Which model determines how much weight each article included in a meta-analysis will have in calculating the estimated overall effect?

• A. random effects model
• B. fixed effects model
• C. Bayesian model
• D. mixed effects model

Answer: (A)

In meta-analysis, how much each study contributes to the overall estimate comes from its weight. The model that directly determines these weights by incorporating variability between studies is the random effects model. It adds a between-study variance term (tau-squared) to each study’s variance, so the weight for a study becomes the inverse of the sum of its within-study variance and this between-study variance. This means that when there is heterogeneity among studies, the random effects model tends to balance the influence of studies more evenly, giving relatively more weight to smaller studies than the fixed effects approach would.

In contrast, the fixed effects model uses only the within-study variance to assign weights, which tends to amplify the influence of larger studies with smaller variances. The Bayesian approach introduces prior information and appears as a probabilistic framework rather than a straightforward, fixed weighting rule for combining study results. Mixed effects blends elements but is not the standard weighting rule used to determine each article’s contribution in the typical meta-analytic estimate.