The performance of two types of forward facing child restraint systems (CRSs), belt-positioning boosters (BPBs) and CRSs with an integral 5-point harness were compared in frontal and side-impact testing. Performance criteria in frontal impacts (head injury criteria (HIC), chest acceleration, head excursion and knee excursion) was evaluated by comparing a large set of NHTSA-run FMVSS 213 compliance test data generated with the 3-year-old-sized anthropomorphic dummy (ATD). Side-impact performance was evaluated by conducting a series of sled tests and comparing the relative head excursion of a 3-year-old-sized ATD.FMVSS 213 compliance test data shows that the average HIC, chest acceleration, and head and knee excursions are comparable for BPBs and harness CRSs. ATDs in BPBs experienced a slightly higher average HIC, and a slightly lower average head excursion than ATDs in harness CRSs without a tether. On average, tests with a harness CRS and a tether resulted in HIC values that were lower than those with a BPB or a harness CRS without a tether. The tests with a harness CRS with a tether resulted in head excursions that were on average between those of the BPBs and harness CRSs without a tether. For side impacts, the ATDs experienced comparable lateral and vertical head excursions in both types of CRS.