- DAG-Stat:
**D**iagnostic &**AG**reement**Stat**istics. This is a freely available Excel workbook, written by Andrew MacKinnon. It provides a comprehensive range of statistics calculable from 2x2 tables that are useful in evaluating diagnostic tests and interrater agreement. - Online exact tests for 2x2 contingency tables
- Online calcuators for contingency tables, crosstabs, and chi-square tests.
- kappa.exe - a program for calculating Cohen's kappa and weighted kappa for 2 raters. (It is stored as a zip-file because Angelfire/Lycos blocks download of .exe files.)
- John Uebersax's Statistical Methods for Rater Agreement. This site is a resource for the analysis of agreement among raters, diagnostic tests, observers, judges or experts. It contains background discussion on different methods, examples, references, software, and information on recent methodological developments.

- Systematic reviews from astronomy to zoology: myths and misconceptions. Article by Mark Pettigrew in BMJ 2001;322:98-101.
- Larry C. Lyons' meta-analysis page. Includes links to other sites.
- The Meta Analysis of Research Studies. Includes the following sections: Overviews, online reports, software, bibliographies, and links to other sites.
- Meta-analysis
programs by Ralf Schwarzer. Check out the manual. Includes the following:
- the analysis of exact probabilities (p values) as suggested by Rosenthal
- the analysis of effect sizes d, following the methods of either Glass or of Hedges and Olkin
- the analysis of effect sizes r as recommended by Hunter, Schmidt, and Jackson

- EasyMA: Free software for meta-analysis of 2x2 tables, developed in the Department of Clinical Pharmacology of the cardiological hospital in Lyon, France. Note that I have not been able to run this DOS-based software on some newer and faster machines--but if it will run on your machine, it's pretty good. Also, the manual is very good.
**My own programs**. I have written 3 programs that implement the procedures described by Fleiss (1993). Zip-files are available for download.- MACONT.EXE is for continuous (or interval scale) variables, and requires one to enter (or read from a text file) a study number, N, Mean, and SD for both the Treatment and Control groups for each trial or study.
- MA2x2.EXE is essentially the same, but is for use with binary outcomes. Instead of entering means and standard deviations, however, you can simply enter the 4 cell counts.
- FLEISS.EXE is a general purpose program that requires you to enter a study number, total N, some measure of effect size (e.g., a logit) and its variance for each study. It requires a bit more data preparation than the others, but can be used for any measure of effect size. An SPSS syntax file version is also available.

*Last modified on 9-Mar-2003.*