Consider a mixed model generated using the lme function in R. How can I consider the Box-cox transformations of this model in R? I have seen similar questions being asked before but they did not give specific references to the R code which is to be used.

Secondly, it is reasonable to consider the same linear model but without the random effects, and then use the transformation from this analysis in the mixed model? What would be explanation behind being able to/ not being able to do this? I would argue that on average the random factors are zero (hence this approach may work), but I do not have a definite answer.

Also, box cox should be considered before or after running significance tests for the different parameters in the model?

  • $\begingroup$ You should probably look at this: dx.doi.org/10.1111/j.1467-985X.2005.00391.x. $\endgroup$ Commented Feb 22, 2015 at 17:45
  • $\begingroup$ stats.stackexchange.com/questions/24115/… $\endgroup$ Commented Feb 22, 2015 at 17:56
  • $\begingroup$ I have had a look at the links but neither give refer to specific R code which I can use. Also, the first link states: "this work gives a clearwarning about the naïve implementation of a transformation to the linear mixed model wheninterest lies in the variance components and/or the random effects. Future work that focuses onimproving these estimators would be beneficial." @StéphaneLaurent @T C / anyone: do you have any suggestions on what would be the way forward? Can anyone suggest/reason why do the transformation on the linear model (without random effects) would/would not be possible? $\endgroup$
    – thankyou
    Commented Feb 22, 2015 at 19:04
  • $\begingroup$ If you seek only R code your question is likely to be off topic. $\endgroup$
    – Glen_b
    Commented Feb 23, 2015 at 2:12
  • $\begingroup$ The boxcoxmix package that provides Response Transformations for Random Effect and Variance Component Models, would answer your question. See cran.r-project.org/package=boxcoxmix . $\endgroup$
    – A M A
    Commented Oct 16, 2017 at 22:35


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.