[R-sig-ME] "slice(SAS)" with R

[Thompson,Paul]

I am a SAS user. 

The SLICE command is a method of performing simple main effects.

If you have a 2-way ANOVA-like model

     A 
   1   2

 1
B
 2

You may wish to examine the interaction more closely by looking at the A factor difference for B=1 and B=2, and the B-factor difference for A=1 and A=2. That is, the slice performs conditional tests of factors.

-----Original Message-----
From: r-sig-mixed-models-bounces at r-project.org [mailto:r-sig-mixed-models-bounces at r-project.org] On Behalf Of ONKELINX, Thierry
Sent: Monday, September 12, 2011 3:08 AM
To: fernando barbero; r-sig-mixed-models at r-project.org
Subject: Re: [R-sig-ME] "slice(SAS)" with R

Dear Fernando,

You will have to tell us what the slice command does. Otherwise it will be hard to give you advice.

Best regards,

Thierry

> -----Oorspronkelijk bericht-----
> Van: r-sig-mixed-models-bounces at r-project.org [mailto:r-sig-mixed-models-
> bounces at r-project.org] Namens fernando barbero
> Verzonden: vrijdag 9 september 2011 16:47
> Aan: r-sig-mixed-models at r-project.org
> Onderwerp: [R-sig-ME] "slice(SAS)" with R
> 
> Hello all, I have been working with data from 8 different tree populations of an
> american beech, currently I have been recording the filling of seeds (binary
> response variable)and have fitted a GLMM with lme4 which has this
> structure:
> 
> 
> 
> model<-lmer(
> seedfilling~population+(1|population:tree)+typewing+typewing*population,fam
> i
> ly=binomial,data1)
> 
> 
> 
> where:
> 
> seedfilling is the binary response variable (1= filled seed,0=empty seed)
> 
> population is a fixed effect of population, I have eight different populations
> 
> tree is a random effect nested under population (I harvested 15 trees randomly
> chosen in every population)
> 
> typewing is the type of wing, as seeds in this tree species can have 2 or 3 wings
> 
> typewing * population is an interaction effect
> 
> I recorded the filling of 150 seeds in every tree
> 
> 
> 
> So far so cool. I have tested the significance of all effects with LRTs and they all
> showed to be significant, now I want to study if there are any differences
> between populations (If I had a linear mixed model I would be using the glht
> function of Multcomp package). Colleagues that work with SAS have told me,
> and in fact they have carried out,  this analysis using the "slice" command but as
> I work with R I want to do it with it, do you know if this can be carried out?
> 
> Best regards
> 
> Fernando
> 
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models

_______________________________________________
R-sig-mixed-models at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models

-----------------------------------------------------------------------
Confidentiality Notice: This e-mail message, including any attachments,
is for the sole use of the intended recipient(s) and may contain
privileged and confidential information.  Any unauthorized review, use,
disclosure or distribution is prohibited.  If you are not the intended
recipient, please contact the sender by reply e-mail and destroy
all copies of the original message.