PURPOSE: The ability of the Stanford 7-Day Recall (7-DR), a well known instrument for surveying work and leisure-time physical activity (PA) in epidemiologic studies, to assess levels of habitual PA in men and women was evaluated.METHODS: The 7-DR was administered twice, one month apart. Its accuracy was studied in 77 men and women, aged 20-59 years, by its repeatability and comparison of both administrations of the 7-DR with: fourteen 48-hour physical activity records; fourteen 48-hour Caltrac accelerometer readings; peak oxygen uptake (VO 2 peak) determinations; and percent body fat. These criteria measures were obtained over a year's duration.RESULTS: One month repeatability correlation coefficients for 7-DR total activity were r = 0.60 p =< 0.01 and r = 0.36 p =< 0.05 for men and women, respectively. Comparison of corresponding indices of activity between the 7-DR and the PA record indicated: 1) a closer relationship in men for total (r = 0.58 for visit 10 7-DR and 0.66 for visit 11 7-DR, p =< 0.01), and very hard (r = 0.44 and 0.60, p =< 0.05) activity then in women (r = 0.32 and 0.33, p =< 0.05, and r = 0.21, ns and 0.43, p =< 0.01, respectively); and 2) in general, lower and less consistent associations for hard, moderate, and light activity. Total PA by the 7-DR was significantly associated with Caltrac readings r = 0.54 and 0.45 in men only. 7-DR results were more consistently related to VO 2 peak in men than women, but were significantly related to percent body fat in women only.CONCLUSIONS: The ability of the 7-DR to assess habitual PA was greater for more vigorous than for lower intensity PA.