A major disadvantage of the mail survey in medical research Is non-response, which may introduce bias. This study measured the effect of two simple factors on the response rate to a questionnaire concerning headache and migraine in a survey of 2,508 electors. The two factors were the class of postage and the colour of envelopes used; neither was found to have a significant effect. It was concluded that cheaper mailing methods are not likely to jeopardise response.