Forest plot
Shows estimates and confidence intervals from several studies or subgroups, with optional fixed- or random-effects pooled estimates (inverse variance; DerSimonian-Laird) and heterogeneity statistics.
Menu: Graphs → More charts → Forest plot…
When should I use it?
- Summarizing a meta-analysis or subgroup analysis.
What are the assumptions?
- Estimates with their CI limits or standard errors; ratios on a log scale.
How do I run it in StatLab?
- Choose the label, estimate and CI (or SE) variables.
- Choose a pooled estimate and tick ratio scale for OR / RR / HR.
How do I read the output?
| Output | Meaning |
|---|---|
| Squares | Study estimates (size ∝ weight). |
| Diamond | Pooled estimate and CI. |
| I² | Share of variation due to heterogeneity. |
What are common mistakes?
- Pooling on the ratio scale instead of the log scale.