Shared Code for Statistical Programs
For comments, questions and suggestions about this page, contact Jesper Frickmann.
This is the place where you might find some useful program code to re-use for your tasks so you don't have to implement and test everything from scratch. If you have created some statistical programs you think have general interest for others, and you do not mind sharing them, then you can do so right here. Please contact Jesper Frickmann. The programs can be anything, as long as they have to do with statistics, and as long as they have some general relevance. The term "program code" shall be perceived in the widest sense, as anything from C/C++ or FORTRAN code over SAS and S-plus or R programs to spreadsheets.
All the code submitted here is provided "as is". ENBIS and the contributors do not assume any liability or guarantees for correctness of the code submitted here. Nor do we assume any liability for other consequences of downloading the programs, such as viruses.
This file implements the REGWQ procedure, which is probably the most powerful procedure for multiple comparison, controlling the experimentwise error rate. Besides, functions are available for computing Dixon's Q-values, which are also used in Tukey's procedure for multiple comparison ("Honestly Significant Difference", HSD). The implementation is done according to the description in the manual for the SAS/STAT system's GLM procedure, and results have been compared with output from this procedure, using the MEANS statement with the REGWQ option.
The Non-Central T-Distribution
Excel only comes with a function for calculating the central t-distribution. The non-central t-distribution is useful for, e.g., power calculations and operating characteristic curves.