[R-sig-ME] Residuals plot for a linear mixed model

[ONKELINX, Thierry]

Dear Manuel,

The points on a line are due to zero's in your dataset. The distance between this line and the other residuals depends on the number that you add to the respons prior to taking the log. Increasing it will make the difference smaller.

A log transformation with zero's is something that I would try to avoid. Maybe a sqrt() trasnformation will do. Or glmm is the respons can be interpreted as some kind of counts. Another option would be to use a hurdle model to find the zero's and the non-zero's seperatly.

Do you have an idea why the Swiftness is either zero or a quite larger number?

Best regards,

Thierry

----------------------------------------------------------------------------
ir. Thierry Onkelinx
Instituut voor natuur- en bosonderzoek
team Biometrie & Kwaliteitszorg
Gaverstraat 4
9500 Geraardsbergen
Belgium

Research Institute for Nature and Forest
team Biometrics & Quality Assurance
Gaverstraat 4
9500 Geraardsbergen
Belgium

tel. + 32 54/436 185
Thierry.Onkelinx at inbo.be
www.inbo.be

To call in the statistician after the experiment is done may be no more than asking him to perform a post-mortem examination: he may be able to say what the experiment died of.
~ Sir Ronald Aylmer Fisher

The plural of anecdote is not data.
~ Roger Brinner

The combination of some data and an aching desire for an answer does not ensure that a reasonable answer can be extracted from a given body of data.
~ John Tukey
  

> -----Oorspronkelijk bericht-----
> Van: r-sig-mixed-models-bounces at r-project.org 
> [mailto:r-sig-mixed-models-bounces at r-project.org] Namens 
> Manuel Spínola
> Verzonden: maandag 21 maart 2011 14:15
> Aan: r-sig-mixed-models at r-project.org
> Onderwerp: [R-sig-ME] Residuals plot for a linear mixed model
> 
> Dear list members,
> 
> I got the attached residuals plot for linear mixed model 
> after logtransformationn of the response variable.
> Is the plot telling me that I have a problem with the model 
> fitted?  If logtransformation didn't help, what are thealternatives?
> 
>  > mode = lmer(log(Swiftness.1 + 0.05) ~ Lure + Sex + 
> Facility.Size + (1|Subject), REML = F, data = otter)  > 
> summary(mode) Linear mixed model fit by maximum likelihood
> Formula: log(Swiftness.1 + 0.05) ~ Lure + Sex + Facility.Size + (1 |
> Subject)
>     Data: otter
>   AIC   BIC logLik deviance REMLdev
>   546 574.9   -262      524   511.5
> Random effects:
>   Groups   Name        Variance Std.Dev.
>   Subject  (Intercept) 0.75907  0.87125
>   Residual             9.33305  3.05500
> Number of obs: 102, groups: Subject, 17
> 
> Fixed effects:
>                     Estimate Std. Error t value
> (Intercept)          1.1968     1.0832   1.105
> Lure[T.2]           -0.4375     1.0479  -0.418
> Lure[T.3]            0.9928     1.0479   0.947
> Lure[T.4]            0.4381     1.0479   0.418
> Lure[T.5]           -0.3059     1.0479  -0.292
> Lure[T.6]           -1.1340     1.0479  -1.082
> Sex[T.1]            -0.2301     0.7411  -0.311
> Facility.Size[T.2]   1.9357     0.9550   2.027
> Facility.Size[T.3]   1.2889     0.9820   1.312
> 
> Correlation of Fixed Effects:
>              (Intr) L[T.2] L[T.3] L[T.4] L[T.5] L[T.6] S[T.1] F.S[T.2
> Lure[T.2]   -0.484
> Lure[T.3]   -0.484  0.500
> Lure[T.4]   -0.484  0.500  0.500
> Lure[T.5]   -0.484  0.500  0.500  0.500
> Lure[T.6]   -0.484  0.500  0.500  0.500  0.500
> Sex[T.1]    -0.342  0.000  0.000  0.000  0.000  0.000
> Fclt.S[T.2] -0.540  0.000  0.000  0.000  0.000  0.000 -0.055 
> Fclt.S[T.3] -0.544  0.000  0.000  0.000  0.000  0.000  0.000  0.617
> 
> predicted<-fitted(mode)
> raw.residuals<-residuals(mode)
> plot(predicted,raw.residuals)
> abline(h=c(-1.96*sd(raw.residuals),0,1.96*sd(raw.residuals)))
> 
> 
> Best,
> 
> Manuel
> 
> 
> 
> --
> *Manuel Spínola, Ph.D.*
> Instituto Internacional en Conservación y Manejo de Vida 
> Silvestre Universidad Nacional Apartado 1350-3000 Heredia 
> COSTA RICA mspinola at una.ac.cr mspinola10 at gmail.com
> Teléfono: (506) 2277-3598
> Fax: (506) 2237-7036
> Personal website: Lobito de río
> <https://sites.google.com/site/lobitoderio/>
> Institutional website: ICOMVIS <http://www.icomvis.una.ac.cr/>
>