Accident Proneness in the Industrial Setting

Abstract

The central notion of the accident proneness concept is that people exposed to equivalent hazards do not have an equal number of accidents. If people were equally accident prone, one would expect accidents to be distributed according to chance. Using accident data collected at Shell Oil Company's Manufacturing Complex in Deer Park, Texas, the present study explored the proneness concept for major (OSHA recordable) and minor accidents by comparing the observed distribution of accidents to a chance distribution. The database contains information on 7131 accidents which occurred between 1981 and 1986. The methodology used to create expected values employed a Poisson distribution and assumed that accidents are distributed randomly among the population at risk. The minor accident data was also analyzed by job family. Chi-square analyses of the differences between the expected and observed distributions were found to be statistically significant, including within each job family. The data for minor accidents indicates a striking difference between the expected and actual distributions. Many more people suffered repeat accidents than would be predicted by chance. Approximately 3.4% of the employees accounted for 21.5% of the accidents. While the differences for major accidents was statistically significant, these results are not nearly so striking. The statistical effects are largely due to five employees who were involved in three major accidents in the five year period. In the context of this very large industrial setting, the problem of individuals having repeated minor accidents is significant and merits attention in developing safety interventions.