Comparing analyses of unbalanced split-plot experiments

Abstract

Several procedures for constructing confidence intervals and testing hypotheses about fixed effects in unbalanced split-plot experiments have previously been presented and discussed by Remmenga and Johnson. They recommended a few of the procedures they considered as useful and reliable procedures. Since the advent of the SAS^® MIXED procedure, mixed model analyses with restricted maximum likelihood (REML) estimates of the variance components are easily accessible to researchers. This paper compares the analysis of unbalanced split-plot experiments using mixed model procedures with REML estimates of the variance components to the previously established procedures by means of additional simulation studies.

Keywords:

• Unbalanced
• Split-plots
• Mixed models
• Variance components
• REML