Increasing Response Rates in Mail Surveys Without Increasing Error

Abstract

Self-administered mail surveys remain one of the most common methods of collecting data in social science research. Not only are mail surveys cost-effective, they also enable researchers to gather both quantitative and qualitative data from elite policy makers as well as everyday citizens. However, no matter how well the research is designed or how interesting the questions are, it is all of little use if the population fails to respond. Accordingly, in an effort to increase response while decreasing the error associated with nonresponse, the following article argues for the careful integration of multiple contacts and endorsements when surveying elite policy makers. Such an approach is consistent with the Tailored Design Method and was used in a recent survey of chief prosecutors in four states. Response rates reached as high as 90% in one of the states, with an overall response rate of 76% across the four states.