causal

Small average effects can mask large opposite effects

A small average treatment effect can mask large effects in opposite directions for different subgroups—for example, private schools may show a small positive average test-score effect, but for kids with ADHD or on the autism spectrum public schools may be far better due to their legal responsibility to provide infrastructure, so slavish adherence to the average would badly mislead.

causalpending

Speaker

Emily Oster

Evidence Quote

“our average of small effects may be made up of large effects in both directions”

Source

Emily Oster on the Family Firm— EconTalk
Created: 6/13/2026, 12:27:29 AM

Your notes

Loading notes...