What is meta-analysis and when would you use fixed vs random effects models?

• A. Fixed effects assume a common true effect across studies.
• B. Fixed effects assume a common true effect; random effects allow study-specific effects and heterogeneity.
• C. Both models assume identical effects and no between-study variability.
• D. Random effects assume no heterogeneity.

Answer: (C)

Meta-analysis combines results from multiple studies to estimate an overall effect, while recognizing that study results may differ. A fixed effects model assumes there is a single true effect shared by all studies and that any differences observed are due to sampling error within studies. This approach is most appropriate when the studies are very similar in design, populations, and interventions, so you’re estimating that common effect for this specific set of studies.

A random effects model, on the other hand, allows the true effect to vary across studies. It includes a between-study variance component to account for heterogeneity, yielding wider confidence intervals that reflect uncertainty about the distribution of effects in the broader universe of possible studies. Use this when there is noticeable heterogeneity among studies or when you want to generalize beyond the exact studies included.

The statement that both models assume identical effects and no between-study variability isn’t accurate, because that second model explicitly accounts for between-study differences. If there’s heterogeneity, a random effects approach is typically more appropriate.