Comparison of Modeling Procedures

Subsections:

The CATMOD, GENMOD, GLIMMIX, LOGISTIC, PROBIT, and SURVEYLOGISTIC procedures can all be used for statistical modeling of categorical data.

The CATMOD procedure treats all explanatory (independent) variables as classification variables by default, and you specify continuous covariates in the DIRECT statement. The other procedures treat covariates as continuous by default, and you specify the classification variables in the CLASS statement.

The CATMOD procedure provides weighted least squares estimation of many response functions, such as means, cumulative logits, and proportions, and you can also compute and analyze other response functions that can be formed from the proportions corresponding to the rows of a contingency table. In addition, a user can input and analyze a set of response functions and user-supplied covariance matrix with weighted least squares. PROC CATMOD also provides maximum likelihood estimation for binary and polytomous logistic regression.

The GENMOD procedure is also a general statistical modeling tool which fits generalized linear models to data; it fits several useful models to categorical data including logistic regression, the proportional odds model, and Poisson and negative binomial regression for count data. The GENMOD procedure also provides a facility for fitting generalized estimating equations to correlated response data that are categorical, such as repeated dichotomous outcomes. The GENMOD procedure fits models using maximum likelihood estimation. PROC GENMOD can perform Type I and Type III tests, and it provides predicted values and residuals. Bayesian analysis capabilities for generalized linear models are also available.

The GLIMMIX procedure fits many of the same models as the GENMOD procedure but also allows the inclusion of random effects. The GLIMMIX procedure fits models using maximum likelihood estimation.

The LOGISTIC procedure is specifically designed for logistic regression. It performs the usual logistic regression analysis for dichotomous outcomes and it fits the proportional odds model and the generalized logit model for ordinal and nominal outcomes, respectively, by the method of maximum likelihood. This procedure has capabilities for a variety of model-building techniques, including stepwise, forward, and backward selection. It computes predicted values, the receiver operating characteristics (ROC) curve and the area beneath the curve, and a number of regression diagnostics. It can create output data sets containing these values and other statistics. PROC LOGISTIC can perform a conditional logistic regression analysis (matched-set and case-controlled) for binary response data. For small data sets, PROC LOGISTIC can perform exact conditional logistic regression. Firth’s bias-reducing penalized-likelihood method can also be used in place of conditional and exact conditional logistic regression.

The PROBIT procedure is designed for quantal assay or other discrete event data. In additional to performing the logistic regression analysis, it can estimate the threshold response rate. PROC PROBIT can also estimate the values of independent variables that yield a desired response.

The SURVEYLOGISTIC procedure performs logistic regression for binary, ordinal, and nominal responses under a specified complex sampling scheme, instead of the usual stratified simple random sampling.

Stokes, Davis, and Koch (2012) provide substantial discussion of these procedures, particularly the use of the FREQ, LOGISTIC, GENMOD, and CATMOD procedures for statistical modeling.