Measuring Differential Safety performance among air carriers

Richard Golaszewski, Earl Bomberger
1986 unpublished
This paper develops a technique suitable for measuring differential levels of safety performance among air carriers and groups of air carriers. It is based on outlier theory-finding groups or individuals whose performance is significantly different from others. Airline accidents are infrequent phenomena which limit the utility of standard statistical analysis techniques in determining whether a particular carrier is more or less safe than other similar carriers. Also, the airline industry has
more » ... dergone significant restructuring since the advent of deregulation in 1978. The methodology developed herein has been tested using the Department of Transportation standards for groupsing air carriers. However, it also could be applied to test for differences in safety performance between "old line" carriers and the post-deregulation new entrants. Using data for the 1978 to 1982 time period, the methodology is applied to identify whether differences in accident rates occur among groups of air carriers. The results show that there are statistically significant differences in accident rates between groups of air carriers. Specifically, DOT groups air carriers on the basis of annual revenues. The results show that accident rates are inversely related to carriers size. The methodology also is applied to hypothetical data for individual air carriers within the majors group (annual revenues of more than $1 billion) to show how it can be applied in this context. (Data for actual carriers was not used so as not to prejudge the safety performance of any particular carriers. This was done because the DOT schema for grouping air carriers may not produce a homogeneous set of carriers with respect to critical safety variables.) The results of the analysis show that it is possible to use relatively sparse data to measure subtle differences in safety performance among air carriers that belong to a homogeneous group.
doi:10.22004/ag.econ.311809 fatcat:xlsacimzxnf37jay42qxaobjl4