The objectives of this study were to: (a) determine if very young children will wear an accelerometer for relatively long periods of time and comply with the protocol for its use; (b) evaluate how well a caregiver can estimate the activity level of his/her infant or toddler when completing an exposure-oriented time activity diary; and (c) compare accelerometer count output with caregiver-provided estimates of children's activity level. This paper describes the statistical relationship between the accelerometer output and the diary activity entries. Results from this study show that young children will wear the accelerometer for multiple data collection days. The children and their caregivers complied in completing the time activity diary, wearing the accelerometer, and videotaping representative activities. The statistical analysis showed a strong and positive relationship between diary entries and accelerometer output (r = 0.42, p less than 0.0001). (Contains 6 tables.)