STUDY OBJECTIVE: To determine the validity of a self administered physical activity questionnaire to be used as part of a screening device for postmenopausal osteopaenia (with additional questions on medical history and calcium intake). DESIGN: A questionnaire was posted to 86 perimenopausal women to enquire about weekly hours spent in non-sedentary activity at work, in the household, and during leisure hours. Subjects who returned the questionnaire were visited at home and asked to complete a four day activity diary and subsequently to undertake a submaximal estimate of VO2 MAX, carried out using a treadmill ergometer. They were interviewed to clarify questionnaire and diary entries. Questionnaire validity was assessed in comparison with the diary estimates of hours of activity and with VO2 MAX. PARTICIPANTS: A total of 86 perimenopausal women aged 43-54 years were randomly selected from a GP list in Hammersmith, London. Thirty five women (41%) returned the questionnaire. They were visited at home, given the diary to complete, and invited to attend the physiology laboratory for VO2 MAX measurements. Twenty six of the 35 (74%) completed the study and were included in the final analysis. MAIN RESULTS: Women spent an average of 51 hours per week in non-sedentary activities. Questionnaire and diary yielded similar results (51.05 versus 51.30 h/wk), and there was a good correlation between diary and questionnaire estimates of total weekly hours of non-sedentary activity (r = 0.45, p < 0.05). Other significant correlations were for standing (r = 0.69, p < 0.01), leisure activities (r = 0.66, p < 0.01), and for light household activities (r = 0.42, p < 0.05). Correlations were better for employed than non-employed subjects. In relation to the diary, the questionnaire correctly classified 60% into the top or bottom half of the distribution activity. Sensitivity and specificity of the questionnaire were both equal to 61.5%. CONCLUSIONS: The questionnaire is useful for classifying subjects according to their level of activity, especially when administered in conjunction with an interview. The four day diary provided a useful reference measure and a focus for discussing activity patterns during an interview related to the questionnaire responses.