It seems clear to me that dueling statisticians is just not a good way to handle such issues. Of course Harvard can recruit the most prestigious stat gun-slingers; doesn't make them right.
-
-
Show this threadThanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
There’s no consensus on what qualifies as fair. Part of what people want out of college admissions is general equilibrium changes, but few want to be explicit about their loss functions, so we get proxy arguments about fairness.
-
I say it would be more fair to pick the outcome measures we want from admitted students. Some of those are fully within Harvard's control, so they could less complain of being polluted by a biased outside society.
-
That doesn’t address how overall composition affects the outcome. There are standard cases for diversity of background and interests, and optimizing overall outcomes of that kind won’t work well with your proposal.
-
If one could specific how to measured the desired outcomes that one hopes diversity will produce, prediction markets could also be made on which policies would best produce those outcomes.
-
That’s true, but this is just a specific application of futarchy.
-
Indeed it is.
End of conversation
New conversation -
-
-
All of those strongly correlate to SAT scores - because all intelligence measures strongly correlate to each other.
-
Sure, but IQ isn't the only relevant signal about key outcomes.
End of conversation
New conversation -
-
-
I'm curious how you would implement that. What bets would people make and how would it translate into admission criteria? How would one bet on GPA or SAT scores?
-
Bets could be on the GPA of a student at the new school, conditional on their being admitted to and coming to that new school.
End of conversation
New conversation -
-
-
Prediction markets wld likely just reflect American biases. Eg over valuing extroversion, which is almost certainly one of the things hurting (E?) Asians trying to get into Harvard (or promoted in Silicon Valley?) (you urself alluded to being less likely to complain)
-
You can allow anyone in the world to trade in this markets. They've shown a great track record so far in resisting biases.
End of conversation
New conversation -
-
-
Transparency would reveal inconvenient things, like the fact that admissions departments consider FFA and 4H achievements to be a negative. Can't have people finding out which demographic is truly the most discriminated against by the Ivies...
-
Prediction markets would be transparent about the outcomes desired, but not necessarily about what is used to infer those outcomes.
End of conversation
New conversation -
-
-
I mean... if I were betting I would probably bet on the white students because of https://www.nber.org/papers/w9873 but that doesn’t really seem like a fair outcome either, especially because a Harvard degree is a large bestower of status. You can perpetuate the problem
-
That's re prediction future wages conditional on admission. But do you really think funny names hurt much re GPA or graduation rates?
-
Oh, sorry. I should have read closer. I think racial bias could play a role when student work is graded in a non-blind fashion, yes... maybe not as big a problem at Harvard but minorities more likely to have to work during college, not have $$ for random expenses
End of conversation
New conversation -
-
-
All of those strongly correlate to SAT scores - by design, it's what SAT scores were designed to do.
-
Sure, but even so, SAT scores don't embody all relevant available info re these outcomes.
-
Second most important measure is Trait Conscientiousness - and you'd be hard pressed to find a better objective measure than 'first-come, first-served'. Third most important measure is "# of parents who graduated from a similar institution" - ie Legacy.
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.