pbkrtest - Test in linear mixed effects models based on parametric
bootstrap approaches and Kenward-Roger modification of F-tests
The pbkrtest package is an R package for tests in linear mixed effects models
based on parametric bootstrap approaches and Kenward-Roger
modification of F-tests.
- The reference to pbkrtest is: Halekoh, U., and Højsgaard,
S. (2014) A Kenward-Roger Approximation and Parametric Bootstrap
Methods for Tests in Linear Mixed Models - the R Package
pbkrtest. J. Stat. Soft. Vol. 59, Issue 9.
- Please do cite the above mentioned
paper if you publish work where pbkrtest has been used.
On the usage of the pbkrtest package
- This demo.R script produces this demo.html file.
1 Development versions of the packages
The pbkrtest package is hosted on github: here
2 Performance issues
Calculation of the the adjusted degrees of freedom for the
Kenward-Roger approximation can be computationally demanding because it
requires inversion of an N ×N matrix where N is the number
Possible remedies for this:
- On linux (ubuntu), changing the BLAS to ATLAS-BLAS
generally provides a speed-up with a factor 3-6 (compared with R's
default BLAS). That helps in some situations.
- Parametric bootstrap is an alternative, and while also
computationally intensive, parametric bootstrap can be parallelized
(facilities exist in pbkrtest).
- Lastly, an alternative to Kenward-Roger is to implement a
Satterthwaites approximation. It is on the todo-list.
3 Reporting unexpected behaviours (bugs)
When reporting unexpected behaviours, bugs etc. in the packages,
- A reproducible example in terms of a short code
- The data. The preferred way of sending the data "mydata" is to copy and paste the result from running
- The result of running the sessionInfo()
4 FAQ (frequently asked questions)
sorenh [at] math [dot] aau [dot] dk
- Do these methods work for generalized linear mixed models ?
- Parametric bootstrap is available for generalized linear mixed models.
We are not aware of any developments for approximate F-tests in the spirit of
Kenward-Roger for generalized linear models.
- Are these models implemented for mixed models fitted with the nlme
- No. We focus attention on the lme4 package.
File translated from