E-mail or snail mail? Randomized controlled trial on which works better for surveys.

OBJECTIVE To compare e-mail with regular mail for conducting surveys of physicians.

DESIGN Randomized controlled trial.

SETTING Ontario, Canada.

PARTICIPANTS A random sample of physicians listed in the College of Family Physicians of Canada's membership database.

INTERVENTIONS Survey delivered by e-mail and by post.

MAIN OUTCOME MEASURES Response rates and times, and completeness and characteristics of responses to the survey.

RESULTS Overall response rate was 44.7% (33.6% of e-mail recipients, 52.7% of post recipients who have e-mail, and 47.8% of post recipients without e-mail). While the e-mail rate was significantly lower than for both post groups, e-mail responses were received much faster. There was no significant difference among groups as to completeness of responses, but e-mail responses had more frequent and longer comments.

CONCLUSION E-mail provides faster but fewer responses to surveys. Content of structured-response questions was similar in all groups, but e-mail provided more and longer responses to open-ended questions. Where a quick response to a survey is required, e-mail is superior.