Physician response rates to mail and personal interview surveys

This paper reports the outcome of a randomized experiment, sponsored by the Food and Drug Administration, which compares mail and personal administration of a 12-page questionnaire dealing with physician antibiotic prescription practices. To overcome the barrier created by the lack of clearly defined and accepted definitions of survey outcomes, the authors introduce six concepts for describing and comparing survey outcomes. Overall, mail surveys appear to require less time, to be less costly, and to generate higher response rates than equivalent personal contacts. The report describes methods for gaining maximum benefit from the mail technique.