What is it about?
There are two ways how to model dependence of an ordered variable on explanatory variables - one is called "generalized linear model", the other is called "compositional data analysis". We show that the two models are very similar, but not the same. Their difference is pointed out.
Featured Image
Why is it important?
People who need to analyze ordinal data are usually familiar only one of the compared approaches. It is good to be informed about the alternative approach as well as about possible weak points of the approaches.
Perspectives
I hope the publication will help practitioners to answer the question "What should I do with ordinal data?" and it will attract attention of statisticians who are familiar with only one of the compared approaches.
Ondrej Vencalek
Univerzita Palackeho v Olomouci
Read the Original
This page is a summary of: A comparison of generalised linear models and compositional models for ordered categorical data, Statistical Modelling, January 2019, SAGE Publications,
DOI: 10.1177/1471082x18816540.
You can read the full text:
Contributors
The following have contributed to this page
