Meta-analysis Essentials

A forest plot is easy to produce and easy to misuse. This level builds the judgement behind meta-analysis: when pooling is valid, which effect measure to use, which model to choose, and how to read heterogeneity honestly, so your synthesis is defensible, not just tidy.

Who this level is for

Anyone who will run or interpret a meta-analysis: researchers, clinicians, analysts and evidence users. Recommended after Levels 1-3.

What you'll learn

When to meta-analyse, and when a narrative synthesis is the safer, more honest choice

Effect measures for binary outcomes (risk ratio, odds ratio, risk difference) and how to choose

Effect measures for continuous outcomes (mean difference vs standardized mean difference)

Fixed-effect vs random-effects models, and why random-effects usually suits global-health reviews

Heterogeneity in plain terms: Q, I-squared and tau-squared, interpreted qualitatively

How to read and explain a forest plot: study lines, weights, and the pooled diamond with its confidence interval

You finish with: Your synthesis plan: chosen effect measure and model, with a heterogeneity strategy.