Background

The validity of the results of mailed surveys is often threatened by nonresponse bias, which is made more likely when response rates are low. However, the effectiveness and cost-effectiveness of several strategies to increase response rates are uncertain.

Objectives

To assess three strategies to increase response rates to mailed physician surveys: including a 10 dollars versus a 5 dollars cash incentive in the initial mailing, including a mint candy or not, and using a large versus small outgoing envelope.

Research design and subjects

Using a 2 x 2 x 2 factorial design, a randomized trial of these strategies was conducted in a survey of 1200 physicians randomly selected from the American Medical Association's Master File.

Results

Including a 10 dollars incentive yielded a significantly higher response rate (60.5% vs. 52.8%) (P = 0.009). The mailing and incentive costs per completed response were 12.24 dollars (95% CI, 11.75 dollars, 13.64 dollars) in the 5 dollars group and 18.48 dollars (95% CI, 17.77 dollars, 20.69 dollars) in the 10 dollars group. Each additional response obtained in the 10 dollars group came at an incremental cost of 61.26 dollars (95% CI, 36.98 dollars, 200.80 dollars). Neither inclusion of a mint nor use of a large envelope influenced the response rate.

Conclusions

Investigators may increase response rates by including more money in the initial questionnaire packet, but there may be diminishing returns to serial increments in incentives greater than 5 dollars. Including smaller incentives in more questionnaires may maximize total responses.