Statistics

The statistical support core can assist CROET scientists and other OHSU researchers with the analysis of existing data or the design of future experiments. Level of assistance ranges from basic consultation and guidance for researchers wanting to analyze their own data, up to a complete analysis that includes a summary of the statistical methods used (and references), the findings, tables and publication-ready figures. 

An initial meeting to discuss the aims of your project and your specific questions of interest can be arranged by contacting Michael Lasarev by e-mail (lasarevm@ohsu.edu) or phone (x4-3867). The initial meeting is free, though further services are charged on an hourly basis according to the fee schedule. You may be referred to another on-campus resource such as Biostatistics Shared Resource, depending on the scope of your project and your specific needs.

Researchers choosing to analyze and explore their own data may want to visit some of the sites shown on the right.

Software Resources

Data visualization

1. GNU Plot - A free high-level plotting program
2. Kyplot - A commercial plotting program with an Excel-like interface; also capable of some statistical analyses [Windows only]. The last free (beta) version of the program is available here.
   

Statistical

1. R - A powerful and free open-source program for statistical computations and visualization of data.
2. Gretl - A useful program for statistical models (regression, non-linear models, time series, panel data) that has point-and-click menus. To allow greater flexibility, the program can interface with R and also uses gnuplot as the graphics engine.
   

Power analysis

1. UnifyPow - Ralph O'Brien's SAS macro for performing power and sample size calculations; the main page has useful resources and documentation from short courses he's given on the subject.
2. Gpower - A free program for power and sample size determination for some common statistical tests.
3. R2 - A free DOS executable created by James Steiger to perform basic sample size and power calculations in regression problems.
4. STPLAN - A free DOS executable for study planning and sample size calculations for some very specific designs / tests.
5. Piface - Russell Lenth's page dedicated to power and sample size issues. The web site runs a java applet (piface.jar) that performs many power/sample size calculations; balance of page has useful information about power analysis in general.
   

Numerical computations

1. Octave - This is an open-source clone of Matlab©, a highly successful commercial program for numerical computations.
2. Scilab - Another open-source Matlab© clone.