TimelyScholarFree Tools › Funnel Plot Generator
Check your meta-analysis for publication bias and small-study effects. Draw a contour-enhanced funnel plot, run Egger's regression test and see how trim-and-fill changes the pooled estimate.
FreeNo account neededYour data stays in your browser
| Study | Odds ratio | Lower 95% CI | Upper 95% CI | |
|---|---|---|---|---|
| Pooled odds ratio | 1.24 (95% CI 1.13 to 1.36) |
|---|---|
| Egger's test intercept | 0.905 (95% CI 0.138 to 1.671)Regression of the standard normal deviate on precision |
| Egger's test | t(35) = 2.40, p = 0.022Evidence of small-study effects (funnel asymmetry) |
| Trim-and-fill: studies imputed | 7 on the left sideL0 estimator (Duval & Tweedie) |
| Adjusted odds ratio | 1.19 (95% CI 1.08 to 1.31)A sensitivity analysis, not a corrected result |
Odds, risk or hazard ratio, mean difference or SMD.
Estimates with 95% CIs, or 2×2 counts.
Egger's test, filled studies and the adjusted estimate.
PNG or SVG, ready for your manuscript.
A funnel plot shows each study's effect against its standard error. Without bias, small studies scatter widely at the bottom and large studies cluster near the pooled effect at the top, forming a symmetrical inverted funnel.
Asymmetry, usually a gap where small studies with null or unfavourable results should be, suggests small-study effects. Publication bias is one cause, but heterogeneity, poor methodological quality in small studies and chance can also produce it. Contour-enhanced plots help tell these apart: if the missing studies would fall in areas of non-significance, publication bias is more likely.
Tests for funnel asymmetry have low power, so they are usually only used when there are at least 10 studies.
A statistically significant intercept (often judged at p < 0.10) suggests funnel plot asymmetry, meaning smaller studies report different effects from larger ones. It does not prove publication bias.
At least 10 is the usual minimum. With fewer studies, the plot and the tests are too imprecise to distinguish chance from real asymmetry.
No. Trim-and-fill is a sensitivity analysis. Report the original pooled estimate and describe how much trim-and-fill changes it.
It shades areas of statistical significance. If studies seem to be missing from non-significant areas, publication bias is a plausible cause of the asymmetry.