Interviews or postal questionnaires? Comparisons of data about women's experiences with maternity services

Abstract

Surveys by personal interview are often assumed to be superior to those conducted by mail questionnaire. An experimental study of experiences and attitudes of 800 newly delivered mothers revealed surprising advantages to postal surveys: they are cheaper, more easily repeatable, and minimize interviewer effects. While response rates differed, the quality of responses was similar, except between middle- and working-class mothers. Postal surveys can be used with considerable assurance in national studies of fairly intimate experiences of pregnancy and delivery.

PIP: Surveys by personal interview are often assumed to be superior to those conducted by mail questionnaire. An experimental study of experiences and attitudes of 800 newly delivered mothers revealed surprising advantages to postal surveys: they are cheaper, more easily repeatable, and minimize interviewer effects. While response rates differed, (92% for the interviews, 75% for the postal study) the quality of responses was similar, except between middle-and working-class mothers. Postal surveys can be used with considerable assurance in national studies of fairly intimate experiences of pregnancy and delivery. These conclusions are based on a comparison of interview responses and postal survey responses in 4 different geographic areas of England where experienced interviewers from the Institute for Social Studies in Medical Care were already in place. Data was collected in 2 phases--October to December 1983, and February to April 1984, based on a random sample of 100 births registered for each area 2 months prior to the collection phases. The postal questionnaire was made up into a 20-page booklet with accompanying personalized letter. 2 reminders with further copies of the questionnaire were sent. In each location, 50 births were assigned to an interviewer and 50 designated for the postal questionnaire. Replies to painful and delicate subjects were similar in the 2 groups. Topics which had been identified as laden appeared to produce a number of differences, but some in 1 direction, others in another. There was some support for the prediction that criticisms would be more often reported at interviews than on postal questionnaires. This was unexpectedly found in the work-class women. The addition of a blank sheet on the postal survey illicited a wealth of illustrative material and comments.