Using Analysis of Covariance with Unequal Slopes to Increase Efficiency and Information Obtained from Designed Experiments

Abstract

Horticulturists often perform experiments involving both qualitative and quantitative factors. Sometimes the quantitative factor is a continuous variable (covariate) measured on each experimental unit and data can be analyzed by analysis of covariance (ANCOVA). ANCOVA is a powerful and flexible technique for extracting maximum information from a data set. Data from an experiment designed to compare the productivity of three peach ( Prunus persicaL. Batsch) genotypes were used to determine if genotype influenced average fruit weight, while accounting for the variation explained by the linear relationship between fruit weight and crop density. The data set was analyzed with SAS’s MIXED procedure to demonstrate a strategy for analyzing experiments with a qualitative variable plus a covariate.