The flail space model is widely used to assess occupant injury risk potential in full-scale crash tests of roadside safety hardware, e.g. guardrail. A fundamental assumption of this model is that the occupant is unbelted and not airbag restrained. In the early 1980s, these were valid assumptions in the US: cars were not airbag equipped and belt usage rates were around 11%. In today's US vehicle fleet, however, these assumptions are questionable: the belt usage rate is approximately 80% and airbags are required equipment. These changes have significant implications on injury risk computed using the original flail space model. The objective of this study is to contrast flail space model injury risk with the widely accepted dummy-based injury criteria in frontal crashes involving unbelted, belted or airbag restrained occupants. In an analysis of 39 frontal crash tests of speeds ranging from 40 to 97 km/hr, the flail space model was unable to account for variations in occupant risk due to the presence and per...