[R-pkgs] First release of package "cg" for comparison of groups

Bill Pikounis billpikounis at gmail.com
Mon Dec 27 14:32:01 CET 2010

Version 0.9.0, a first release of cg, is published on CRAN and its
mirrors. The "cg" name stands for "compare groups." Its genesis and
evolution are driven by common "in-practice" needs to compare samples,
treatments, administrations, conditions, etc. in medicine research &
development. The current version provides comparisons of unpaired
samples, i.e. a linear model with one factor of at least two levels.
Good data graphs, modern statistical methods, and useful displays of
results are emphasized.

The cg package is developed with the overall goal of a comprehensive
analysis of data when comparison of groups is a primary interest. A
flow of wrapper functions are contained within it to guide the full
analysis and interpretation of the data. The wrapper functions
encompass functions in base R and other packages. Some features

* Housekeeping of log-scale analysis to express multiplicative effects;

* Use of resistant & robust models to accommodate potential outliers;

* Use of accelerated failure time models to handle limits-of-detection values.

Once the package is loaded, the call

> example(cg)

illustrates the package with two included data sets.

A slide deck presentation of the package can be found at
http://user2010.org/slides/Pikounis+Oleynick.pdf . A manuscript on the
package has just recently been submitted to the Journal of Statistical
Software and is under early review.

Bug reports and questions are welcome and can be directed to at "cg at

Thanks to all of the R Development Core Team for base R and its
packages, all authors and maintainers of the packages Hmisc, grid,
lattice, MASS, survival, multcomp, mvtnorm, nlme and VGAM on which cg
depends, and all of the R community for inspiration.

Bill Pikounis and John Oleynick

More information about the R-packages mailing list