validity of combining results from these different methods. DESIGN: A cross-sectional community survey of a local population using postal questionnaires with interview questionnaires for non-respondents. Postal questionnaires were in English and interview questionnaires were prepared in South Asian languages. A sub-sample completed both postal and interview questionnaires. SETTING: Two general practices in Tameside, Greater Manchester, UK. PARTICIPANTS: Questionnaires were mailed to 1,267 people. People were included if they defined their ethnicity as Indian, Pakistani, Bangladeshi or a combination of these. Fifty-five people who returned postal questionnaires were also interviewed. MAIN RESULTS: Overall response rate was 75%. Comparison of questionnaire and interview responses produced values of kappa ranging from marginally below zero to one. Equivalence was greater with a shorter time between postal completion and interview and where questions were more objective. CONCLUSIONS: It is possible to achieve a good response rate for an epidemiological study among ethnic minorities by using both postal questionnaires and interviews. Care should be taken when results from these two methods are combined, since equivalence is uncertain.