Response rates to mail surveys published in medical journals.

摘要: Objective. The purpose of this study was to characterize response rates for mail surveys published in medical journals; determine how the rate among subjects who are typical targets varies; and evaluate contribution several techniques used by investigators enhance rates. Methods. One hundred seventy-eight manuscripts 1991, representing 321 distinct surveys, were abstracted survey techniques. In a follow-up survey, 113 authors these provided supplementary information. Results. mean journals is approximately 60%. However, vary according subject studied tech- niques used. Published physicians have only 54%, those non-physicians 68%. addition, multivariable models suggest that written reminders with copy instrument telephone each associated about 13% higher than do not use Other techniques, such as anonymity financial incentives, Conclusions. Although rates, tend be moderate. survey's at best an indirect indication extent non-respondent bias. Investigators, journal editors, readers should devote more attention assessments bias, less specific thresholds. j CLIN EPIDEMIOL 50;10:1129-1136, 1997. 0 1997 Elsevier Science Inc.