Hiya,
I'm fairly new to MuMIn (and plotting predicted values in R, for that
matter) and I have quite a basic question:
I want to plot model-averaged estimates with their 95% CI's for the
supported interaction (Infection x ecs). I used the following data:
> head(summ12b)
ld sum12 Infection ecs
1 13 78.0 Infected Enlarged
2 15 134.5 Uninfected Enlarged
3 14 91.5 Infected Control
4 16 62.0 Infected Reduced
5 17 93.6 Infected Control
6 15 104.6 Uninfected Control
>s12b<-lm(sum12~(Infection+ecs)^2+ld,data=summ12b)
>ss12b<-dredge(s12b,trace = TRUE, rank = "AICc")
>top.ss12b<- get.models(ss12b, subset = delta<2)
>ad<-summary(model.avg(top.ss12b))
and here's what I've got:
>ad
Call:
model.avg.default(object = top.ss12b)
Component models:
df logLik AICc Delta Weight
1234 8 -112.95 249.10 0.00 0.47
124 7 -115.31 249.95 0.86 0.31
1 4 -120.44 250.55 1.45 0.23
Term codes:
ecs Infection ld ecs:Infection
1 2 3 4
Model-averaged coefficients:
Estimate Std. Error Adjusted SE z value
Pr(>|z|)
(Intercept) 112.737 30.847 31.505 3.578
0.000346 ***
ecsEnlarged 23.817 13.376 13.798 1.726
0.084324 .
ecsReduced -10.597 8.352 8.740 1.212
0.225364
InfectionUninfected 7.493 9.381 9.893 0.757
0.448793
ld -3.477 1.763 1.865 1.864
0.062252 .
ecsEnlarged :InfectionUninfected 23.308 14.740 15.510 1.503
0.132892
ecsReduced :InfectionUninfected -22.112 12.589 13.307 1.662
0.096572 .
---
Signif. codes: 0 *** 0.001 ** 0.01 * 0.05 . 0.1 1
Full model-averaged coefficients (with shrinkage):
(Intercept) ecsEnlarged ecsReduced InfectionUninfected ld
ecsEnlarged :InfectionUninfected
112.7373 23.8169 -10.5966 5.7952
-1.6285 18.0261
ecsReduced :InfectionUninfected
-17.1008
Relative variable importance:
ecs Infection ecs:Infection ld
1.00 0.77 0.77 0.47
I know that I can obtain 95% Ci's by using ad$avg.model, however, how to
derive the estimates and CI's for ALL the groups and their factor levels
included in the interaction?
Thanks for your help!
cheers,
Kasia
[[alternative HTML version deleted]]
______________________________________________
[email protected] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
