4 ms·

> Your comment makes no sense to me. That's because you didn't go to Harvard (/s) and you need to learn yourself what Selection Bias is. If Harvard selected i

> Your comment makes no sense to me. That's because you didn't go to Harvard (/s) and you need to learn yourself what Selection Bias is. If Harvard selected its students randomly from the general population of all 18-year-olds (including those who dropped out of high school), then it would be statistically relevant that Harvard grads make 299% more than the national median. But this is not the case. Harvard chooses its students from the brightest, highest achieving, most highly motivated kids in the nation (and whose families tend to be wealthy.) It is reasonable to think that had Harvard vanished and these kids never got a college education, they would still succeed greatly in life. So to mitigate the selection bias, we might compare Harvard grads versus the population of kids were in the top 5% of their high school class (but even that's not perfect.) The measured boost of the "Harvard Super Drug" would be considerably lower than 3x--maybe 10% if that.

cs702 3y ago
FYI, robocat added his note about selection bias as an edit, AFTER I had posted my response. With the new edits, robocat's comment now make sense to me. That said, there's still a big caveat: We should look at the change in income for students drawn from the low-income population. And please follow the guidelines: There's no need to lob insults.
- dudeinjapan 3y ago
  > We should look at the change in income for students drawn from the low-income population You'd really have to compare outcomes among a population of low-income students who equally "could have gone to Harvard" but some of them didn't for whatever reason. And even here there's a hazard of selection bias--as one example, a "reason" for not going to Harvard might be teenage pregnancy, which is certainly going to put a dent in economic outcome. If you compare low income students in the top 5% of their HS class who went to state school vs. Harvard I'll bet the gap is smaller still. For example, here's a list of where Fortune 500 CEOs went to college--it's mostly state schools. Warren Buffett went to U. of Nebraska, Tim Cook to Auburn, Larry Page to U. Michigan, etc. https://twitter.com/daveckang/status/1667946905966637061 https://twitter.com/daveckang/status/1667946905966637061
- robocat 3y ago
  I had to add Selection bias afterwards, because I had thought it was bleeding obvious but user li_am made a snarky remark so I made the comment clearer.: I just don’t understand how you and li_am could totally miss the point, The 98% pass rate shows it so clearly. Note: I am not from the USA but I know this shit.

cs702 3y ago

FYI, robocat added his note about selection bias as an edit, AFTER I had posted my response. With the new edits, robocat's comment now make sense to me. That said, there's still a big caveat: We should look at the change in income for students drawn from the low-income population. And please follow the guidelines: There's no need to lob insults.

dudeinjapan 3y ago
> We should look at the change in income for students drawn from the low-income population You'd really have to compare outcomes among a population of low-income students who equally "could have gone to Harvard" but some of them didn't for whatever reason. And even here there's a hazard of selection bias--as one example, a "reason" for not going to Harvard might be teenage pregnancy, which is certainly going to put a dent in economic outcome. If you compare low income students in the top 5% of their HS class who went to state school vs. Harvard I'll bet the gap is smaller still. For example, here's a list of where Fortune 500 CEOs went to college--it's mostly state schools. Warren Buffett went to U. of Nebraska, Tim Cook to Auburn, Larry Page to U. Michigan, etc. https://twitter.com/daveckang/status/1667946905966637061 https://twitter.com/daveckang/status/1667946905966637061
robocat 3y ago
I had to add Selection bias afterwards, because I had thought it was bleeding obvious but user li_am made a snarky remark so I made the comment clearer.: I just don’t understand how you and li_am could totally miss the point, The 98% pass rate shows it so clearly. Note: I am not from the USA but I know this shit.