[R-sig-ME] p-value for the fixed effect
Patrick Onyango
pogola at Princeton.EDU
Thu Sep 17 15:41:31 CEST 2009
Hi,
First, here are some texts on R that you might find useful:
1. Pinheiro & Bates, 2000. Mixed Effects Model inn S and S-PLUS
2. West et al., 2007. Linear Mixed Models: A Practical Guide USing
Statistical Software.
3. Crawley 2005: The R Book.
These books are gems!!
Second, and briefly: There are 3 approaches known to me; they are as
follows
1. Obtain conditional tests (t-test and F-tests) using summary ( ) or
anova ( ). These tests have some drawbacks (see Pinheiro & Bates,
2000; West et al, 2007; Baayen et al., 2008). Also, the F test in R
is a Type 1, which means from what I have gathered that p values
obtained are sequential, dependent on the order of your fixed effects
in the model. The t-test is not.
2. Likelihood ratios test using the anova( ) function, which you can
use to compare two models that differ in either the random effects
structure in which case you specify REML as method; or the fixed
effects structure in which case you specify method as ML. Please see
above refs for further details. Here the p value is based on a chi-
square distribution, which has its own drawbacks.
3. In lmer, you can also us MCMC commands, which are in the package
languageR. So you will need to load that library. I will let the
experts tell you more about what it does; because I am still reading
about it myself.
Good luck
Patrick
On Sep 17, 2009, at 5:48 AM, FMH wrote:
> Dear All,
>
> Could someone advice me the way to extract the p-value for the
> fixed effect in lme and lmer functions?
>
> Thank you
> Fir
>
>
>
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
More information about the R-sig-mixed-models
mailing list