Wednesday assorted links

1. Cryptocurrency pump-and-dump schemes.

2. No Coase theorem: “Opinion polls show more than 50 per cent of Greeks oppose the June deal because the government controversially accepted the existence of Macedonian ethnicity and language in return for the change of name.”  FT link.

3. Chocolate is older than we had thought.

4. The world’s Scrabble champion in French cannot speak French (NYT).

5. “A Russian scientist working in Antarctica is facing attempted murder charges after allegedly stabbing a colleague for telling him the endings of books he wanted to read.

6. David Card vs. Peter S. Arcidiacono.  Good piece, Card is wrong.

7. MRU Halloween video.

Comments

No 5. must have been a case of cabin fever.

Due to the absence of sun in the poles six months of the year, it's a wonder more of this doesn't happen. I read that paranoia is not unknown to the permanent staff that lives there during the dark half of the year (I've read the scientists don't stay when it's dark, just the low-level support staff and technicians).

6. But what if an applicant is an Girl Named Florida!

" “A Russian scientist working in Antarctica is facing attempted murder charges after allegedly stabbing a colleague for telling him the endings of books he wanted to read.”"

So, it's justifiable homicide then.... ;)

Clearly the wrong one of the two men is on trial.

Yes, the only regret is that the murder was only attempted and not completed. The aggrieved scientist should be jailed for incompetence.

#4 is probably true for many languages. Even many top scrabble players in English cannot speak English.

Hasn't it been the case the spelling bee champion us not a native English speaker? Maybe I'm imagining it...

"6. David Card vs. Peter S. Arcidiacono. Good piece, Card is wrong."

"David Card .... used words like “nonsensical” to describe another economist’s analysis."

I strongly doubt that the other economists work was "nonsensical". So, this strike me as theater more than analysis.

"But Harvard had only about 1,700 spots to offer. ... Harvard admissions officials have vigorously denied that they discriminate against Asian-American applicants. An applicant’s race, they said, can help, but not hurt, his or her chances of admission."

This stance does strike me as nonsensical. There are a fixed number of slots and it's a zero sum game. So, If you help one group, you hurt another group.

So, you having a job that pays harms everyone else?

Wrong use of "zero sum". Zero sum apllied tol economics says costs-benefits=0. So, you can pay one person $50,000 and then sell $50,000 in production to one person, or pay one million $50,000 each and sell $50 billion in production to one million customers.

After all economists look at markets, and Harvard is only one of thousands of producers.

Apple picks who gets an iPhone, which means at time you will get the iPhone you want and hundreds of people won't until China's production and Trump's trade policy allows enough iPhones to be available for sale.

For example, early on, you had to buy from AT&T, a version of affirmative action, to get an iPhone. AT&T was being discrimated against with Verizon favored.

wowow this is so stupid it makes my head hurt

maybe that's the flu

zero sum:

> in which each participant's gain or loss of utility is exactly balanced by the losses or gains of the utility of the other participants.

There are only 1,700 spots. It's zero sum.

Suppose there are two schools of equal size, A and B, and equal amounts of two types of students, S and T. Also assume A has nicer facilities and S students have better applications.

If S students benefit from having some interaction with T students, the admissions team at school A might choose to reject some of the group S applicants in favor of the best applications from group T. This allows A to matriculate the applicants from S with the best applications and the applicants from group T with the best applications.

If A accepted all applicants from group S, they do not have control over which accepted applicants will choose a diverse experience at B. It may be the S students with the best applications choose B (to signal a lack of need of the nice facilities).

The applicants from group S with the best applications may benefit from A's decision to accept the best applications from group T over the worst applications from group S.

Right, actually Harvard is doing everyone they reject a favor so the court should award them a bonus instead of punishment

> If S students benefit from having some interaction with T students,

From the data from CollegeData.com, the cognitive range of the Harvard offered applicants was up to about 4 SD IQ i.e. 60 IQ points. The range could be much more since SAT and GRE can track IQ up to only 140. There are many reports that people with 1 SD difference have problem communicating with one another, what can T students learn from S students and vice versa? Irrespective of the course level, Glenn Loury said that he preferred smart PhD class cognitive difference of less than 2 SD i.e. IQ 2 SD to 4 SD. With greater than 4 SD the European PhD solely by research might be better. For undergrad that might be from 1 SD and up. Graduate with less than 1 SD most probably ended up under-employed in jobs that do not require any degrees, e.g. Criminal Justice major, 75% under-employed, most probably working as bouncers or low level security guards.

http://i63.tinypic.com/jqhvls.png

Look at the level of "special invite" and offers, compare that to the Harvard SAT 50 percentile which is much higher than 2 SD (IQlike=130). On average, Harvard students at or above the 50 percentile (half the class) might not be able to communicate effectively with about 40% of the remainders.

