Analysing percentage data using SAS procedures for generalized linear models

摘要

In many field trials visually assessed data are collected as percentages. Such data can be analysed by a Generalized Linear Model (GLM). The Quasi-Likelihood approach allows an analysis, which adequately accounts for the presence of experimental error in .field trials. Experimental designs with more than one error term can be accomodated by introducing random effects. This leads to so-called Generalized Linear Mixed Models (GLMMs). The present paper discusses implementation of these methods using SAS.