Abstract
The likelihood ratio chi-square criterion for testing goodness-of-fit in k cell multinomials is known to overestimate significance for small and moderate sample sizes (see, e.g., Larntz (1978)). Therefore, the usual chi-square approximation to the upper tail of the likelihood ratio statistic G 2, is not satisfactory. Several authors have derived adjustments (e.g., Williams (1976), Smith et al. (1981), Hosmane (1987b)), so that the asymptotic mean of G 2 matches the mean of the asymptotic chi-square distribution in the hope that the distribution of G 2 would improve. In this paper, a new adjustment to G 2 is determined on the basis of the n -1-order term (n being the total number) of the Edgeworth expansion of the distribution of smoothed G 2. Monte Carlo results indicate that the modified G 2 outperforms the unadjusted G 2.
Similar content being viewed by others
References
Barndorff-Nielsen, O. E. and Cox, D. R. (1984). Bartlett adjustments to the likelihood ratio statistic and the distribution of the maximum likelihood estimator. J. Roy. Statist. Soc. Ser. B, 46, 483–495.
Bednarski, T. and Ledwina, T. (1978). A note on biasedness of tests of fit, Mathematische Operationsforschung und Statistik, Series Statistics, 9, 191–193.
Chapman, J. W. (1976). A comparison of the X 2,-2 log R and multinomial probability criteria for significance tests when expected frequencies are small, J. Amer. Statist. Assoc., 71, 854–863.
Cochran, W. G. (1952). The χ2 test of goodness of fit, Ann. Math. Statist., 23, 315–345.
Cohen, A. and Sackrowitz, H. B. (1975). Unbiasedness of the chi-square, likelihood ratio and other goodness-of-fit tests for the equal cell case, Ann. Statist., 3, 959–964.
Cressie, N. A. and Read, T. R. C. (1984). Multinomial goodness-of-fit tests, J. Roy. Statist. Soc. Ser. B, 46, 440–464.
Good, I. J., Gover, T. N. and Mitchell, G. J. (1970). Exact distributions for χ2 and for the likelihood-ratio statistic for the equiprobable multinomial distribution, J. Amer. Statist. Assoc., 65, 267–283.
Hosmane, B. S. (1986). Improved likelihood ratio tests and Pearson chi-square tests for independence in two-dimensional contingency tables, Comm. Statist. A—Theory Methods, 15, 1875–1888.
Hosmane, B. S. (1987a). On the adjusted Wald statistic for testing log-linear hypothesis in categorical data, Sankhyā Ser. B, 49, 199–217.
Hosmane, B. S. (1987b). Improved likelihood ratio test for multinomial goodness of fit, Comm. Statist. A—Theory Methods, 16, 3185–3198.
Knusel, L. and Michalk, J. (1987). Asymptotic expansion of the power function of the two-sample binomial test with and without randomization, Metrika, 34, 31–44.
Koehler, K. J. and Larntz, K. (1980). An empirical investigation of goodness-of-fit statistics for sparse multinomials, J. Amer. Statist. Assoc., 75, 336–342.
Larntz, K. (1978). Small-sample comparisons of exact levels for chi-squared goodness-of-fit statistics, J. Amer. Statist. Assoc., 73, 253–263.
Lawal, H. B. (1984). Comparison of the X 2, Y 2, Freeman-Tukey and Williams' improved G 2 test statistics in small-samples of one-way multinomials, Biometrika, 71, 415–418.
Lawley, D. N. (1956). A general method for approximating to the distribution of likelihood ratio criteria, Biometrika, 43, 295–303.
Siotani, M. and Fujikoshi, Y. (1984). Asymptotic approximations for the distributions of multinomial goodness-of-fit statistics, Hiroshima Math. J., 14, 115–124.
Smith, P. J., Rae, D. S., Manderschield, R. W. and Silbergeld, S. (1981). Approximating the moments and distribution of the likelihood ratio statistic for multinomial goodness of fit, J. Amer. Statist. Assoc., 76, 737–740.
Spruill, M. C. (1977). Equally likely intervals in the chi-square test, Sankhyā Ser. A, 39, 299–302.
Williams, D. A. (1976). Improved likelihood ratio tests for complete contingency tables, Biometrika, 63, 33–37.
Author information
Authors and Affiliations
About this article
Cite this article
Hosmane, B.S. Smoothing of likelihood ratio statistic for equiprobable multinomial goodness-of-fit. Ann Inst Stat Math 42, 133–147 (1990). https://doi.org/10.1007/BF00050784
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF00050784