ABSTRACT

This book uses mixed linear models as a syntax for richly parameterized linear models. At least two other systems have been proposed, one based on Gaussian Markov random fields, proposed by Rue & Held (2005), and another proposed by Lee et al. (2006), which might be viewed as extending mixed linear models although the analytic approach is somewhat different. This chapter gives a brief overview of these two syntaxes and their associated analytic approaches, to contrast them with mixed linear models and to give interested readers a place to begin. Each alternative approach is treated briefly, so it may seem that I’ve given undue emphasis to their controversial aspects. However, this is inevitable in a brief overview and disgruntled readers might recall that I’ve been more critical of mixed-linear-model theory and methods than any other treatment I’ve ever seen. Section 7.1 gives a schematic comparison of the three competing syntaxes; this is followed by a section for each alternative syntax.