I am unaware if there is a good repository for psychometric software. These are free or freely-distributable titles that I have used or looked for. I have accumulated them here so as to be on-hand for myself and anyone else who can be helped. I am generally not involved with the packages, nor can I support them or necessarily get in touch with the authors.

• noharm - Colin Fraser's parameter estimation software based on MacDonald's multidimensional IRT work. Coming soon!
• equate 2.0 - Baker, Al-Karni, and Al-Dosarys' implementation of Lord-Stocking TCC equating--a standard tool in IRT DIF analysis (82.6k; zipped)
• simulate - A simple DOS program to generate dichotomous IRT data based on BILOG outputted item parameters. Coming soon!
• horn - A simple DOS program to help determine the statistical significance of successive eigenvalues after performing a principal components analysis. Based on Horn's parallel analysis. (15.6k; zipped)
• Lord_DIF - Computes Lord's Chi-square (widely used in DIF analysis). Coming soon!
• David Thissen's software page - Contains links to his commercial MultiLog program, a program for graphical item analysis, and a program for analysis of local item independence

I love using UNIX and Liunx makes it easy to have a real UNIX-like environment on my desktop computer for simulation and analysis. The only hitch is that actual comsumer programs for Linux for data analysis are somewhat rare (compared to, say, compilers or games or shells). This small set of links is the result of my investigation into options. Please let me know at amead@soltec.net if you write or find some cool new ap.

• R - S/S-Plus seems to be the hip new way to analyze data. R is a S-like language distributed freely. It is close enough to S that a large amount of the S code has been ported to R.
  http://www.ci.tuwien.ac.at/R/
• PSPP - A GNU-style competitor to SPSS the popular statistical package used by social scientists and, increasingly, business people for analysis of all sorts of data. I think this project is really exciting because SPSS is a major tool in my area of psychology. Having it available on Linux could mean a lot to struggling graduate students and researchers. Even more exciting, having it be GNU means that it can be improved upon in ways that SPSS, inc. seems reluctant to do. Unfortunately, the latest version as of this writing (early April, 99) is still quite empty.
  http://www.gnu.org/software/pspp/manual/pspp_toc.html

Visit the psychometric links page

Return to the main page