David B. Wilson’s Meta-Analysis Page

Updated 12 July 2021

This website contains information about tools I’ve created for conducting a meta-analysis, including an effect size calculator and analysis macros for SPSS and Stata. I also provide links to web resources related to meta-analysis. Finally, I provide a link to my CV and Google Scholar Profile that list my publications.

Effect Size Calculator

I’ve created a web-based calculator for four standard effect size types: Cohen’s d, the odds ratio, the risk ratio, and the correlation coefficient (r). Several methods are available for calculating each effect size type, particularly Cohen’s d. This calculator is housed on the Campbell Collaboration website at the following link.

Documentation for the equations used by the calculator are available in this PDF.

SPSS Macro for Performing Meta-analysis

The macro MetaAnal.sps (link) is a complete re-write of the prior three meta-analysis macros I had written: MeanES.sps, MetaF.sps, and MetaReg.sps. Detailed information on installing and using the new macro is provided. This macro performs inverse variance weighted meta-analysis of any user-specified effect size type under either a fixed-effect (common-effect) model or a random-effects model. Eight estimators for the random effects variance component are available. This macro can also run both categorical and regression-based moderator analyses.

You can download the original macros at this link for archival purposes: archived SPSS macros.

Stata Macro for Performing Meta-analysis

As of Version 16.0, Stata has built-in meta-analysis capabilities, as summarized here: link. For those still using prior versions of Stata, a completely re-written version of my macros meanes, metaf, and metareg, are available from the ssc package repository and can be installed via the following command:

    ssc install Meta_Analysis

This command will install three programs, masum, maanova, and mareg. The names were changed from the original macros because these are not backward compatible in terms of syntax. As with the SPSS macros, eight different estimators for the random effects variance component are available. You can run both categorical and regression-based moderator analyses using maanova and mareg, respectively. Documentation is available through help pages within Stata after installation.

    help masum
    help maanova
    help mareg

You can download the original Stata macros here.

CV, Google Scholar, and Such