Making Sense of the Mixed Model Analyses Available in R.

The R programming language and software environment is becoming increasingly popular among statisticians and researchers, and this has driven a proliferation of functions available to fit linear mixed models. This provides a great deal of flexibility in choosing a mixed model solution appropriate for data set, but this comes at the expense of simplicity and ease of use for experimentalists.
Many package authors, particularly those of the well-established packages nlme and lme4, support a standard model syntax and common set of generic functions for extracting results. This allows code based on previous analyses to be easily updated to take advantage of different algorithms. Other package authors choose idiosyncratic syntax for specifying fixed and random effects and produce model objects that are not interoperable with other widely used packages.
Standard mixed model analyses based on the nlme and lme4 libraries are illustrated using data sets representing common agricultural experiments. These analyses are contrasted with analyses from a selection of newer packages using different parameter estimation algorithms. The ease by which results from different packages can be incorporated into reusable code is demonstrated by plotting mean estimates with error bars and compact letter displays using custom code integrating multiple R functions.