Many studies of accident data assign a large portion of culpability to the human part of the driver-vehicle-environment system. However, detailed and quantitative analyses of the driver's contribution to accident causation or accident avoidance are not generally available. Field accident data have not been able to provide sufficient insight to driver behavior. A possible supplement to field data is clinical data on driver behavior acquired from carefully planned experimental scenarios in a proving ground environment.

The considerations important to planning of clinical driver behavior studies are discussed, together with limitations of this approach. Recent studies conducted by industry and government are summarized. A study conducted by Calspan Corporation under General Motors sponsorship is reviewed in detail.

Clinical driver behavior studies will continue since there are few altematives. However, the results should be compared with field experience and design practice to support their applicability. The clinical approach may be useful for evaluating advanced driver training.