Reviewing the Research on Mail Survey Response Rates: Meta-Analysis.

Meta-analysis, a way of cumulating the results of research studies, focuses on the magnitudes of effect sizes and seeks to explain effects through study characteristics. This meta-analysis used the methods developed by G. V. Glass to summarize the research on mail survey response rate manipulations. A narrative review using the same studies set the stage for this review and a related review using descriptive methods. Results will demonstrate the similarities and differences one may see in analyzing the same data in different ways, as well as indicating the effectiveness of treatments meant to enhance mail survey response rates. Studies were coded and repeated measure main effects were summarized, as were absolute mean differences between the two levels of the variables and the 95% absolute confidence intervals for the absolute differences. Results were clearcut in that treatments were either strong or far from significant. Primary findings suggest that class of postage, incentives (especially when enclosed), followup, precontact, personalization of the cover letter, personalization of postage, and sponsorship to be effective methods of inducing responses to mail surveys. Questionnaire length also has an effect on response rate. Found to be ineffective were use of anonymity, color, cover letter appeals, deadlines, personalization of labels, and the status of the sender. Appendixes give category codings and correlations among dependent variables. (Contains 10 tables and 19 references.) (SLD)