The outcomes of crash tests can be influenced by the initial posture and position of the anthropomorphic test devices (ATDs) used to represent human occupants. In previous work, positioning procedures for ATDs representing adult drivers and rear-seat passengers have been developed through analysis of posture data from human volunteers. The present study applied the same methodology to the development of positioning procedures for ATDs representing six-year-old and ten-year-old children sitting on vehicle seats and belt-positioning boosters. Data from a recent study of 62 children with body mass from 18 to 45 kg were analyzed to quantify hip and head locations and pelvis and head angles for both sitter-selected and standardized postures. In the present study, the 6YO and 10YO Hybrid-III ATDs were installed using FMVSS 213 procedures in six test conditions used previously with children. Data were gathered on a vehicle seat with and without a backless belt-positioning booster at 19-, 23-, and 27-degree seat back angles. The data analysis showed that children sat with their hips further forward and with more reclined pelvis orientations in the sitter-selected posture than in the standardized posture. On the booster, the 6YO ATD head and pelvis positions and orientations corresponded well to the data from similar-size children seated in the standardized posture on the booster, but the 6YO ATD pelvis was about 20 mm rearward of the expected location for similar size children on the booster in sitter-selected postures, and the ATD head was about 20 mm higher than the expected location. Without the booster, the average 6YO ATD hip location was 40 mm rearward of the expected location in sitter-selected postures and the ATD head was 40 mm above the heads of similar-size children. The trends for the 10YO ATD hip- and head-location trends were similar, but the 10YO head was substantially forward of the head locations of similar-size children. Based on these findings, the current ATD positioning procedure was modified by inserting a foam block behind the pelvis and adjusting the 10YO ATD neck angle. A preliminary evaluation using the 10YO ATD indicates that the new procedure produces ATD postures that better represent child passenger postures with respect to hip and head locations and pelvis and head angles.