An analysis of crash data is presented using the National Automot Sampling System (NASS) and the Fatality Analysis Reporting System (FARS) for fatality counts. The population is drivers in frontal collisions with air bag restraints. Using NASS, frontal impact mo are grouped into general "test" conditions which will best repres the real world impact environment. These general test conditions include full barrier, left and right offset, and other impact mod Using these general groupings of impact conditions, the analysis further assesses degree of overlap and impact direction to determ more specifically which crash conditions result in highest injury/fatality to drivers with air bags. Injury/fatality risk is also assessed by driver size and body region, with a more detaile analysis of leg injuries. Finally, a preliminary benefits analysi presented for a future frontal, left, offset test procedure. A te procedure has been developed, and is reported on in a separate pa Collinear and oblique, offset, frontal crash testing, at differen widths of overlap, has been conducted with several current model, "target" cars into a standard "bullet" car at closing speeds of a 110 kph. Dummy injury measurements and structural responses provi basis for determining which impact conditions produce the most se environment for occupants with air bags. It appears that the obli impact with over 50 percent overlap produces the most severe responses on the "target" car. Development of this impact configuration into a potential frontal test procedure has been completed using a moving deformable barrier (MDB). For the cover abstract see IRRD E102514.