This is the third report on traffic accident statistics in the United States from the National Accident Sampling System (NASS) of the National Center for Statistics and Analysis (NCSA). NASS began collecting data from detailed accident investigations in 1979. It is designed to provide basic information on a representative sample of police-reported accidents in the U.S. For 1981, 5,986 such accidents were investigated and reported. When the system reaches its full, planned maturity, it is expected to have data on 15,000 crashes per year. While fewer than half of all motor vehicle accidents are reported to the police they include almost all of the more serious crashes: those in which there is significant personal injury or major property damage. Since the majority of accidents not reported to the police involved no (or only very minor) injury, this report focuses on those accidents reported by the police. They are important to safety research and detailed information on their circumstances and outcomes can be obtained in most cases by investigators. The information in this report is based on investigations of accidents that took place in 1981. In some cases (which are identified in the report itself), these data have been combined with data from the 3,413 crashes investigated in 1979 and the 2,973 crashes investigated in 1980. NASS and the Fatal Accident Reporting System (FARS) provide the Department of Transportation (DOT) with detailed information on traffic accidents. Both programs are operated by the National Highway Traffic Safety Administration (NHTSA). NASS data are used to identify traffic safety problems; and to support research, development, and evaluation of motor vehicle and highway safety countermeasures. Appendix 4 shows the number of accidents investigated from 1979 through 1981, and the primary national estimates derived from them. Estimates of year-to-year changes in accident statistics are not yet possible from NASS data because such changes are masked by the uncertainty of individual annual estimates. Appendix 5 describes NASS estimation procedures and provides illustrative examples of sampling variability for 1981. These measures of variability should be used to estimate the probable error of the estimates in the body of this report. Because of the uncertainty in national projections from NASS data, most of these projections are valid to no more than two significant figures. Thus, most of the information in this report that is derived from NASS data is rounded two significant figures except where additional figures are useful to show sums. Nevertheless, sums are not always exact because of rounding.
Abstract