T students may be:

1) Wealthy and therefore able to fund the ideas of the S students
2) From powerful/connected families and able to support the S students with their connections
3) Have specialized knowledge of about a region/culture/religion
4) Able to bring the school prestige through other means such as athletics or community service

The lowest SAT percentile students are often enrolled in different classes than the highest students. The lower SAT percentile students have non-classroom contributions to campus life/student experience.

If you were a billionaire, would you prefer to donate a building (buy your child's admission) to a school that accepted students using SAT scores alone? What admission process would you want to see?

Jokes aside, it is like the Florida Girl problem. Arcidiacono looks at 95% of applicants who take 71% of slots. Card looks at 100/100%.

I'm out of my depth, but I'd kinda think 100/100 tells the full story of admissions.

Right, because that 29% represents the legacy admissions, people on a "special list" (donors and vip's), athletic recruitments, etc. Why would they not count?

We know why they don't count for Arcidiacono's side: because it would show that the "special list" program is affirmative action for whites, and that truly every race is getting a special bonus except asians, who are hurt by all of the above. And though "special list" affirmative action isn't a constitutional issue for the courts, the optics would look pretty bad, and who knows maybe it causes Harvard to cut both types of affirmative action if plaintiffs win. Then the whites funding this case have only ended up helping the asian pawns they're using to kill race-based affirmative action. Just a thought. But I'm all for it.

Seems strange to call that affirmative action for whites. It's only for a very special subset of whites. I certainly didn't benefit from the existence of any special list like that when I was applying to colleges, and I'm all kinds of white.

It's affirmative action for those on the "special list", which doesn't include probably >99.99% of whites. Let's just call it affirmative action for rich people.

These groups could be predominantly white, but not exclusively, though, right? It's donors [ok, maybe most likely to be white of this subgroup, and also arguably maybe the least deserving group to be given an advantage], athletes, and family of alumni and faculty. Children of alumni would reflect past admission trends, including any historical bias depending on how far back their admission practices go, and for the latter the Harvard Gazette states [https://news.harvard.edu/gazette/story/2018/04/harvard-reports-sustained-progress-on-faculty-diversity/]:

"Harvard’s faculty is more diverse than ever, with women making up 30 percent of tenured and tenure-track faculty and minorities making up 23 percent."

So, yeah, majority white.

...why would you discard evidence of anti-asian affirmative action in order to help make your case that asians are discriminated against?

Because meeting the burden of proof requires comparison of similarly situated applicants. Athletes, legacies and donors are not similarly situated. After meeting a minimum academic requirement, they are guaranteed admission. They are never ranked along with other applicants according to selection factors.

You are making the same sophistic argument as Card that the judge apparently bought. While Legacies and Donors might frustrate an overall diversity goal, this data isnt at all useful in comparing similars for the disparate treatment analysis.

Blacks could argue that Legacy and Donor admissions discriminate against them, but that is a separate (and failing) lawsuit. Legacy and donor admissions are a racially neutral policy with a nondiscriminatory purpose. The fact that Harvard has openly friendly policies toward minorities would defeat such a claim.

I don't see that the article gives enough data to support that claim.

The 95% of applicants produce 71% of freshmen, and Asians comprise 23% of the whole class.

If you assume Asians came in with the 95%, that selection process is doing heavy lifting, and not limiting.

There are other articles about this case.

Asians score pathetically low on the "personality" factor which is wholly subjective and most often without any evidentiary basis.

The key to unravelling a lie is to first understand the motives of the liar. Harvard has an expressed desire to admit more disadvantaged minorities. It isn't going out of it's way to admit Mormons for the sake of "diversity." Diversity is and always has been code for racial quotas.

As one person suggested, what is the percentage of high achieving blacks denied admission? The answer is most likely close to zero percent.

By the way, this discrimination also enables them to maintain a leftist ideological tilt.

Not true, they actually do like to bring in 'hicks from the sticks' like Mormons as part of the diversity goal. My wife's cousin, white as they come, got a full ride to Stanford for being basically lower middle class and semi-rural. Basically almost 'white trash'. He had good grades and played water polo but was by no means a star genius.

97% of Stanford admissions have an SAT of 1430 or higher & 99% had a 3.25 GPA or better. Have you asked him what is SAT scores were? My guess is he's a lot brighter than you are giving him credit for.

No, because those slots given to donors, athletes and legacies are essentially non discretionary. They have a minimum academic constraint to meet but are otherwise guaranteed admission. Grades are part of the eligibility criteria but not ranking criteria. It is thus proper to exclude them from the analysis of how ranking criteria are adversely applied.

