An intuitive explanation of generalised linear models

Abstract

This paper offers an intuitive description of generalised linear models so that agricultural researchers may capitalise on the potential of this approach for data analysis and not be deterred by its apparent sophistication. The familiar analysis of variance and regression techniques are used as a basis upon which to develop new concepts, these being presented in a non-mathematical manner whenever possible by relying on diagrammatical representations in conjunction with the stepwise analysis of a simple set of agronomic data. The paper concludes by taking a more detailed look at the logit analyses performed on 3 large sets of data concerning lamb survival as reported in another paper in this issue.

Keywords:

• variance analysis
• data analysis
• models
• mathematical models
• lambs