[R-sig-ME] Is it possible to compute pairwise differences of LS-mean for all factors in a Generalized linear mixed model.
Guillaume Adeux
gu|||@ume@|mon@@2 @end|ng |rom gm@||@com
Thu Jun 20 19:05:13 CEST 2019
Hi Julian,
You should look into the {emmeans} package and the function emmeans. The
vignettes are very nice, the package has a lot of features and from what I
can remember, glmmTMB is handled. Maybe time should be treated as
continuous and you should test slopes between CAP (with the emtrends()
function)?
Nevertheless, you could try different contrasts like:
emmeans(fit_zipoisson2, CAP|time)
or
emmeans(fit_zipoisson2, time|CAP)
or
emmeans(fit_zipoisson2, time*CAP)
You could the wrap the call with cld() from the {multcomp} package to
obtain a compact letter display.
Cheers,
GA2
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
Virus-free.
www.avast.com
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
Le jeu. 20 juin 2019 à 17:55, Julian Gaviria Lopez <
Julian.GaviriaLopez using unige.ch> a écrit :
> Hi everyone,
>
>
> I am fitting the next model with the glmm TMB package:
>
> > fit_zipoisson2 <- glmmTMB(Observations ~ CAP * Time + (1|ID), data=mDATA,
>
> ziformula=~ 1 , family=poisson)
>
> and I obtain as output the next one:
>
>
> > summary(fit_zipoisson2)
>
> Family: poisson ( log )
>
> Formula: Observations ~ CAP * Time + (1 | ID)
>
> Zero inflation: ~1
>
> Data: mDATA
>
>
>
> AIC BIC logLik deviance df.resid
>
> 1252.8 1339.5 -604.4 1208.8 358
>
>
>
> Random effects:
>
>
>
> Conditional model:
>
> Groups Name Variance Std.Dev.
>
> ID (Intercept) 0.03325 0.1824
>
> Number of obs: 380, groups: ID, 19
>
>
>
> Conditional model:
>
> Estimate Std. Error z value Pr(>|z|)
>
> (Intercept) 1.29264 0.14364 8.999 <2e-16 ***
>
> CAPinsC5 -0.70525 0.29149 -2.419 0.0155 *
>
> CAPpreC1 -0.21058 0.21474 -0.981 0.3268
>
> CAPpreC3 -0.14764 0.24991 -0.591 0.5547
>
> Timem1 -0.06505 0.22609 -0.288 0.7736
>
> Timem3 -0.11959 0.21359 -0.560 0.5755
>
> Timem4 -0.37762 0.21934 -1.722 0.0851 .
>
> Timem5 -0.31970 0.22905 -1.396 0.1628
>
> CAPinsC5:Timem1 0.19259 0.47302 0.407 0.6839
>
> CAPpreC1:Timem1 -0.73915 0.49687 -1.488 0.1368
>
> CAPpreC3:Timem1 -1.02677 0.54285 -1.891 0.0586 .
>
> CAPinsC5:Timem3 0.11425 0.43376 0.263 0.7923
>
> CAPpreC1:Timem3 -0.42517 0.35701 -1.191 0.2337
>
> CAPpreC3:Timem3 -0.14454 0.42445 -0.341 0.7335
>
> CAPinsC5:Timem4 0.76611 0.39757 1.927 0.0440 *
>
> CAPpreC1:Timem4 0.26068 0.32745 0.796 0.4260
>
> CAPpreC3:Timem4 -0.06192 0.41907 -0.148 0.8825
>
> CAPinsC5:Timem5 0.30779 0.50160 0.614 0.5395
>
> CAPpreC1:Timem5 0.42848 0.37898 1.131 0.2582
>
> CAPpreC3:Timem5 0.37346 0.35116 1.063 0.2876
>
>
> Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
>
>
>
> Zero-inflation model:
>
> Estimate Std. Error z value Pr(>|z|)
>
> (Intercept) -0.2467 0.1282 -1.924 0.0543 .
>
> ---
>
> Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
>
> How can I observe the changes of the same CAP across time? Regardless if
> it is the intercept or not? E.g.:
> preC5 m1 vs preC5 m2
> preC5 m1 vs preC5 m3
> preC5 m1 vs preC5 m4
> preC5 m1 vs preC5 m4
>
> preC1 m1 vs preC1 m2
> … etc.
>
> And how can I observe the changes
> of the CAPs in relation to the predictor CAP (preC5) over time? E.g.:
> preC1 m1 vs preC5 m1
> preC1 m1 vs preC5 m2
> preC1 m1 vs preC5 m3
> preC1 m1 vs preC5 m4
> preC1 m1 vs preC5 m5
>
> I look for something similar to the Least-Square means and pairwise
> differences functions implemented in other packages for mixed models, such
> as
> the lmerTest:
> https://cran.r-project.org/web/packages/lmerTest/lmerTest.pdf
>
>
> Thanks a lot in advance
>
> Julian Gaviria
> Neurology and Imaging of cognition lab (Labnic)
> University of Geneva. Campus Biotech.
> 9 Chemin des Mines, 1202 Geneva, CH
> Tel: +41 22 379 0380
> Email: Julian.GaviriaLopez using unige.ch
>
> [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-mixed-models using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
Virus-free.
www.avast.com
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
[[alternative HTML version deleted]]
More information about the R-sig-mixed-models
mailing list