You are not currently logged in.
Access JSTOR through your library or other institution:
If You Use a Screen ReaderThis content is available through Read Online (Free) program, which relies on page scans. Since scans are not currently available to screen readers, please contact JSTOR User Support for access. We'll provide a PDF copy for your screen reader.
On the Chi-Square Goodness-of-Fit Statistic for Bivariate Discrete Distributions
S. Loukas and C. D. Kemp
Journal of the Royal Statistical Society. Series D (The Statistician)
Vol. 35, No. 5 (1986), pp. 525-529
Stable URL: http://www.jstor.org/stable/2987967
Page Count: 5
Since scans are not currently available to screen readers, please contact JSTOR User Support for access. We'll provide a PDF copy for your screen reader.
Preview not available
Assessing goodness-of-fit for bivariate discrete distributions, using the classical Pearson chi-squared statistic, is much less straightforward than in univariate situations due to the relatively large number of classes with low expectation and the two-based structure. In this paper three procedures for grouping into chi-squared classes are examined (one row-based, one column-based, one using ordered expected-frequencies). Approximate asymptotic power has been computed for each procedure, using two minimum-expected-frequency criteria (4 and 1), applied to a variety of bivariate Poisson and compound Poisson distributions. In the light of these comparisons, the 'ordered-frequencies' method is recommended as a simple systematic standard procedure. The choice of minimum group size had relatively little effect.
Journal of the Royal Statistical Society. Series D (The Statistician) © 1986 Royal Statistical Society