Citation

Bhattacharya SK. Biometrika 1967; 54(1/2): 324.

Copyright

(Copyright © 1967, Biometrika Trust, Publisher Oxford University Press)

DOI

10.2307/2333881

PMID

unavailable

Abstract

A bivariate correlated Poisson model for the number of accidents sustained by a set of individuals exposed to a risk situation in two successive periods of time is considered. It is shown that a selection of individuals free from accidents in the first period reduces the average number of accidents to be expected in the next period, provided that the accident proneness varies from individual to individual with an arbitrary non-degenerate distribution.