[R-sig-ME] Looking for residuals in clmm summary output

[Ben Bolker]

On 13-11-12 02:40 PM, aline.frank at wsl.ch wrote:
> Hello
> 
> I am working with ordinal data using the clmm() mixed models approach
> of the R package "ordinal". My interest is in analyzing the random
> effects of my model, inclusive the residual term. However, the
> summary of my model does not inlude the residuals. Does this mean
> that my residuals are "hidden" in one of the random effects, or is
> there a way to get the residuals anyway? Below you see my model
> summary output.
> 
> Thanks for every hint!
> 
> Aline

My first thought is that since fitted(model) works, you might be able to
used fitted(model)-observed, but on second thought, you're going to have
to figure out what scale the 'fitted' value is on and how it relates to
the predicted value of the response ...






> 
> 
> 
> model <-
> clmm(trait~1+(1|Block_Nr)+(1|Pop_Nr)+(1|Fam_Nr)+(1|Block_Nr:Pop_Nr),data=dat)
>
>  Data: frost damage on the plants in levels 0-5
> 
> Output R:
> 
> Cumulative Link Mixed Model fitted with the Laplace approximation
> 
> formula: trait ~ 1 + (1 | Block_Nr) + (1 | Pop_Nr) + (1 | Fam_Nr) +
> (1 |      Block_Nr:Pop_Nr) data:    dat
> 
> link  threshold nobs logLik   AIC     niter     max.grad cond.H logit
> flexible  4018 -2015.57 4047.14 660(2644) 2.93e-03 3.6e+01
> 
> Random effects: Groups          Name        Variance Std.Dev. 
> Block_Nr:Pop_Nr (Intercept) 0.0000   0.0000 Fam_Nr
> (Intercept) 0.1085   0.3294 Pop_Nr          (Intercept) 0.2694
> 0.5191 Block_Nr        (Intercept) 0.1351   0.3675 Number of groups:
> Block_Nr:Pop_Nr 1435,  Fam_Nr 258,  Pop_Nr 90,  Block_Nr 16
> 
> No Coefficients
> 
> Threshold coefficients: Estimate Std. Error z value 0|1   1.9701
> 0.1221   16.14 1|2   3.7404     0.1488   25.13 2|3   4.5441
> 0.1791   25.37 3|4   5.3417     0.2322   23.00 (103 observations
> deleted due to missingness)
> 
> _______________________________________________ 
> R-sig-mixed-models at r-project.org mailing list 
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>