Meta-Regression
Also known as: Meta-Analytic Regression, Weighted Regression in Meta-Analysis, Moderator Analysis, Meta-regresyon
Meta-regression is a statistical technique that extends conventional meta-analysis by regressing study-level effect sizes on one or more study characteristics (moderators) to explain between-study heterogeneity. Formalized by Thompson and Higgins in 2002, it uses weighted least squares — weighting each study by the inverse of its variance — within a mixed-effects framework, allowing researchers to identify which study features systematically account for variation in observed effects across the literature.
Key highlights
- Quantifies how much of between-study heterogeneity is explained by specific study characteristics
- Accommodates multiple moderators simultaneously in a single model
- Weights studies by precision, making full use of available statistical information
- Provides an R-squared analog that communicates explanatory power in an interpretable metric
Intuition
This section is available to Pro members. Upgrade to Pro
How it works
This section is available to Pro members. Upgrade to Pro
When to use it
Meta-regression is appropriate when a meta-analysis reveals significant heterogeneity (I-squared > 25–50%) and the analyst has a priori hypotheses about study-level moderators that may explain it. It requires a sufficient number of studies — commonly at least ten studies per moderator to avoid overfitting. Assumptions include linearity of moderator effects and availability of moderator data for most included studies. When moderators are individual-level rather than study-level, individual participant data (IPD) meta-analysis is preferable. Subgroup analysis is a simpler alternative when moderators are categorical and studies are few.
Strengths & limitations
- Quantifies how much of between-study heterogeneity is explained by specific study characteristics
- Accommodates multiple moderators simultaneously in a single model
- Weights studies by precision, making full use of available statistical information
- Provides an R-squared analog that communicates explanatory power in an interpretable metric
- Requires a large number of studies (typically 10+ per moderator) to yield reliable estimates
- Ecological fallacy risk: study-level associations may not reflect individual-level relationships
- Results are sensitive to the method chosen for estimating residual between-study variance
- Cannot establish causality; confounding among study characteristics is common
Common pitfalls
This section is available to Pro members. Upgrade to Pro
Applications
This section is available to Pro members. Upgrade to Pro
Frequently asked
How many studies do I need to run a meta-regression?
There is no absolute minimum, but the widely cited rule of thumb is at least ten studies per moderator variable. With fewer studies the model is prone to overfitting, standard errors are underestimated, and results are unlikely to replicate. When the number of available studies is small, subgroup analysis or simple heterogeneity exploration is more appropriate than a full regression model.
What is the difference between fixed-effect and mixed-effects meta-regression?
Fixed-effect meta-regression assumes that after accounting for moderators all residual heterogeneity is zero — that is, only sampling error remains. Mixed-effects meta-regression adds a random-effects term (tau-squared) to allow for unexplained between-study variability beyond what the moderators capture. Mixed-effects models are generally preferred because the assumption of zero residual heterogeneity is rarely tenable in practice.
Can meta-regression prove that a moderator causes differences in effect sizes?
No. Meta-regression is observational at the study level and cannot establish causation. Moderator variables are often correlated with each other — for example, more recent studies may also use larger samples — making it difficult to isolate the unique contribution of any single characteristic. Findings should be interpreted as exploratory or hypothesis-generating unless supported by strong prior theory and pre-registration.
Sources
- 1.Thompson, S. G., & Higgins, J. P. T. (2002). How should meta-regression analyses be undertaken and interpreted? Statistics in Medicine, 21(11), 1559–1573.
You have read it. What now?
Cite this page
ScholarGate. (2026, June 2). Meta-Regression. ScholarGate. https://scholargate.app/meta-analysis/meta-regression