PURPOSE: The Kaiser Physical Activity Survey (KPAS) was evaluated for test-retest reliability and comparison with direct and indirect measures of physical activity (PA) in 50 women (ages 20-60 yr) with a broad range of physical activity (PA) habits. METHODS: The KPAS, an adaptation of the Baecke usual physical activity survey, was designed specifically to assess activity in women. It provides four summary activity indexes: housework/caregiving, active living habits, sports, and occupation. Summary indexes were compared against direct (Caltrac accelerometer and PA records) and indirect (cardiorespiratory fitness (VO2 peak) and percent body fat) criterion measures of PA. Participants kept detailed PA records for two, 7-d periods, separated by 1 month. Caltrac accelerometers were worn concurrently with the PA records. RESULTS: Intraclass correlations for 1-month test-retest reliability were high for all KPAS indexes (r = 0.79 to 0.91, P < 0.01). Age-adjusted Spearman rho correlations between the KPAS sports/exercise and active living habits indexes were of moderate magnitude for VO2 peak (r = 0.34 to 0.76, P < 0.01) and percent body fat (r = -.30 to -0.59, P < 0.05). KPAS caregiving and occupation indexes were related to Caltrac kcal x d(-1) (r = 0.30 to 0.44, P < 0.05). Correlations between similar activities from the KPAS and PA records ranged from r = 0.03 to 0.64. Daily, habitual activities from the KPAS and PA records had the highest correlations (r > 0.28). Correlations among infrequent activities were lower (r < 0.05). CONCLUSION: The KPAS demonstrated good reliability and was reasonably accurate in detecting regular housework/caregiving, occupation, sports/exercise, and active leisure activities among women with a broad range of physical activity habits.