Which statement best contrasts fixed-effects and random-effects models in meta-analysis and when each is appropriate?

Which statement best contrasts fixed-effects and random-effects models in meta-analysis and when each is appropriate?

• A. Fixed-effects assume variation across studies; random-effects assume a single true effect.
• B. Fixed-effects assume a single true effect; random-effects allow variation; use random-effects when heterogeneity is present.
• C. Fixed-effects are always better than random-effects.
• D. Random-effects cannot be used when there is heterogeneity.

Answer: (B)

In meta-analysis the key idea is how the model treats the true effect across studies. A fixed-effects approach assumes there is one true effect size shared by all studies, so any differences in study results come from sampling error. A random-effects approach, by contrast, allows the true effect to vary from study to study, reflecting real differences in populations, interventions, and settings, and it adds a between-study variance component to capture that heterogeneity. You’d choose a random-effects model when there is heterogeneity among study results and when you want to generalize beyond the exact studies included. If there’s little or no heterogeneity, a fixed-effects model can be appropriate, but it’s not inherently better—it’s about the assumptions you’re willing to make and the scope of inference. The statement that best fits this understanding is the one that says fixed-effects assume a single true effect, random-effects allow variation, and use random-effects when heterogeneity is present.

In meta-analysis the key idea is how the model treats the true effect across studies. A fixed-effects approach assumes there is one true effect size shared by all studies, so any differences in study results come from sampling error. A random-effects approach, by contrast, allows the true effect to vary from study to study, reflecting real differences in populations, interventions, and settings, and it adds a between-study variance component to capture that heterogeneity. You’d choose a random-effects model when there is heterogeneity among study results and when you want to generalize beyond the exact studies included. If there’s little or no heterogeneity, a fixed-effects model can be appropriate, but it’s not inherently better—it’s about the assumptions you’re willing to make and the scope of inference. The statement that best fits this understanding is the one that says fixed-effects assume a single true effect, random-effects allow variation, and use random-effects when heterogeneity is present.