Approaches to Analyzing Experiments with Factorial Arrangements of Treatments Plus Other Treatments

Abstract

Experiments with factorial arrangements of treatments plus one or more other treatment(s) are sometimes analyzed with a one-way analysis of variance (ANOVA) and means are separated with a multiple comparison. A set of single degree-of-freedom contrasts in a one-way ANOVA, provides formal tests for main effects and interactions. Data from a 2 × 3 factorial experiment that also contained a control were analyzed with a one-way ANOVA with a multiple comparison. Results from this analysis were compared to results obtained from a two-way ANOVA, a one-way ANOVA with pre-planned contrasts, a two-way ANOVA with least squares means comparisons obtained with SAS/general linear models procedure, and a regression model with an indicator variable and random blocks obtained with SAS/Mixed procedure. Results and interpretation differed depending on how the data were analyzed and these differences are discussed.