On the Power of the X Squared Goodness of Fit Test at Signal Plus Noise Alternatives.
CALIFORNIA UNIV DAVIS DEPT OF MATHEMATICS
Pagination or Media Count:
The chi square goodness of fit test is widely used as a means of validation of a probability model. Acceptance of a probability model on the basis of a chi square test is a dangerous procedure in the absence of a thorough knowledge of the power of the test for a broad class of reasonable alternatives. The power of the chi square test is examined via simulation at signal plus noise alternatives to the geometric and the Poisson distribution, and the power is seen to compare unfavorably with that of the likelihood ratio test. Likelihood ratio tests are thus advanced as goodness of fit criteria when signal plus noise alternatives are deemed relevant. Author
- Statistics and Probability