Evidence guide

What Is a Meta-analysis?

A meta-analysis combines numerical results from multiple studies to estimate an overall effect.

The promise

Pooling can improve precision and reveal patterns too small for individual studies.

The trap

A precise average can still be misleading when studies measure different things, contain bias or omit unpublished negative results.

Read beyond the diamond

Look at heterogeneity, study quality, publication-bias checks, sensitivity analyses and whether the review protocol was registered before results were known.

A practical reading checklist

  1. Was the search comprehensive?
  2. Were weak studies separated from strong ones?
  3. How different were the included studies?
  4. Did conclusions change in sensitivity analyses?