Psychological research on accidents: some methodological considerations

Abstract

Deviation of an obtained accident-frequency distribution from the Poisson may merely reflect differences in exposure; lack of deviation may be consistent with a high correlation between two time periods. Correlation studies suffer because of the coarseness of the criterion. Four refinements in accident research are suggested: systematic collection of information on "near-accidents" and critical behaviors; better differentiation between "personal" and "situational" accidents; more complete information on exposure; and computation of correlations between individuals' accident records and scores on theoretically-related predictor variables. (PsycInfo Database Record (c) 2025 APA, all rights reserved)