What Is Meta-Regression?

Looking for the full method? Read the detailed guide: What Is Subgroup Analysis?.

Meta-regression is a statistical method that relates the effect estimates from the studies in a meta-analysis to study-level characteristics, such as average participant age, dose or risk of bias, to explore whether they explain heterogeneity.

Example

Invented for teaching: In a meta-analysis of an exercise programme, meta-regression asks whether studies with longer programme duration show larger effects. The unit of analysis is the study, not the individual participant.

Common misunderstanding

Meta-regression findings are observational across studies, so they can be confounded and should be treated as hypothesis-generating. It needs enough studies (a common guide is at least ten per characteristic) and pre-specified characteristics, and it can mislead if many are tried.

Learn more

See What Is Subgroup Analysis? for related ideas.

This page is general educational information about research methods. It is not medical, legal or regulatory advice; follow your institution's ethics board and your target journal's instructions.