What CMA costs
Comprehensive Meta-Analysis is a Windows desktop package, now sold as a one- or two-year subscription in tiered editions: academic licenses start around 295 USD per year as of 2026, with corporate rates higher. It has a long history, and for years it was the default answer for meta-analysis without code. The analysis it sells is now free in the browser.
The same analysis
- Fixed-effect and random-effects pooling, with DerSimonian-Laird and REML estimators for between-study variance
- Data entry as events and totals, means and SDs, or precomputed effects with a standard error or confidence interval (generic inverse variance)
- Odds ratio, risk ratio, hazard ratio, mean difference, and standardized mean difference
- Forest plots from a publication SVG renderer, funnel plots with Egger's and Begg's tests
- Cochran's Q, I², and tau² with every analysis
- Subgroup analysis, meta-regression, and leave-one-out sensitivity analysis
The engine is validated against R's metafor, every method shows its formula and citation, and you can download R replication code for your analysis.
What CMA does not have
Network meta-analysis: network diagram, league table, P-score ranking, and inconsistency checks, validated against R's netmeta. And the review around the numbers: literature screening with two blinded reviewers, risk of bias, GRADE, and a PRISMA flow diagram that counts itself from your decisions.
An honest note
CMA accepts an unusually long list of effect-size input formats, and if your studies report something exotic, its converters are genuinely useful. Covexe covers the common cases and includes an effect converter (OR, RR, RD, SMD, and standard error from a confidence interval) for much of the rest. If your data are events, means, or reported effects with a CI, nothing is missing.