Simultaneously Modeling Joint and Marginal Distributions of Multivariate Categorical Responses

Abstract We discuss model-fitting methods for analyzing simultaneously the joint and marginal distributions of multivariate categorical responses. The models are members of a broad class of generalized logit and loglinear models. We fit them by improving a maximum likelihood algorithm that uses Lagrange's method of undetermined multipliers and a Newton-Raphson iterative scheme. We also discuss goodness-of-fit tests and adjusted residuals, and give asymptotic distributions of model parameter estimators. For this class of models, inferences are equivalent for Poisson and multinomial sampling assumptions. Simultaneous models for joint and marginal distributions may be useful in a variety of applications, including studies dealing with longitudinal data, multiple indicators in opinion research, cross-over designs, social mobility, and inter-rater agreement. The models are illustrated for one such application, using data from a recent General Social Survey regarding opinions about various types of government s...

[1]  J. Aitchison,et al.  Maximum-Likelihood Estimation of Parameters Subject to Restraints , 1958 .

[2]  S. D. Silvey,et al.  The Lagrangian Multiplier Test , 1959 .

[3]  S. D. Silvey,et al.  Maximum-Likelihood Estimation Procedures and Associated Tests of Significance , 1960 .

[4]  John Aitchison,et al.  Large‐Sample Restricted Parametric Tests , 1962 .

[5]  M. W. Birch.  Maximum Likelihood in Three-Way Contingency Tables , 1963 .

[6]  P. McCullagh,et al.  Generalized Linear Models , 1972, Predictive Analytics.

[7]  S. Haberman.  The Analysis of Residuals in Cross-Classified Tables , 1973 .

[8]  P. Holland,et al.  Discrete Multivariate Analysis. , 1976 .

[9]  G G Koch,et al.  A general methodology for the analysis of experiments with repeated measurement of categorical data. , 1977, Biometrics.

[10]  Juni Palmgren,et al.  The Fisher information matrix for log linear models arguing conditionally on observed explanatory variable , 1981 .

[11]  Leo A. Goodman,et al.  Association models and the bivariate normal for contingency tables with ordered categories , 1981 .

[12]  Log-linear models for correlated marginal totals of a contingency table , 1985 .

[13]  M. Haber.  Maximum likelihood methods for linear and log-linear models in categorical data , 1985 .

[14]  Morton B. Brown,et al.  Maximum likelihood methods for log-linear models when expected frequencies are subject to linear constraints , 1986 .

[15]  J. Dale.  Global cross-ratio models for bivariate, discrete, ordered responses. , 1986, Biometrics.

[16]  DEPENDENCE FUNCTIONS FOR CONTINUOUS BIVARIATES DENSITIES , 1987 .

[17]  P. Holland,et al.  Dependence function for continuous bivariate densities , 1987 .

[18]  A log-linear model for binary crossover data , 1988 .

[19]  G G Koch,et al.  Some general methods for the analysis of categorical data in longitudinal studies. , 1988, Statistics in medicine.

[20]  P. McCullagh,et al.  Generalized Linear Models , 1992 .

[21]  M. Becker.  On the bivariate normal distribution and association models for ordinal categorical data , 1989 .

[22]  N. Saitou,et al.  Maximum likelihood methods. , 1990, Methods in enzymology.

[23]  N M Laird,et al.  Maximum likelihood regression methods for paired binary data. , 1990, Statistics in medicine.

[24]  S. Zeger,et al.  Multivariate Regression Analyses for Categorical Data , 1992 .