The judge may be correct that the preponderance of whites among legacies and donors would hurt a diversity goal, but that is not justification for outright discrimination. Blacks, Hispanics and Asians are statistically underrepresented among legacies and donors, but they certainly aren't barred from this benefit because of race. Apparently rich Blacks, Hispanics and Asians need to donate more to Harvard. As Bill Cosby said, "How big a hospital do you need?"

I suspect that Arcidiacono looked at the other 95%, because that's the group in question.

There is no question about academic discrimination among the 5%. That is the group that is openly discriminatory.

"legacy admissions, people on a "special list" (donors and vip's), athletic recruitments, etc"

"That is the group that is openly discriminatory."

Exactly, but Harvard makes no guarantee that they'll be 30% legacy and 70% Stanford style merit.

They are using their own weird method which is .. 100% discriminatory.

They must, to hide the legacies in the population, and ensure that they aren't the bottom scorers in each class.

"They are using their own weird method which is .. 100% discriminatory."

Yes. Which makes their claims of not discriminating against Asians particularly laughable. It's an open secret. Everyone knows they do it.

To be fair to Harvard, they probably wouldn't deny it, if they didn't know the result would be a huge lawsuit payout. Sometimes you can't afford to be honest.

Harvard accepts 23% Asians, George Mason 19%, the UC system 33%.

What's the natural level? UC is higher, but it's also drawing from a more Asian state (California 14.9% vs the national 5.6%)

("Stanford" above was a goof)

"What's the natural level?"

Why would that possibly matter? By those grounds Harvard discriminates against both whites and hispanics.

Most people assume that academic institutions discriminate along academic grounds. Harvard doesn't like that outcome, so they tinker with the metrics to get the desired output.

What Card's argument is really showing, I bet, is that there is a scoring function that realizes virtually any set of racial profiles.

I wonder if that is an argument he really wants to make.

#1. This is my shocked face.

😱

#1 Never before have so many been bamboozled out of so much by so few...wait...tulips anyone?

#2 It is well known that in ancient history the Greeks viewed the Macedonians as "bumpkins, barbarians, uncivilized" even after getting repeatedly trounced by Alexander's father Philip. It looks as though old attitudes die hard. Stubborn Attachments

#5 Apparently they were both described as extremely avid reader. If I was isolated for years (yes. apparently they had been posted for years) with very intermittent supplies in subzero temps cooped up with someone vindictive enough to spoil at every opportunity the thing I love most, yes, I'm going to shank them.

Modern Macedonia is not the ancient Kingdom of Macedon, which in fact was located in modern northern Greece. Modern Macedonia appropriated its name. Obnoxious.

6. Since Asian-Americans have the strongest academic ratings, any admission criteria that includes factors other than academic ratings necessarily discriminates against Asian-Americans. Harvard’s admissions staff rates applicants in four categories: academic achievement, extracurricular activities, personal qualities, and athletic abilities. Moreover, Arcidiacono excludes recruited athletes, the children of alumni, the children of Harvard faculty and staff members, and students on a special list that includes children of donors. That further exacerbates the discrimination against applicants with the strongest academic ratings; athletes, the children of alumni, the children of Harvard faculty and staff members, and students on a special list that includes children of donors aren't likely to be Asian-American. Cowen's comment suggests that he would prefer an admissions criteria that has one factor: academic ratings. I know many very smart people I wouldn't want to be in a position of responsibility or authority. Does Cowen?

Arcidiacono excludes a large amount of data (those with special relationships with Harvard and make up about 5 percent of applicants and 29 percent of those accepted) that are probably more likely to be white. Card utilizes all the data. Here's what I don't get. By including a group that is more likely to be white and more likely to accepted, Card finds no discrimination. By excluding this subset of the data, Arcidiacono finds there is discrimination. Given the skewed nature of this subset of data, it seems that including it would be detrimental to Card's case and beneficial to Arcidiacono. What gives?

Well, is MLB biased because the Red Sox won the world series and the Yankees didn't? As a kid in Indiana, I remember the teachers who put the game on each year when the Yankees won yet again, and no one else seemed allowed to win.

The Yankees kept winning and thus had the money and status to get the players and managers to keep winning. Then the sports industry adopted affirmative action to discriminate against the Yankees. MLB, including the Yankees ended up better off. As an older kid, who didn't like baseball, I thought the Yankees not winning was unfair. Thinking as an economist, the Yankees always willing was bad for the economy. Thus affirmative action is good for the economy.

