# Idea A random-effects meta-analysis assumes the existence of a distribution of true effects applicable to a set of different studies and populations. Deviations of individual studies from the center of such distribution represent true heterogeneity—a degree of between-study variability beyond what is expected to occur by chance. As with [[fixed-effects meta-analysis]], sampling error still contributes to explain deviations between study-specific estimates and the assumed "true" effect for each particular study. Study weights will need to consider both sources of variance, and the meta-analytic pooled estimate can only be regarded as the mean of a distribution of effects, and not as a true effect for any real population. [[Knapp-Hartung adjustments]] can be applied to random-effects models. # References - [How to Perform a Meta-Regression | Columbia Public Health](https://www.publichealth.columbia.edu/research/population-health-methods/meta-regression)