Motor Vehicle Inspection and Accident Mortality: A Reexamination
用多元回归模型重新检验机动车检测与事故死亡率的关系，发现随机检测和定期检测均能降低死亡率，且随机检测更有效，并进行了成本效益分析。
The relationship of motor vehicle inspection to accident mortality is examined using a multiple regression model. Three inspection schemes are explicitly considered: periodic, random, and no inspection. The fatality rate model indicates that both random and periodic schemes are effective in reducing fatality rates when compared to states with no inspection. Random inspection appears to be the more effective of the two systems. Although tentative in nature, the results of benefit-cost calculations provide additional confirmation of the greater effectiveness of the random system. Research concerning the relationship of motor vehicle inspection to accident mortality has been relatively minor since a flurry of articles appeared in the late 1960's [1][2][5]. At that time automobile safety was a subject of intense national interest fostered by Ralph Nader and his Unsafe at Any Speed published in 1965. This period also saw the creation of the National Highway Traffic Safety Administration (NHSTA) and federally-required energyabsorbing steering columns, lap and shoulder belts, etc. Today, efforts continue at the national level to require all states to enact compulsory periodic motor vehicle inspections. This paper reexamines the potential reduction in highway fatalities of such a program and then evaluates its cost-effectiveness. Motor vehicle inspection began in 1926 with a voluntary program in Massachusetts. Similar programs were initiated in a few other states; and in 1929, Pennsylvania, Maryland, and Delaware began compulsory inspections which were performed in officially designated inspection garages or service stations [21]. In 1970, 30 of the 48 contiguous states had passed laws Joseph G. Van Matre is Professor, School of Business, University of Alabama, Birmingham. He holds the Ph.D. degree. Professor Van Matre is senior author of Statistics for Business and Economics (BPI, 1980), co-editor of Census 1980: Continuing the Factfinder Tradition (Bureau of the Census, 1980); President, Southeastern Region, American Institute for Decision Sciences, and has published in The American Statistician, Decision Sciences, and other journals. George A. Overstreet, Jr. is Associate Professor of Finance, University of Alabama, Birmingham. He holds the Ph.D. degree and has published in the Journal of the American Dental Association, Journal of Dental Education, and Journal of Bank Research. Dr. Overstreet was a Huebner Post-doctoral Fellow, 1978-79. The authors acknowledge the helpful comments of Professors J. David Cummins, John Neter and John Swan on an earlier draft of this paper. The authors also appreciate the work of Graduate Assistants Richard Bums, David Spaulding, and Don Chastain.