Q from
@deankarlan
: do people make too much about what nudges can do?
Kahneman: everybody is excited about being right on the direction of the hypothesis, but ignore the magnitude of the effect, which is often quite small, or comes with very high variance
Your final judgement shouldn't be given to a human it should be the algorithm, except in very rare circumstances. When people overrule the algorithm, they're usually wrong.
@poverty_action
@deankarlan
Yes. A perpetual disappointment in public policy research is the small size of many significant +ve effects. We need to keep reminding policy makers of this awkward 'it works!' factoid.
《Muted Eureka!》Archimedes lost the top few inches of his bathwater, not the whole lot.
@poverty_action
@deankarlan
In line with Kahneman's thoughts, I propose a new approach to assess the economic importance of regressors in a recent CSAE working paper