OBJECTIVE: The aim was to study the most effective method for increasing response rates to postal questionnaires by comparing normal post with registered mail and to assess the cost implications of the two mailing methods. STUDY DESIGN AND SETTING: General Practitioners in Victoria (n=1,550) were randomized to receive a research questionnaire via either normal post or registered mail. RESULTS: The overall response rate for the study after two reminders was 76.1% (1,179/1,550). After the initial mail out the response rate for the registered mail arm was 55.9% (n=433) compared to the normal post arm, 40.1% (n=311, P-value <0.001). After the first and second reminders the response rates dropped, registered mail obtained a response rate of 47.4% (n=162) and 37.8% (n=68), respectively, compared to a response rate of 28.0% (n=130) and 22.5% (n=75) for normal post. Overall, the cumulative response rate for the registered mail was greater, 85.6% (n=663) compared to 66.6% (n=516) for the normal post arm (P-value <0.001) There was a total cost difference between the registered and normal method of AUD $1,531.50. CONCLUSION: Registered post when used for mail out of the initial questionnaire and all subsequent reminders is a more effective yet more expensive method for achieving a high response rate.