15. Ordinal Logistic Regression

Introduction to Statistical Modelling
With Dr Helen Brown, Senior Statistician at The Roslin Institute, December 2015
*Recommended TH-cam playback settings for the best viewing experience: 1080p HD
************************************************
Content:
Ordinal Logistic Regression
-Suitable when outcome is ordinal
---an ordered categorical scale
---eg mild, moderate, severe
Ordinal Logistic Regression - Example
-Trial of 3 treatments to lower blood pressure
-Side effect ‘cold feet’ measured on scale of 1-5
-Assess whether severity of cold feet differs between the treatments
Incorporating pre-treatment cold feet as a covariate
-Improves efficiency if pre-treatment cold feet correlated with post-treatment cold feet
Generalised linear models
-Suitable for outcomes with various non-normal distributions
-Note: Not same as general linear models (GLMS) which are for normal outcomes
Types of generalised linear model