[R-sig-ME] Questions on the results from glmmPQL(MASS)

David Duffy David.Duffy at qimr.edu.au
Mon Dec 8 06:40:33 CET 2008

On Sun, 7 Dec 2008, zhijie zhang wrote:

> Dear Rusers,
>   Sorry for re-post my question in this list. Some person has recommend
> that this list is more specific for me.
>   I have used R,S-PLUS and SAS to analyze the sample data "bacteria" in
> MASS package. Their results are listed below.
> Q3:In SAS and other softwares, we can get p-values for the random effect
> 'sigma', but i donot see the p-values in the results of R/SPLUS. I have used
> attributes() to look for them, but no p values. Anybody knows how to get
> p-values for the random effect 'sigma',.

Someone answered your specific questions before.
Because this is a simple random intercept model, you can also use the
glmmML R package, which gives:

glmmML(y ~ trt + I(week > 2), cluster=ID, family=binomial, data=bacteria)

                    coef se(coef)      z Pr(>|z|)
(Intercept)      3.5482   0.6933  5.118 3.09e-07
trtdrug         -1.3667   0.6768 -2.020 4.34e-02
trtdrug+        -0.7826   0.6830 -1.146 2.52e-01
I(week > 2)TRUE -1.5986   0.4754 -3.363 7.71e-04

Standard deviation in mixing distribution:  1.242
Std. Error:                                 0.4024

Residual deviance: 192.3 on 215 degrees of freedom      AIC: 202.3

You can then construct a test for the random effect.

| David Duffy (MBBS PhD)                                         ,-_|\
| email: davidD at qimr.edu.au  ph: INT+61+7+3362-0217 fax: -0101  /     *
| Epidemiology Unit, Queensland Institute of Medical Research   \_,-._/
| 300 Herston Rd, Brisbane, Queensland 4029, Australia  GPG 4D0B994A v

More information about the R-sig-mixed-models mailing list