A specialized scatter plot primarily used in systematic reviews and meta-analyses to visually check for the existence of publication bias and other reporting errors.
- Horizontal Axis (X-axis): Displays the effect size of individual studies (e.g., Odds Ratio, Risk Ratio, Mean Difference).
- Vertical Axis (Y-axis): Displays a measure of study precision or size, most commonly the Standard Error (SE). The axis is typically inverted, meaning larger studies with higher precision and lower standard error sit at the top, while smaller studies with lower precision cluster near the bottom.
- Center Line: A vertical line represents the pooled or overall effect size calculated from the meta-analysis. [1, 2, 3, 4]
How to Interpret the Shape
The ideal layout forms an inverted, symmetrical pyramid or funnel around the overall pooled effect line. [1, 2]
| Plot Shape | Indication | Statistical Meaning |
|---|---|---|
| Symmetrical Funnel | No Publication Bias | Large studies (at the top) cluster tightly around the true population effect because they have less sampling error. Small studies (at the bottom) have higher variability and scatter widely on bothsides of the center line purely due to chance. |
| Asymmetrical Funnel | Potential Bias Present | A chunk or "bite" of the funnel is missing (usually at the bottom left or right). This suggests that smaller studies showing negative or statistically insignificant results were intentionally or unintentionally left unpublished. |
