This study examined stability and reliability of free-living physical activity assessed by pedometer in 69 young female college students (M age = 18.7 years, SD = 1.2, range: 18-25 years; body mass index = 23.2 kg/m2, SD = 0.6) for two complete weeks (Week 1 and Week 2) separated by 12 weeks. During Week 1, participants took an average of 8880 steps/day, SD = 3219, range: 1858-19480; during Week 2 9088 steps/day, SD = 3299, range 1736-16837; t (68) = -.568, p = .572. In both weeks, only Sunday differed significantly (repeated measures analysis of variance; Week 1: p < .0001; Week 2: p < .01) from all other days of the week. The computed intraclass correlation between weeks was moderate (.72). Group mean pedometer-determined physical activity was stable across 2 weeks separated by 12 weeks in this young, healthy sample of women. Individuals within these groups held their rank order to a moderate extent over time. These findings provide important evidence of the reliability of pedometer-determined physical activity data and are of practical importance to study design.