[R-sig-ME] statistical significance of group level intercepts in glmer

[Maaz Gardezi]

Thanks for the clarification.

Maaz

On May 30, 2017 9:00 AM, "Ben Bolker" <bbolker at gmail.com> wrote:

> Unfortunately, you *can't* test significance of conditional modes (the
> values that ranef()) returns; this is the price you pay for treating
> them as random variables. However, you *can* extract the 'conditional
> variances' of the conditional modes (see p. 28 of the vignette that
> comes with lme4).  You can visualize these conditional variances via
> `lattice::dotplot(ranef(fitted_model,condVar=TRUE))`.  The development
> version of lme4 (on Github) has an as.data.frame method that makes it
> easier to extract these values ...
>
>
> On Tue, May 30, 2017 at 12:16 AM, Maaz Gardezi <maaz.gardezi at gmail.com>
> wrote:
> > Hello,
> >
> > I am using glmer to model a mixed effects logistic regression. The output
> > below shows the random intercepts and slope (zcwcexp) for 22 groups. I
> was
> > wondering if there is any way of finding whether these are statistically
> > significant?
> >
> >> ranef(addRandomCrossLevelInteraction)
> > $subject
> >    (Intercept)      zcwcexp
> > 1   0.08913319 -0.007245956
> > 2  -0.06997405  0.009194926
> > 3  -0.77408045  0.247743815
> > 4   0.07087457  0.048671025
> > 5   0.16529502 -0.025846944
> > 6   0.21978190  0.020381977
> > 7   0.22628857 -0.060421227
> > 8   0.67790616 -0.183998656
> > 9   0.25113625 -0.109811064
> > 10 -0.14569194 -0.039863882
> > 11  0.13872081 -0.074646309
> > 12 -0.23401794  0.069596589
> > 13  0.96193273 -0.195829039
> > 14 -0.46834542  0.092182158
> > 15 -0.59465656  0.158177344
> > 16 -0.06963680 -0.019127508
> > 17 -0.25015416  0.067321201
> > 18  0.32001648 -0.100452172
> > 19 -0.40396733  0.120731775
> > 20  0.07413768 -0.061477287
> > 21 -0.08520228  0.044519621
> > 22 -0.09427640  0.002166830
> >
> > Thanks!
> >
> > Maaz
> >
> >         [[alternative HTML version deleted]]
> >
> > _______________________________________________
> > R-sig-mixed-models at r-project.org mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>

	[[alternative HTML version deleted]]