Abstract
Statistical methods for comparing two or more injury incidence rates are presented as an aid in selecting operations for ergonomic intervention to reduce work place injuries. The methods assume that injuries occur according to a Poisson process and the analysis is conditional on the total number of injuries. Methods presented include hypothesis tests to compare the ratio of two injury incidence rates and hypothesis tests to compare more than two rates. The power of these comparisons increases with the total number of injuries observed, the Type I error probability, and as the ratio of the injury exposure periods approaches one. Expressions for simultaneous confidence intervals for two or more rates and for the ratios of two or more rates are also given. These general statistical methods can be implemented easily on a personal computer.
Notes
Handled by the Department of Work Measurement/Methods/Ergonomics.