What's new

Dan Kahneman - Gates Foundation bad statistics research


New Member
Hi All,

While studying the linear regression chapter I remembered something I read on Dan Kahneman's book "Thinking Fast and Slow". He mentioned a research funded by Gates Foundation with ~1bi USD that investigated the best schools in the US and it concluded that the best ones were the smaller ones, but this was a mistake made by a heuristic way of thinking. The worst schools were also the smaller ones. The problem is that is easy to think in a causal relation to explain it and that small schools are more variable.

Somehow i think this relates to regression, but i cannot exactly say how. Maybe they used an linear regression and it produced an estimate something like the example of Stock and Watson test score example and tried to attribute a causal relationship to it. What do you think about it?

You can find a part of this reading in the last section of the following article: https://theweek.com/articles/478388/dangers-quick-thinking