Felix Schönbrodt

Dr. Dipl.-Psych.

Installation of WRS package (Wilcox’ Robust Statistics)

Update Feb 17, 2014: WRS moved to Github – This installation procedure has been updated and still is valid

Some users had trouble installing the WRS package from R-Forge. Here’s a method that should work automatically and fail-safe:

# first: install dependent packages
install.packages(c("MASS", "akima", "robustbase"))

# second: install suggested packages
install.packages(c("cobs", "robust", "mgcv", "scatterplot3d", "quantreg", "rrcov", "lars", "pwr", "trimcluster", "parallel", "mc2d", "psych", "Rfit"))

# third: install an additional package which provides some C functions
install.packages("devtools")
library("devtools")
install_github("WRScpp", username="mrxiaohe")

# fourth: install WRS
install_github("WRS", username="nicebread", subdir="pkg")

WRS cannot be hosted on CRAN, as CRAN demands help files for every user-visible function. This has not been done for WRS (yet). For the time being, this somewhat more complicated installation routine has to be used.

Comments (10) | Trackback

10 Responses to “Installation of WRS package (Wilcox’ Robust Statistics)”

  1. Peter Ellis says:

    Hi – doesn’t seem to work at the moment (July 2013). Has it been locked off?

  2. FelixS says:

    Yup, R-Forge is down until 12 July 2013: http://r-forge.r-project.org/forum/forum.php?forum_id=4407
    If you can’t wait, download the functions from http://dornsife.usc.edu/labs/rwilcox/software/

  3. Heidi says:

    Hi Felix, just to say many thanks for the helpful code.

  4. […] [Update Aug 2013: For a fail safe installation routine, see this up-to-date post: Installation of the WRS package] […]

  5. […] For a fail-safe installation of the package, follow this instruction. […]

  6. FelixS says:

    Update Sep 19 (v. 0.23.2): Due to an error in the NAMESPACE file the functions were not exported. Works with 0.23.2.

  7. […] from the WRS package compares user-defined quantiles of both distributions using a Harrell–Davis estimator in conjunction with a percentile bootstrap. The method seems to improve over other methods: “Currently, when there are tied values, no other method has been found that performs reasonably well. Even with no tied values, method HD can provide a substantial gain in power when q ≤ .25 or q ≥ .75 compared to other techniques that have been proposed”. The method is described in the paper “Comparing two independent groups via the upper and lower quantiles” by Wilcox, Erceg-Hurn, Clark and Carlson (2013). You can use the function as soon as you install the latest version of the WRS package following this installation instruction. […]

  8. […] Update: For installation of the WRS package, please use only the fail-safe installation procedure described here! […]

  9. Shu Fai says:

    Thanks for the instruction. Just a simple question. I noticed that there is also a project called WRScppWin. If I run R for Windows, should I install that package instead of WRScpp?

Leave a Reply