Forest Plot Generator for Meta-Analysis
A forest plot generator turns your study effect sizes into a publication-ready figure with a pooled estimate, confidence intervals, and heterogeneity statistics. Enter your data, pick a model, and download a clean PNG or SVG in seconds.
The plot is free and yours to keep. If you are unsure which model to pool under, what your heterogeneity is telling you, or whether the data supports a meta-analysis at all, ask a PhD methodologist and get an answer in minutes. No account needed. How to read a forest plot
Build your forest plot
Live preview updates as you type. Switch between binary, continuous, and pre-computed effect sizes.
Move data between tools automatically. Compute effect sizes, then send results to Forest Plot, Funnel Plot, or Heterogeneity analysis with one click.
No data in pipeline yet. Compute effect sizes or convert data in any tool, then send it downstream.
Drag & drop a file or
CSV, TSV, Excel (.xlsx/.xls) - max 500 rows
Load sample data to see how the tool works, or clear all fields to start fresh.
Enter each study's effect size and confidence interval directly.
| Study | Effect | CI Lower | CI Upper | Subgroup | |
|---|---|---|---|---|---|
Built on the pooling reviewers expect
The numbers behind a forest plot matter more than the picture. This generator pools effect sizes with inverse-variance and Mantel-Haenszel weighting, estimates between-study variance with the DerSimonian-Laird method, and reports the full set of heterogeneity statistics a methods reviewer will look for. Ratio measures are pooled on the natural-log scale and back-transformed, exactly as RevMan and the metafor package do, so your results line up with what a statistician would reproduce.
- Fixed-effect and random-effects models with one click
- Cochran’s Q, I-squared, tau-squared, H-squared, and prediction intervals
- Continuity correction for zero cells and Hedges’ g small-sample correction
- Weight-scaled study squares and a pooled summary diamond
Correct statistics
Standard Cochrane and Borenstein formulas, validated against metafor and statsmodels.
Live preview
The plot redraws instantly as you edit studies, labels, and model options.
Journal-ready export
High-resolution PNG or scalable vector SVG, sized for manuscript figures.
Any effect measure
Odds ratios, risk ratios, mean differences, hazard ratios, and more.
How to make a forest plot
Building a forest plot takes five steps once your studies are selected and your effect sizes are ready. The preview above updates live, so you can follow along as you go.
- 1. Pick your effect measure. Decide whether your outcome is an odds ratio, risk ratio, mean difference, standardised mean difference, or a pre-computed value such as a hazard ratio. This sets the scale of the axis and the line of no effect.
- 2. Enter each study. Add one row per study with either raw counts and totals or a published estimate and its confidence interval. Give every row a clear label, usually the first author and year, so the plot reads cleanly.
- 3. Choose fixed-effect or random-effects. Use a random-effects model when your studies differ in population or design, which is the common case. The generator estimates the between-study variance for you.
- 4. Review the pooled diamond and statistics. The summary diamond shows the combined effect and its confidence interval, with Cochran's Q, I-squared, and tau-squared reported beside it. Confirm the result matches what you expect before exporting.
- 5. Export a publication-ready figure. Download a high-resolution PNG for a draft or a scalable vector SVG for final submission. The vector file prints sharply at any journal size.
For a fuller walkthrough with screenshots, read the guide on making a forest plot step by step.
Anatomy of a forest plot
Every forest plot follows the same layout, and learning to read it takes only a minute. Each row is one study. The square marks that study's effect estimate, and its size reflects the study's weight in the pool, so larger, more precise studies draw a bigger square. The horizontal line through the square is the confidence interval: a short line means a precise estimate, a long line means an uncertain one.
The vertical line of no effect sits at 1.0 for ratio measures such as odds ratios and risk ratios, and at 0 for differences. When a study's confidence interval crosses that line, its result is not statistically significant on its own. At the bottom, the summary diamond shows the pooled estimate from every study combined. The centre of the diamond is the pooled effect and its width is the pooled confidence interval. If the diamond sits clear of the line of no effect, the combined evidence points to a real effect even when individual studies were inconclusive.
The spread of the squares tells you about consistency. Squares clustered tightly together suggest the studies agree, while squares scattered across the axis signal heterogeneity, which is the moment to trust the random-effects diamond and its prediction interval over a single neat number. For a deeper walkthrough see the guide on interpreting each part of the plot.
Need the systematic review or meta-analysis behind these numbers?
A forest plot is the last step. If you want a PhD methodologist to run the search and screening, pool the data, choose the right model, assess risk of bias, or reviewer-proof your methods and results, we can take it from wherever you are now, across systematic reviews, meta-analyses, scoping reviews, and biostatistics.
Tell us about your reviewNo payment now and no obligation. You describe the project, a PhD methodologist replies with a tailored quote.
Convert and pool any effect measure
A forest plot is the last step of a longer chain, and the supporting calculators handle the steps before it. You can convert a two-by-two table into an odds ratio, work out the relative risk for the same data, or recover a hazard ratio and its standard error from a published survival result. For continuous outcomes you can standardise a mean difference into Cohen's d, and for clinical reporting you can find the number needed to treat or build a confidence interval from scratch. When the selection process needs documenting, you can also draw a PRISMA flow diagram. Browse the full suite of statistics calculators to see them together.
Frequently asked questions
What is a forest plot?
A forest plot is a graph that displays the effect estimate and confidence interval of each study in a meta-analysis, alongside a pooled summary estimate shown as a diamond. Each study appears as a square sized by its weight, with a horizontal line for its confidence interval, so readers can see at a glance how consistent the evidence is and where the combined effect falls relative to the line of no effect.
Is the forest plot generator free?
Yes, completely free. You can build a forest plot, run the pooled analysis, and download a high-resolution PNG or vector SVG at no cost. No watermark, no sign-up, no payment.
Which effect measures does it support?
You can pool odds ratios, risk ratios, and risk differences from binary outcomes; mean differences and standardised mean differences from continuous outcomes; and any pre-computed effect with a confidence interval, including hazard ratios and incidence rate ratios. Ratio measures are analysed on the log scale and back-transformed for display.
Does it calculate heterogeneity statistics?
Yes. Every analysis reports Cochran’s Q, the I-squared statistic, tau-squared, H-squared, and the p-value for heterogeneity. When you choose a random-effects model with three or more studies and real between-study variance, it also reports a 95 percent prediction interval.
What is the difference between fixed-effect and random-effects models?
A fixed-effect model assumes every study estimates one shared true effect, so differences between studies are due to chance alone. A random-effects model allows the true effect to vary between studies and uses the DerSimonian-Laird estimate of between-study variance. When heterogeneity is moderate or high, a random-effects model is usually the more defensible choice.
Can I use the figure in a published manuscript?
Yes. The export is a clean, high-resolution figure suitable for journal submission. The vector SVG can be scaled to any size without losing quality, and the pooling follows standard Cochrane and Borenstein methods that reviewers expect.