To gain population norms for the short form 36 health survey questionnaire (SF36) in a large community sample and to explore the questionnaire's internal consistency and validity. Postal survey by using a booklet containing the SF36 and several other items concerned with lifestyles and illness. The sample was drawn from computerised registers of the family health services authorities for Berkshire, Buckinghamshire, Northamptonshire, and Oxfordshire. 13,042 randomly selected subjects aged 18-64 years. Scores for the eight health dimensions of the SF36. The survey achieved a response rate of 72% (n = 9332). Internal consistency of the different dimensions of the questionnaire was high. Normative data broken down by age, sex, and social class were consistent with those from previous studies. The SF36 is a potentially valuable tool in medical research. The normative data provided here may further facilitate its validation and use.