Sub-analysis testing whether a pooled effect varies by a study-level factor.
In meta-analysis, moderator analysis examines whether the pooled effect size differs according to a characteristic such as programme duration, sex, or recovery period, typically via meta-regression or subgroup comparison. These comparisons are usually based on between-study (ecological) correlations rather than within-study randomisation, so they are hypothesis-generating and vulnerable to confounding rather than definitive evidence of effect modification.
This guide was auto-drafted and is pending editorial review.