[R-sig-ME] Announcement: afex 0.6-82 now on CRAN

Henrik Singmann henrik.singmann at psychologie.uni-freiburg.de
Thu Sep 26 22:01:18 CEST 2013


Dear list members,

With the release of R 3.0.2 and lme 1.0.4, I have pushed
a new version of afex (Analysis of Factorial EXperiments,
version 0.6-82) to CRAN which was accepted today and
should appear on your CRAN mirror within the next day.

It contains some new features that may be of interest:

- mixed() (for obtaining p-values for LMMs and GLMMs) now
   has method = "LRT" to obtain p-values from likelihood
   ratio tests (via the anova function). This method is
   computationally inexpensive (i.e, fast) and
   recommended by e.g., Barr et al. (2013, JML).

- fitting models in mixed() can now be parallelized using
   a cluster via package parallel (e.g., one can use
   multiple cores on one computer).

(more: http://cran.r-project.org/web/packages/afex/NEWS)

Note, mixed() now only works with the new version of
lme4 (>= 1.0.4). If there is interest I will provide a
version that is backwards compatible, so please mail me
if you need such a version.

To use method = "PB", for p-values based on parametric
bootstrap, you will need pbkrtest 0.3-6 which also was
just pushed to CRAN (thanks to Søren Højsgaard).

Best,
Henrik

-- 
Dipl. Psych. Henrik Singmann
PhD Student
Albert-Ludwigs-Universität Freiburg, Germany
http://www.psychologie.uni-freiburg.de/Members/singmann



More information about the R-sig-mixed-models mailing list