[R] lme vs. lmer, how do they differ?

[Ben Bolker]

John Sorkin <jsorkin <at> grecc.umaryland.edu> writes:

> 
> windows Vista
> R 2.10.1
> 
> What is the difference (or differences) between lme and lmer? Both appear to
perform mixed effects
> regression analyses.
> Thanks
> John 
> 

  in a nutshell:

  lmer is newer, much faster, handles crossed random effects well (and
generalized linear mixed models), has some support for producing
likelihood profiles (in the development version), and is under rapid development.
It does not attempt to estimate residual degrees of freedom and
hence does not give p-values for significance of effects.
  lme is older, better documented (Pinheiro and Bates 2000), more
stable, and handles 'R-side' structures (heteroscedasticity,
within-group correlations).

  r-sig-mixed-models is a good place for questions about these
packages.

  Ben Bolker