Multinomial-Poisson homogeneous models for contingency tables. (English) Zbl 1105.62352

Lang, Joseph B.

Ann. Stat. 32, No. 1, 340-383 (2004).

Summary: A unified approach to maximum likelihood inference for a broad, new class of contingency table models is presented. The model class comprises multinomial-Poisson homogeneous (MPH) models, which can be characterized by an independent sampling plan and a system of homogeneous constraints, \(h(m) = 0\), where \(m\) is the vector of expected table counts. Maximum likelihood (ML) fitting and large-sample inference for MPH models are described. The MPH models are partitioned into well-defined equivalence classes and explicit comparisons of the large-sample behaviors of ML estimators of equivalent models are given. The equivalence theory not only unifies a large collection of previously known results, it also leads to useful generalizations and many new results. The practical, computational implication is that ML fit results for any particular MPH model can be obtained directly from the ML fit results for any conveniently chosen equivalent model. Issues of hypothesis testability and parameter estimability are also addressed. To illustrate, an example based on statistics journal citation patterns is given for which the data can be used to test the hypothesis that a certain model holds, but they cannot be used to estimate any of that model’s parameters.

MSC:

62H17	Contingency tables
62F12	Asymptotic properties of parametric estimators
62F05	Asymptotic properties of parametric tests

Keywords:

approximate normality; categorical data; equivalent models; estimability; homogeneous constraint; homogeneous statistic; large-sample inference; restricted maximum likelihood; sampling plan; testability

Software:

SAS; R

References:

[1]	Agresti, A. (1990). Categorical Data Analysis . Wiley, New York. · Zbl 0716.62001
[2]	Aitchison, J. and Silvey, S. D. (1958). Maximum-likelihood estimation of parameters subject to restraints. Ann. Math. Statist. 29 813–828. · Zbl 0092.36704 · doi:10.1214/aoms/1177706538
[3]	Aitchison, J. and Silvey, S. D. (1960). Maximum-likelihood estimation procedures and associated tests of significance. J. Roy. Statist. Soc. Ser. B 22 154–171. · Zbl 0096.34403
[4]	Andersen, A. H. (1974). Multidimensional contingency tables. Scand. J. Statist. 1 115–127. · Zbl 0307.62039
[5]	Apostol, T. M. (1974). Mathematical Analysis , 2nd ed. Addison–Wesley, Reading, MA. · Zbl 0309.26002
[6]	Baker, S. G. (1994). The multinomial–Poisson transformation. The Statistician 43 495–504.
[7]	Bergsma, W. P. (1997). Marginal Models for Categorical Data . Tilburg Univ. Press. · Zbl 0881.62003
[8]	Bergsma, W. P. and Rudas, T. (2002). Marginal models for categorical data. Ann. Statist. 30 140–159. · Zbl 1012.62063 · doi:10.1214/aos/1015362188
[9]	Birch, M. W. (1963). Maximum likelihood in three-way contingency tables. J. Roy. Statist. Soc. Ser. B 25 220–233. · Zbl 0121.14001
[10]	Birch, M. W. (1964). A new proof of the Pearson–Fisher theorem. Ann. Math. Statist. 35 817–824. · Zbl 0259.62017 · doi:10.1214/aoms/1177703581
[11]	Bishop, Y. M. M., Fienberg, S. E. and Holland, P. W. (1975). Discrete Multivariate Analysis . MIT Press. · Zbl 0332.62039
[12]	Chambers, R. L. and Welsh, A. H. (1993). Log-linear models for survey data with nonignorable nonresponse. J. Roy. Statist. Soc. Ser. B 55 157–170. · Zbl 0780.62010
[13]	Chen, T. and Fienberg, S. E. (1974). Two-dimensional contingency tables with both completely and partially cross-classified data. Biometrics 30 629–642. · Zbl 0294.62073 · doi:10.2307/2529228
[14]	Christensen, R. R. (1990). Log-Linear Models . Springer, New York. · Zbl 0711.62050
[15]	Cormack, R. M. and Jupp, P. E. (1991). Inference for Poisson and multinomial models for capture–recapture experiments. Biometrika 78 911–916. · Zbl 0754.62084 · doi:10.1093/biomet/78.4.911
[16]	Ferguson, T. S. (1996). A Course in Large Sample Theory . Chapman and Hall, London. · Zbl 0871.62002
[17]	Fienberg, S. E. (2000). Contingency tables and log-linear models: Basic results and new developments. J. Amer. Statist. Assoc. 95 643–647.
[18]	Fienberg, S. E. and Larntz, K. (1976). Log linear representation for paired and multiple comparisons models. Biometrika 63 245–254. · Zbl 0339.62051 · doi:10.1093/biomet/63.2.245
[19]	Fleming, W. (1977). Functions of Several Variables , 2nd ed. Springer, New York. · Zbl 0348.26002
[20]	Glonek, G. F. V. and McCullagh, P. (1995). Multivariate logistic models. J. Roy. Statist. Soc. Ser. B 57 533–546. · Zbl 0827.62059
[21]	Grizzle, J. E., Starmer, C. F. and Koch, G. G. (1969). Analysis of categorical data by linear models. Biometrics 25 489–504. · Zbl 1149.62317 · doi:10.2307/2528901
[22]	Haber, M. (1986). Testing for pairwise independence. Biometrics 42 429–435. · Zbl 0657.62065 · doi:10.2307/2531063
[23]	Haberman, S. J. (1973). The analysis of residuals in cross-classified tables. Biometrics 29 205–220.
[24]	Haberman, S. J. (1974). The Analysis of Frequency Data . Univ. Chicago Press. · Zbl 0325.62017
[25]	Ihaka, R. and Gentleman, R. (1996). R: A language for data analysis and graphics. J. Comput. Graph. Statist. 5 299–314.
[26]	Ireland, C. T. and Kullback, S. (1968). Contingency tables with given marginals. Biometrika 55 179–188. · Zbl 0155.26701 · doi:10.1093/biomet/55.1.179
[27]	Kullback, S. (1971). Marginal homogeneity of multidimensional contingency tables. Ann. Math. Statist. 42 594–606. · Zbl 0215.54305 · doi:10.1214/aoms/1177693409
[28]	Lang, J. B. (1996a). On the comparison of multinomial and Poisson loglinear models. J. Roy. Statist. Soc. Ser. B 58 253–266. · Zbl 0850.62567
[29]	Lang, J. B. (1996b). Maximum likelihood methods for a generalized class of loglinear models. Ann. Statist. 24 726–752. · Zbl 0859.62061 · doi:10.1214/aos/1032894462
[30]	Lang, J. B. (2000). Maximum-likelihood analysis for useful classes of multinomial–Poisson homogeneous-function models. Technical Report 298, Dept. Statistics and Actuarial Science, Univ. Iowa.
[31]	Lang, J. B. and Agresti, A. (1994). Simultaneously modeling joint and marginal distributions of multivariate categorical responses. J. Amer. Statist. Assoc. 89 625–632. · Zbl 0799.62063 · doi:10.2307/2290865
[32]	Lang, J. B., McDonald, J. W. and Smith, P. W. F. (1999). Association-marginal modeling of multivariate categorical responses: A maximum likelihood approach. J. Amer. Statist. Assoc. 94 1161–1171. · Zbl 1072.62590 · doi:10.2307/2669932
[33]	Lipsitz, S. R., Parzen, M. and Molenberghs, G. (1998). Obtaining the maximum likelihood estimates in incomplete \(R\times C\) contingency tables using a Poisson generalized linear model. J. Comput. Graph. Statist. 7 356–376.
[34]	Lyons, N. I. and Hutcheson, K. (1986). Estimation of Simpson’s diversity when counts follow a Poisson distribution. Biometrics 42 171–176.
[35]	Matthews, J. N. S and Morris, K. P. (1995). An application of Bradley–Terry-type models to the measurement of pain. Appl. Statist. 44 243–255. · Zbl 0821.62081 · doi:10.2307/2986348
[36]	Palmgren, J. (1981). The Fisher information matrix for log linear models arguing conditionally on observed explanatory variables. Biometrika 68 563–566. · Zbl 0477.62039
[37]	Ross, S. M. (1993). Introduction to Probability Models , 5th ed. Academic Press, San Diego, CA. · Zbl 0781.60001
[38]	Serfling, R. J. (1980). Approximation Theorems of Mathematical Statistics . Wiley, New York. · Zbl 0538.62002
[39]	Silvey, S. D. (1959). The Lagrangian multiplier test. Ann. Math. Statist. 30 389–407. · Zbl 0090.36302 · doi:10.1214/aoms/1177706259
[40]	Stigler, S. M. (1994). Citation patterns in the journals of statistics and probability. Statist. Sci. 9 94–108.
[41]	Stokes, M. E., Davis, C. S. and Koch, G. G. (2000). Categorical Data Analysis Using the SAS System , 2nd ed. SAS Institute, Cary, NC.
[42]	Wald, A. (1949). Note on the consistency of the maximum likelihood estimate. Ann. Math. Statist. 20 595–601. · Zbl 0034.22902 · doi:10.1214/aoms/1177729952