Editor,—The poor response rate of GPs to postal surveys has long been a barrier to successful research. We were pleased that Stocks and Gunnell1 systematically studied this longstanding problem.

There are a number of factors contributing to a poor response rate. Stocks and Gunnell1 studied the influence of the characteristics of the GPs and confirmed that there are such differences between responders and non-responders. However, other factors may also influence response rates, for instance the method of distribution of the questionnaires.

We would like to share the results of a study carried out to investigate the effect of different methods of distribution on response rates. We performed a questionnaire study related to psychiatry, investigating GPs in two similar geographical areas, Solihull and Warwickshire, both in the West Midlands.

The questionnaires were posted to GPs in Warwickshire, with prepaid envelopes provided for reply. The GPs in Solihull received their questionnaires delivered by hand by one of the researchers, and replies were collected from the surgery by the same person two weeks later.

The postal response rate was 50%, while the response rate from the hand delivered area was 60.3%. These results were significant (p=0.05, χ² = 5.860, Yates's correction = 5.330).

These results show that the method of delivering and collecting questionnaires by hand significantly improves response rates. It also has a number of other advantages, such as an improvement in the researcher's knowledge of the area under study, their relationship with GPs and reducing costs of the study. We would encourage others to consider this method of questionnaire distribution in the future.