regression models for ordinal responses a review of methods

18 was utilized to fit both the constrained and the

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: d further to allow for higher dimensions. Model Fitting and Statistical Software With the exception of the Stereotype logistic model, all models described above were fit to the data of Table 3 using the SAS system (SAS Institute, Cary, NC). The proportional odds and continuation ratio models were fit using the LOGISTIC procedure with the logit (logistic) and the cloglog (complimentary log-log) link function specifications, respectively. The polytomous and adjacent category logistic models were fit using the CATMOD procedure with the logit and the alogit specifications for link functions, respectively. The LOGIST procedure of SAS’s supplemental library (version 5.18) was utilized to fit both the constrained and the unconstrained partial proportional-odds models. Peterson and Harrell11 describe the SAS code required to fit the partial proportional odds models. All models were fit through the procedure of maximum likelihood estimation, while the adjacent category model was fit using the weighted least squares procedure. Peterson and Harrell6,12 however, warn against the use of...
View Full Document

Ask a homework question - tutors are online