Theory 01/Refinement
Aggregates Are Smaller
The claim
Pooled effect estimates are smaller than the studies that dominate a literature, because quality-weighted pooling downweights small selective samples.
01Mechanism
Selection on significance inflates small studies; pooling by precision corrects it.
02Prediction
Pooled estimates should fall as study quality rises.
03Refutation
Pooled estimates tracking the largest studies would remove the correction.