But what was good for the economy, was costly to fans who saw their costs go up year after year. Ie, higher prices for everything. If the Yankees were just celebrating their 50th win, baseball at the local sandlot would be more popular, tickets cheaper, more seats, less TV, players paid very little.

I think Card includes the legacies and athletes and concludes that (much of) the difference in admission rates between whites and Asians is due to the fact that whites make up a large fraction of athletes and legacies, not discrimination. Arcidiacono excludes the legacies and athletes to conclude that there is still residual discrimination against Asians after excluding legacies and athletes.

That's why Card is wrong. Legacy and athletic factors are legal admissions criteria even if they have a disparate impact on admissions rates. He includes these factors to make it more difficult to identify other factors that also contribute to differences in admissions rates (among non-athletes and non-legacies), other factors like whether an applicant is white or Asian that would not be legal. He's deliberately keeping the confounding factors of legacies and athletics so that he can conclude that the difference in admissions rates is (mostly) due to those confounders. That makes no sense if one can just exclude the confounders from the outset.

Thanks. I have successfully outsourced my thinking.

The only way Card's approach could be semi-defensible would be if the admissions office were for some reason favoring Asian legacies and athletes over white legacies and athletes. Suppose, for example, that the admissions office boosted the "friendliness" scores of Asian athletes vs. white athletes but marked down the friendliness scores of Asian non-athletes vs. white non-athletes. Then, Card could find that discrimination against white athletes is offset by discrimination against Asian non-athletes. Arcidiacono would find just the discrimination against Asian non-athletes. Even in this case, though, the solution would be to stop manipulating friendliness scores of white and Asian athletes and non-athletes. Racial balancing among non-athletes does not suddenly become legal simply because one is also trying to achieve racial balancing among athletes.

"That's why Card is wrong. Legacy and athletic factors are legal admissions criteria even if they have a disparate impact on admissions rates."

This is why I don't really get Card's analysis: the case is about whether race is being directly used as a target, not about whether certain criteria have a racially disparate impact. For instance, my university (not Harvard but another elite private school) had and may still have a program to make sure every state in the country had a reasonable amount of representation, rather than simply the elite public and private schools of the Northeast and Los Angeles. Obviously, some of these states (like North Dakota, where one of my friends was from) are whiter than the national average, so it would be reasonable to conclude that a state origin program could produce a disparate impact in terms of race. Now, you could "correct" for this disparate impact by *explicitly considering race* of applicants from the big applicant-producing schools in the Northeast and Los Angeles, such that no white or Asian person from those schools who is not on the "special list" has a chance. But you can't be explicit about it, so then you're back in the area where you come up with a criterion (personality) as a cover for race, which is what the suit is about. Card is trying to argue that all factors other than SAT/GPAs are suspect if they produce racially disparate outcomes, regardless of their intent, which is a different issue than what the suit is about ("personality" is a proxy for setting racial targets for black and Hispanic students). I can't really see how recruiting elite rowers for the crew team is discriminatory just because white kids tend to be the people who do crew.

No, you are confusing attainment with some unstated, overall diversity goal with the disparate treatment/impact analysis.

The Legacies and Donors cannot be included in the pool of analysis because they are not similarly situated with the Plaintiffs. Harvard has a legitimate, nondiscriminatory reason for admitting the children of legacies and donors - as a reward and incentive for attending and donating. If wealthy Blacks (and there are lots of them) wanted to buy their kids' way into Harvard, nothing is stopping them provided their kids meet Harvard's minimal academic standards.

The 800 lbs gorilla in the room is that there is a paucity of high achieving Black students, and there are not enough for every top 30 ranked school to have 13% Black students and have them pass their classes.

Right, Harvard isn't racist, they just accurately assess Asians to have very poor personal qualities

Black people are better at being black than Asian people are at being Asian.

20 Koreans and 20 Irish are on standby for a flight to Seoul, which has 10 empty seats. 6 Koreans and 4 Irish get a seat.

Also 20 Irish are on standby for a flight to Dublin, which has 10 empty seats. 10 lucky Irish get a seat.

The Irish admission rate is 35% while the Korean is only 30%. Is this racism?

I guess it depends whether anyone is discriminating on the basis of race

Do the gate agents give personal ratings for being a desirable passenger that consistently favors one group?

If Asians are not satisfied, maybe they should go back home. There is no reason for giving into the demands from an ungrateful minority.

"That further exacerbates the discrimination against applicants with the strongest academic ratings; athletes, the children of alumni, the children of Harvard faculty and staff members, and students on a special list that includes children of donors aren't likely to be Asian-American. "

Yes rayward. And logically this means that Ardiacono understates the discrimination.

