Hacker News Reader

The author gets at the ecological fallacy [0] several times - the idea that group characteristics are not always useful for prediction of individual characteristics.

I wonder how much of the heat in this conversation generally comes from the assumption that group level tendencies are prescriptive to individuals, though in reality individuals can display a high level of variance within the group.

[0] https://en.m.wikipedia.org/wiki/Ecological_fallacy

View on HN

Is that really even a fallacy? In medicine, they infer your risk for certain diseases based on age, race, gender, lifestyle habits etc. Individual risk varies of course, but that doesn't mean that generalization is wholesale a bankrupt concept. He never says group traits == individual traits, he states 15x "on average." Yes, there's Simpson's paradox as described in your citation, but that doesn't mean averages, sigmas are useless as a whole. They have to be applied carefully, which Kaufman addresses.

They generalize in that case because there usually isn't a way to know your individual risk factor. I don't really see how that same principle applies to personality, which is obviously much more visible.

Reply on news.ycombinator.com