Abstract: To explore the radical causes of deviations in dynamic test for child restraint systems in different laboratories, both simulation and test verification are conducted on universal and semi-universal child restraint systems with P3 dummy based on national standard GB27887, to quantitatively analyze the degree of effects of each factor on the deviations in dynamic tests.  The results show that the difference of impact waveform in different laboratories, the foam stiffness attenuation coefficient of sled cushion, the initial installation angle of supporting leg and the friction coefficient of safety belt guide have significant effects on the scatter of test data. In the process of subsequent national standard revision, it is necessary to have a quantitative control on key factors, to reduce the test discrepancy between laboratories, and enhance the consistency of tests.

Key words: child restraint system, test, injury criteria, affecting factor, discreteness