Of course. And if Ardiacono had made his statistics with 100% of applicants, Card would have said the opposite at sure: "he doesn't make any sense to count together the special admissions on sports or heritage with the bur ok normal admissions, etc." Actually, I didn't find anything interesting or stated in good faith in anything Card is reported to have said.

Harvard wants to admit applicants that will do well at Harvard and do well in life, the combination of which will benefit Harvard via alumni contributions. Thus, it wouldn't surprise me (should it surprise an economist) that Harvard would design an admissions criteria that would accomplish those goals. Now, the designers of the admissions criteria might be idiots, but I suspect not: the Harvard endowment has $36.5 billion, by far the largest of any college. Doesn't economics matter at this blog?

Then allow them to explicitly discriminate against Asians, and remove tax exempt status.

But if Google can get sued over not having an exact demographic match to the US of their software engineers, then we can sure as hell tear Harvard apart for having a racial quota.

You’re a liberal. Racial quotas should outrage your sense of justice.

Tax exempt universities should not allow legacy or donor consideration. And I say this as someone whose children will be double legacies. Ivies sure as hell don’t need more money.

Every application should be name and demographic blind. Or, make an explicit racial floor for underrepresented minorities and be done with it.

They did this with Jews. Now they’re doing it with Asians. It was despicable then and it’s despicable now. Especially in the aftermath of antisemitic violence, we should strive to overcome this blatant discrimination.

If liberals get up in arms about Ellen Pao, they should get up in arms about yellow quotas.

"Tax exempt universities should not allow legacy or donor consideration"

I used to disagree with this (it's a private school and all), but am beginning to agree that it's extraordinarily unseemly for massively endowed bastions of privilege like Harvard and Yale to reap the benefits of tax-exempt status, as though they are providing a social service, and then give highly coveted spots virtually guaranteeing an extraordinarily privileged life to the children of donors, famous and powerful people, and legacies because they want to make sure the donations keep coming in and boost their marketing prestige by claiming "the sons and daughters of Presidents go here". I am not sure it should be illegal, and it definitely should not be illegal on the grounds that it produces some variety of racially disparate impact, but it is very unseemly.

1. Of course, any market is potentially subject to the risk of pump-and-dump schemes. Why is the crypto market especially subject to such schemes? I'd start by identifying the major participants in crypto markets. Who might they be?

#5 “Some reports *suggest* that alcohol was involved” .... hee hee hee

"Of the 37,000 applicants for admission to the Class of 2019, for instance, 8,200 had perfect grade-point averages, and more than 2,700 had perfect scores on the verbal section of the SAT."

American examinations are far too easy. That's hardly Harvard's fault. But, on the other hand, what does Caltech do about it?

No, the standardized tests sufficiently separate students by ability. Grades may be inflated at some schools but they possess informative value. Having a "harder" test would run students through an additional sieve, but that is overkill in determining likelihood to succeed. Probably close to 100% of students with perfect GPA and Verbal SAT will succeed at Harvard.

Ironically, the lack of discernment at the top hands Harvard and other schools a tool for discrimination. If tests were harder as you suggest, then admissions would more closely approach pure meritocracy and diversity would plummet. As it is, every Black valedictorian in the nation has guaranteed admission to Harvard. Not true for Asian or White valedictorians.

6. Harvard argues "Of the 37,000 applicants for admission to the Class of 2019, for instance, 8,200 had perfect grade-point averages, and more than 2,700 had perfect scores on the verbal section of the SAT. But Harvard had only about 1,700 spots to offer. Even if the university wished to consider only grades and test scores, it would be hard-pressed to select a freshman class using those variables alone." I wonder how many applicant's of their favored racial categories with perfect academic scores are rejected.

That is a very good question, and would probably serve as prima fascia for discrimination. In other words, I posit the number with perfect scores or close not admitted is statistically insignificant, and likely due to other factors like criminal record etc.

So it would seem that a perfect GPA and perfect verbal SAT should be the minimum necessary but still not sufficient admission criteria for all applicants.

Since they don’t mention perfect math SAT scores, one presumes ranking applicants by that additional criteria would allow the class to be filled.

Bonus - you can get rid of all those tiresome essays, and replace the admission staff slots with something useful.

Excellent question, and likely close to zero.

"Of the 26,000 American applicants to the Class of 2019, roughly 3,500 had perfect SAT math scores, 2,700 had perfect SAT verbal scores, and nearly 1,000 earned a perfect overall score on the SAT or ACT. Typically, admitted Harvard classes number around 1,700 students."

In other words, merely by using both SAT components, the number of applicants with perfect scores is far fewer than the number of available slots. Strange coincidence that Harvard would like to pretend otherwise.

