Randomized control trials solve the fundamental evaluation problem of selection bias: because participants (e.g., schools getting computers, women getting microcredit) are normally self-selected or chosen by motivated actors, observed outcomes conflate the intervention's effect with the pre-existing differences of those who opt in; randomization makes the treatment and control groups statistically identical so they can be validly compared.
causalpending
Speaker
Abhijit BanerjeeEvidence Quote
“what a randomized control trial does is it kind of solves that problem of infrared this it basically says you know this school and the other school... they're all all put into their names I put into a hat and we draw out 50 of the names at random”
Created: 6/13/2026, 7:00:20 PM
Your notes
Loading notes...