Bayesian Posterior Estimation of Logit Parameters with Small Samples

When the sample size is small compared to the number of cells in a contingency table, maximum likelihood estimates of logit parameters and their associated standard errors may not exist or may be biased. This problem is usually solved by “smoothing” the estimates, assuming a certain prior distribution for the parameters. This article investigates the performance of point and interval estimates obtained by assuming various prior distributions. The authors focus on two logit parameters of a 2 × 2 × 2 table: the interaction effect of two predictors on a response variable and the main effect of one of two predictors on a response variable, under the assumption that the interaction effect is zero. The results indicate the superiority of the posterior mode to the posterior mean.

[1]  Dale J. Poirier,et al.  An Empirical Investigation of Wagner's Hypothesis by Using a Model Occurrence Framework , 1995 .

[2]  F. J. Anscombe,et al.  On estimating binomial response relations , 1956 .

[3]  A. Agresti,et al.  Approximate is Better than “Exact” for Interval Estimation of Binomial Proportions , 1998 .

[4]  H. Jeffreys,et al.  The Theory of Probability , 1896 .

[5]  Gary Koop,et al.  Bayesian analysis of logit models using natural conjugate priors , 1993 .

[6]  Richard A. Berk,et al.  Death Penalty Charging in Los Angeles County , 1999 .

[7]  Purushottam W. Laud,et al.  On Bayesian Analysis of Generalized Linear Models Using Jeffreys's Prior , 1991 .

[8]  L. Brown,et al.  Interval Estimation for a Binomial Proportion , 2001 .

[9]  Jeroen K. Vermunt,et al.  'EM: A general program for the analysis of categorical data 1 , 1997 .

[10]  L. A. Goodman.  The Multivariate Analysis of Qualitative Data: Interactions among Multiple Classifications , 1970 .

[11]  Shelby J. Haberman,et al.  Log-Linear Models for Frequency Data: Sufficient Statistics and Likelihood Equations , 1973 .

[12]  Robert E. Kass,et al.  Formal rules for selecting prior distributions: A review and annotated bibliography , 1993 .

[13]  Clifford C. Clogg,et al.  Some Common Problems in Log-Linear Analysis , 1987 .

[14]  J. Gart,et al.  On the bias of various estimators of the logit and its variance with application to quantal bioassay. , 1967, Biometrika.

[15]  T H Burns,et al.  Death penalty. , 1970, Lancet.

[16]  Walter R. Gilks,et al.  A Language and Program for Complex Bayesian Modelling , 1994 .

[17]  Douglas G. Bonett,et al.  Estimating the ratio of two Poisson rates , 2000 .

[18]  Dale J. Poirier,et al.  Jeffreys' prior for logit models , 1994 .

[19]  A Agresti,et al.  On Logit Confidence Intervals for the Odds Ratio with Small Samples , 1999, Biometrics.

[20]  Donald B. Rubin,et al.  Multiple Imputation of Industry and Occupation Codes in Census Public-use Samples Using Bayesian Logistic Regression , 1991 .

[21]  Peter E. Rossi,et al.  Bayesian analysis of dichotomous quantal response models , 1984 .

[22]  J. Lingoes.  The Multivariate Analysis Of Qualitative Data , 1968 .

[23]  David E. Booth,et al.  Analysis of Incomplete Multivariate Data , 2000, Technometrics.

[24]  Jeroen K. Vermunt,et al.  LEM: A general program for the analysis of categorical data. Users manual , 1997 .

[25]  Peter Congdon.  Bayesian statistical modelling , 2002 .

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