Yeah, that is a shocking and bizarre statement. Excluding math scores to try to make a point that is the opposite of reality.

From Cal tech admissions: "We require all applicants take the SAT or ACT, both of which we will superscore. We also require 2 SAT Subject Exams, the Math Level II and your choice of at least one of the Science Subject Exams (physics, chemistry, or biology)."
Superscore means to take the highest score in each category from multiple tests.

6. "An applicant’s race, they [Harvard admission officials] said, can help, but not hurt, his or her chances of admission." Amusing.

I think that the larger issue, beyond races, is the unwillingness of the
top-tier universities to give more importance to the academic ability of their students. Sure, they say it is because they have no tool to measure academic ability among their applicants, since too many get perfect scores in the SAT and GPA is too much school-dependent. Maybe, but then they should create a much harder test (Harvard can do that alone, or with a consortium of 10 or 50 universities which could use it), where almost no one would get the maximum grade. This lack of interest for academic quality of admitted students is cause of a clear academic mediocrity in the American system.

Harvard, doesn't really want the absolute best academic students. They want to be able to discriminate and not be sued over it.

As I said to another commenter, the "easy" test allows for discrimination in favor of minorities. A test that more adequately and finely ranked students for pure meritocracy, would cut the balls off diversity programs.

Moscow University used to do this in the opposite. They would give so-called "coffin tests" or "Jewish tests" that no one could pass. The tests had solutions that were easily verifiable but nearly impossible to solve (P#NP). Giving the tests only to Jews enabled them to discriminate and have plausible deniability- the Jews failed. But even if they gave the same test to everyone, everyone would fail and they could then use more subjective and discriminatory criteria.

The key to getting away with discrimination when people are looking for it is to obscure the difference between the victims and their similarly situated comparators or to come up with an ironclad but pretextual nondiscriminatory selection factor.

They certainly could and arguably should. They may be too sclerotic as an organization to make such a jump. Too many of their subunits focused on exploitation to be able to muster the buy in for an exploration?

Why should analytical ability be the sole determinant of admission? Is analytic ability the most important predictor of who produces the best literature? Expands philosophy? Does the best work in archeology? Why do you naively assume that all slots of the university are fungible among students with the same level of analytical ability.

We as a society seem comfortable giving companies autonomy in deciding hiring decisions. Perhaps a coding test is sufficient for a dev, but for most positions there isn't a simple test one can take. We don't demand that the Navy only admit recruits into the Seals based on physical fitness. Would it be reasonable to build a team composed entirely of snipers, if they happened to be the fittest? We don't select our mates solely based on looks, intelligence or wealth, even though that might simplify the search. Even for graduate school, in many fields letters of recommendation matter much more than tests and sometimes even GPA. So why should we expect undergraduate education be any different?

Why can't we let the universities decide which class composition is best for them, and if you don't like it, feel free to lower their status as much as you are able? Seems like a better alternative than a quantitative admission process simply to elevate the status of "merit", as if any of us really knows what that is.

I haven't mentioned "analytical ability", just harder tests, where there would not be a large bulk of students having perfect scores. I haven't said what these tests should be, and there are many ways to imagine them.

As I see them, first they should not be multiple choice tests, but real essays where you need to write long texts (several pages) in answer to each question (of course there would not be 100 questions).
There should be such tests in many fields of academic interest, math, sciences, literature, history, perhaps one foreign or ancient language,
perhaps philosophy, etc. And finally, again, the tests should be hard -- it should be almost impossible for even the best students to answer every questions correctly in the given time. With this the universities would have a good way to judge the academic abilities of their applicant. After that, if they want to temperate this with other non-academic criteria, they are free to do so, but right now the position at Harvard is "we can't use academic criteria because we have no way to measure them", which is quite hypocritical.

Now of course the universities decide what is good for them (whatever that may mean). I don't really care anymore about changing them. Trying to lower their status, semi-anonymously because I work for one of them, is what I am dong here.

I came in the US because I was attracted, first by the country as a whole, and second by the prestige (and riches, in particular the immense libraries) of the US universities. Now I have realized that the prestige just hide a terrible mediocrity, and that no one here was really interested in elevating the academic level (as I thought was the mission I was hired for), just defending advantages for their pet groups (rich kids of rich alumni, or transgender asian people, or whatever). I will just certainly not send my kids in any American university, even though my benefits include much reduced tuition.

Where would you send them?

In Europe: France or Switzerland, UK maybe.

> but right now the position at Harvard is "we can't use academic criteria because we have no way to measure them", which is quite hypocritical.

Is that really Harvard's position? Or Card's? I'm not sure.

