[R-sig-ME] Help on Multiple comparisons with lmer

[David Winsemius]

On Aug 10, 2011, at 3:30 PM, Alejandro Wainselboim wrote:

> Hi everyone,
>
>                     I am currently running a mixed model on response  
> type ("correct" or "incorrect") with the following model:
>
> ModelAngel=lmer(response == "correct" ~ 1 + typeF * group + (1| 
> subject), data=DatosAngel, family="binomial")
>
> This works fine.
>
> I then try to run a multiple comparison on the results on the factor  
> typeF with the following script:
>
>  glht(ModelAngel, linfct=mcp(grupo="Tukey"))
>
> But when I do so, I get the following error message:
>
>
> In mcp2matrix(model, linfct = linfct) :
>   covariate interactions found -- default contrast might be  
> inappropriate
>

Looks more like a warning message than an error.

-- 

David Winsemius, MD
West Hartford, CT