Overview

This book is a step-by-step data story for analysing ordinal data from start to finish. The book is for researchers, statisticians and scientists who are working with data sets where the response is ordinal. This type of data is common in many disciplines, not just in surveys (as is often thought). For example, in the biological sciences, there is an interest in understanding and predicting (growth) stage (of a plant or animal) based upon a multitude of factors. This is true in environmental sciences (for example: stage of a storm), chemical sciences (for example: type of reaction), and physical sciences (for example: stage of damage when force is applied), medical sciences (for example: degree of pain) and social sciences (for example: demographic factors like social status categorized in brackets) as well. There has been no complete text about how to model an ordinal response as a function of multiple numerical and categorical predictors. There has always been a reluctance and reticence towards ordinal data as it lies in a no-man's land between numerical and categorical data. Examples from health sciences are used to illustrate in detail the process of how to analyse ordinal data, from exploratory analysis, to modelling, to inference and diagnostics. This book also show how Likert type analysis is often used incorrectly and discuss the reason behind it. Similarly, it discusses the methods related to Structural Equations and talk about appropriate uses of this class of methods. The text is meant to serve as a reference book and to be a how-to" resource along with the "why" and "when" for modelling ordinal data"--