Your proposal only solves the issue of providing a rank to every applicant at the top of the distribution. It does not justify using that rank for admission, however. Sure, every department of every university might have it's own, un-aceable test, but is that even remotely feasible? No. Tests for a particular discipline shared by multiple universities might work for physics and math (these already exist by the way, perhaps they are not hard enough), but these tests will likely elicit complaints from different schools of thought. But none of this answers why we should tests to assign rank for admissions versus using test to impose a floor on performance. The latter is probably more appropriate outside of the hard sciences.

I'll add that any test where minorities can't achieve highly would become immediately suspect of discrimination, whether that is true or not. The current standardized tests are already accused of that. Greater dispersion of scores would only increase that sentiment.

As BC and anonymous point out, there's no requirement to use analytical ability as the only criterion for admission. Unitversities (or universities if you want to be pedantic) can use as admissions criteria how fast you can run 100 meters, how well you can play the clarinet, how many tennis ball you can juggle successfully for 30 seconds, how well you play basketball or hockey, how much money your family has or is expected to donate, how tall you are. What you can't use as a criterion is race, per the 1964 civil rights act.

Playing basketball?
How well you can rap?
Knowing the proper cooking time for chitterlings?

Dont laugh at these because they have actually been proposed to increase minority enrollment.

What rational basis do any of these criteria have at 1) determining the probability of success or 2) bringing diverse ideas to the school?

Diversity has always been an equivocation. In one sense, Harvard vaunts the value of different ideas and experiences. But what it actually means by diversity is skin color.

5. It's highly likely that the books in question were fiction.

From the chocolate story:

"It's another gift of the people of Amazonia to the world," said Prof Blake. "It highlights the importance of protecting this habitat."

This is a most tendentious understanding of the concept of “gift”. A gift must be given to a recipient with the express intention of benefiting them in some manner. This is more like the function of an invisible hand, an unintentional consequence.

The bloody red hand of Hernán Cortés, more like.

#2 - the Greeks reject the Greek government Macedonia deal. - as a Greek citizen, I could care less. I recall that on a jar of Nutella type macadamia butter in Greece, back in the 1990s, there was a slogan, in Greek, "Macedonia is not a country" which I thought should have read "macadamia is not a nut" (crazy thought I know).

As for the referendum, they are non-binding. Twice the Greeks have voted with large majorities to leave the EU, and twice their elected leaders have defied them and remained. The word on the street, when I was there, is that the GR pols have been paid off by the German and French bank interests to remain in the EU. I wouldn't be surprised.

#5 I knew somebody who used to work for the British Antarctic Survey and he said that being cooped up with one or a few other people for long periods was a real problem. In theory Navies should have at least an institutional memory of this problem. The very interesting book "Royal Navy Way of Leadership" encourages qualities such as cheerfulness in difficult circumstances.

Which is why navies had strict discipline and harsh punishment. And grog. And raping and pillaging of natives. And plundering of prizes.

Naval life was so harsh, the British used impressment.

5 - talk about information hazard!

#6 Does not Harvard and every undergraduate program have quotas for intended majors? Even if a major is not specified, many times it's easy to infer from the application. It seems college Asian enrollment is strongly correlated with the number of STEM graduates they produce. Caltech the highest, then MIT, then all the Ivies. Not to say Asians or whites averages aren't higher, but without taking major into account, admitting simply based on the strength of the application is unlikely to match well with the courses offered or the professors.

So we should treat a 40 billion dollar tax exempt entity as a jobs program for Proust scholars?

For the life of me I don’t understand why the left defends yellow quotas. The left is supposed to be the bleeding heart that points out injustice and demands reform. We need a healthy left wing to balance out the right wing.

Instead the left acquiesces to a hidden racial quota. Why?

What madness 2018 is.

The obvious answer is that whites don't find it as easy to look down on asians, so doing them favors doesn't satiate their need to signal virtue.

The guy who filed this suit on behalf of Asians is White. He's definitely virtue signaling.

No, not really. It is legal tactics.

While Harvard's system discriminates against Whites, it does so more virulently against Asians. The case is stronger together. The evidence more weighty for Asians. Whites get to tag along.

There is also an argument that Asians are a more sympathetic plaintiff because they are historically oppressed in the US.

You don't? Here's a hint: the so called left is not caring, it is codependant. This is a drama triangle.

Perhaps "discrimination" is inevitable in a matching problem where the distribution of applicants does not match the distribution of slots available? If you want to study engineering, apply to Caltech not Harvard. There are more slots available to study that there.

So if a “holistic” assessment of applicant quality and fit is OK for Harvard, how about for selecting applicants for the Fire and Police academy?

