This study aimed to determine the criterion validity of the Short Form 36 health survey questionnaire (SF-36) in a large community sample, and to explore the instrument's internal consistency and validity in groups reporting different levels of ill-health. A postal survey was undertaken using a questionnaire booklet, containing the SF-36 and a number of other items concerned with lifestyles and illness. The questionnaire booklet was sent to 13,042 randomly selected subjects between the ages of 18-64 years, drawn from Family Health Services Authority (FHSA) computerized registers for Berkshire, Buckinghamshire, Northamptonshire and Oxfordshire. This paper is based upon the 9332 (72%) responses gained. Scores for the functional status and well-being scales of the SF-36 were used as outcome measures. The response rate for the questionnaire booklet was 72%. Internal consistency of domains was found to be high, both for the sample as a whole, and when broken down by specific subgroups. Criterion validity was assessed by comparing scores for the seven multi-item dimensions assessing functional status and well being with a single global health question. The global question was the first item of the SF-36 and asks respondents to evaluate their health 'overall'. Statistically significant trends were observed for decreasing SF-36 scores (i.e., those indicating greater health problems) with worsening self-rated general health. These results provide further psychometric evidence for the use of the SF-36 when used with groups reporting varying extents of ill-health.