The R package MultOrd implements composite likelihood estimation in the class of multivariate ordinal regression models. A flexible modelling framework for cross-sectional as well as longitudinal observations is set up, which allows different error structures. Two different link functions are employed, by assuming a multivariate normal and a multivariate logistic distribution for the latent variables underlying the ordinal outcomes. In order to deal with the issue that absolute location and absolute scale are not identifiable in ordinal models, several restrictions on the parameter space are imposed either by fixing location parameters and/or by restricting the full covariance matrix to be a correlation matrix. In addition, constraints on both coefficient as well as threshold parameters can be imposed. Standard errors are computed using the Godambe information matrix.
see multord
, CMOR
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.