This paper summarizes work relating to the assessment of societal benefits of side impact protection. National Crash Severity Study (NCSS) and National Accident Sampling System (NASS) accident data technigues were reviewed with respect to the reliability of output information concerning the distribution of side impact accidents by impact severity and relationships between injury and impact severity.NCSS and NASS are confounded by errors and inadequacies, primarily as a result of improper accident reconstruction based upon the CRASH computer program. Based on review of several sample cases, it is believed that the NCSS/NASS files underestimate Lower severities and overestimate higher severities in side impact, with delta-V errors probably overestimated by 25-30 percent in the case of the more serious accidents. These errors cannot be properly quantified except on a case-by-case basis. They introduce unknown biases into NCSS/NASS.A review of a sampling of case materials from the NCSS/NASS files indicates that it will probably not be possible to recover enough basic data on sufficient side-impact NASS cases to allow accurate estimates of accident severity. In addition to the inaccuracies associated with the struck vehicle C.G. delta-V estimates of CRASH, occupant exposure severity, at least for near-side occupants, should be measured in terms of the door velocity. No straightforward basis is available to derive these estimates except through a detailed case-by-case evaluation. Such a procedure is not possible in a significant percentage of cases, due to inadequate data collection procedures.