A random effects analysis makes the assumption that individual studies are estimating different effects. We assume that the different effects have a distribution with some mean value and some degree of variability dictated by the between-study variance. The idea behind a random effects meta-analysis is to learn about this distribution of effects across different studies.
