Meta-Analysis Calculator with Forest Plot and Funnel Plot
Enter your studies and get the pooled effect, heterogeneity statistics, a publication-ready forest plot and a funnel plot in seconds. It handles continuous outcomes (Hedges’ g), binary outcomes (odds ratio or risk ratio) and any ready-made effect size.
How to use the meta-analysis calculator
- Choose your type of data: means and SDs for continuous outcomes, events and totals for yes/no outcomes, or a ready-made effect size with its 95% CI or SE.
- Pick the effect measure (for example Hedges’ g or risk ratio) and the model. Random effects is the usual choice when studies differ in population, dose or design.
- Type one study per row, or use Paste from Excel or Upload CSV / Excel to load your data extraction sheet and match its columns. Click Load example to see the format.
- Click Run meta-analysis. You get the pooled effect, 95% CI, I², τ², a prediction interval, a forest plot and a funnel plot with Egger’s test.
- Download the plots as PNG (high resolution) or SVG, and copy the ready-made results paragraph into your paper.
What this meta-analysis calculator does
A meta-analysis combines the results of several studies into one overall estimate, giving more weight to larger and more precise studies. This calculator uses the inverse-variance method, the same method used by RevMan, the R package metafor and most published reviews. It is great for learning, checking a supervisor’s numbers, quick scoping analyses and class assignments. For a full review you will usually move to R, and that is exactly what we teach in our course.
Methods and formulas
- Hedges’ g uses the pooled SD with the small-sample correction J, and its variance is 1/n₁ + 1/n₂ + g²/(2(n₁+n₂)).
- Odds and risk ratios are pooled on the log scale and shown back on the ratio scale. If a study has a zero cell, 0.5 is added to all four cells. Studies with no events in either arm are left out, as recommended by the Cochrane Handbook.
- Generic data: when you give a 95% CI, the SE is (upper − lower) / 3.92. For ratios, enter the values as they appear in the paper (for example OR 1.8, CI 1.2 to 2.7). The tool takes the logarithm for you.
How to interpret heterogeneity
| I² | Cochrane rough guide | What to do |
|---|---|---|
| 0–40% | Might not be important | Fixed and random results will be similar. |
| 30–60% | May be moderate | Use random effects and explore reasons (subgroups). |
| 50–90% | May be substantial | Report the prediction interval; run subgroup or sensitivity analyses. |
| 75–100% | Considerable | Ask whether pooling makes sense at all. A narrative synthesis may be better. |
The example binary dataset is the classic BCG vaccine trial dataset (Colditz et al., 1994). The calculator reproduces the DerSimonian–Laird results from R’s metafor: pooled RR ≈ 0.49, τ² ≈ 0.31, I² ≈ 92%.
Frequently asked questions
Should I use a fixed-effect or random-effects model?
Use a fixed-effect model only if you believe all studies estimate exactly the same true effect, for example identical trials. In most reviews, populations, doses and methods differ, so a random-effects model is the safer default. This calculator shows both so you can compare them.
What is a good I² value?
There is no ‘good’ value, because I² describes how much of the variation between studies is real rather than due to chance. Below 40% is usually low and above 75% is considerable. Always report it alongside τ² and the prediction interval, and look at the forest plot.
How many studies do I need for a meta-analysis?
Technically two, but random-effects estimates of τ² are very unreliable with fewer than about five studies, and funnel plots and Egger’s test need at least ten.
Is this the same as RevMan or R metafor?
Yes for the standard inverse-variance fixed-effect and DerSimonian–Laird random-effects models; results match to rounding. Other estimators (REML, Hartung-Knapp adjustment, Mantel-Haenszel) are not included. Use R for those.
Can I paste my data from Excel or upload a file?
Yes. Click Paste from Excel and paste your rows, or Upload CSV / Excel to open a .csv or .xlsx extraction sheet. You then confirm which column holds the study name, means, SDs, events and so on. The file is read in your browser and is never uploaded or stored.
Is my data uploaded anywhere?
No. The entire meta-analysis runs in your browser, so unpublished data stays private.
