Abstract
When the chi-square test is used in the manner suggested by Mr. Dropkin, the hypothesis tested is that all insureds have the same Poisson frequency distribution (with identical means). Rejection of the hypothesis does not necessarily imply a non-Poisson frequency distribution, if individual insureds have different mean frequencies. The “Binomial Variance” test suggested by Mr. Dropkin is invalid.
Volume
LXXIV
Page
424-428
Year
1987
Categories
Actuarial Applications and Methodologies
Ratemaking
Classification Plans
Financial and Statistical Methods
Loss Distributions
Frequency
Business Areas
Automobile
Personal
Financial and Statistical Methods
Credibility
Publications
Proceedings of the Casualty Actuarial Society