When have I ever claimed a holistic admission process?? For a thread filled with so many commenters on statistical aberrations in admissions, it's astonishing none of you have even thought to apply Simpson's paradox here. Asian outperformance yet under-admittance is textbook Simpson's paradox. There is nothing discriminatory aboug universities not wanting their STEM lectures to be packed and their professors to be swamped, while literature classes sit empty. Asian admittance rates across the US correlate very strongly to the proportion of STEM graduates at the university (controlling for the percentage of Asians in the state population).

If you have any evidence that Asians are admitted at the same rate as equally qualified whites, blacks and hispanics for STEM majors, please let us know

7 was fun, but I'm not sure what the last costume was. A crash-test dummy, but what is the economic concept? I could only guess "dummy variable".

6. Of course Card is wrong. He is wrong about everything for which his liberal confirmation bias desires a certain outcome.

The "personal characteristics" rating was heavily in favor of Blacks and Hispanics and blatantly discriminatory against Asians. It is so subjective, it could not possibly be replicated. And in most cases the personality rater never met the applicant.

And we all know that Blacks herd themselves into athletics.

Two selection criteria were rigged in favor of choosing Black applicants to the detriment of everyone else, and especially Asians.

UC Berkeley, which is forbidden by law from having quotas, has only 3% Black students and 40% Asians. Nearby Stanford without such restrictions has 13% Blacks. Notwithstanding the fact that Stanford is a much higher ranked school that Black valedictorians are more likely to choose, it is clear that legal restrictions on reverse discrimination lower Berkeley's diversity numbers. By the same token, we can see that private schools like Harvard pack their classes with minorities for feigned diversity. Of course, there simply aren't enough high achieving Blacks to go to every school, so some suffer.

Checking one city in the neighborhood, San Francisco is 33% Asian.

Don't you think that might influence the makeup of Berkeley just a bit?

Of course it wouldn't. Why wouldn't it? Because Berkeley is not just part of the UC system, but the crown jewel of it, and serves the entirety of California, and everyone applying to non-community college schools in California is aware of Berkeley's prestige. Berkeley isn't a commuter school that serves the local area like Golden Gate or U of San Francisco.

Sorry, but I see a number of kids here in California choosing either you sees that are far away, but part of them close to home, for their preference.

I admit that out-of-state students May treat it completely differently.

Actually my sister went to UC Berkeley and she reported that she was like the only Orange County Republican those kids have ever seen.

While the majority of the UC Berkeley students might be liberal, the Berkeley College Republicans are the largest student organization on campus, both today and in 1968.

The city of Berkeley and nearby Oakland has always been more leftist than the university, and townies often infest the protests.

What I'm saying is, "East Coast colleges should have Asians at the same concentration as West Coast Colleges" seems a suspicious argument.

are you saying California colleges should reduce their Asian representation to better match National averages?

Sorry that was kind of unfiltered speech to text.

Look, Harvard is in a unique situation being the number one destination School in the country, but I'm just not seeing a solid grounding for the idea that it should have 30% preferential students, and more then 23% Asian at the same time.

It is combining two unrelated things and expecting an unspecified result.

Look, all we want is for Harvard not to discriminate against Asians solely for being Asian. Is that too much to ask?

You have no direct evidence they do, you're only claiming that because an intermediate score in the process tilts a certain way for a certain group it must imply discrimination in the overall process.

This despite Harvard selecting Asians at four times their representation in the general population.

No one is discriminating 'against' Asians (or whites), they are discriminating 'for' the other minorities. The effect is the same but it's a distinction that matters. It is attempting to help groups with problems that are not primarily self-inflicted.

I wish they would just switch to socio-economic quotas, plenty of the preferred minorities would get a leg up that way including poor whites from the sticks.

How can you write such sophistry with a straight face?

According to the article, Card's argument is that if you include the "personal qualities" ratings (one of the "four categories") as part of the criteria, there's no discrimination against Asians.

That, of course, is blatant question-begging. Of course if you follow the system as designed, you get the results of the system. The argument is that admissions officials depressing the "personal qualities" ratings of Asians is the mechanism of discrimination against Asians. Simply pointing to the low "personal qualities" ratings assigned by admissions officials who have never met the applicants doesn't defend the ratings given or their use in the admissions system.

By analogy, consider if Harvard's "four categories" were academics, extracurriculars, percentage of Germanic genetic heritage as measured by AncestryDNA, and athletics. Well of course whites would be more likely to do well under the system than Asians; it would also be blatantly illegal to use the system.

So, what Card's doing when he talks about "multidimensional" admissions is obfuscating things in hopes of confusing the issue.

The argument over what population you analyze is a similar case of using word clouds to obfuscate. It's easier to isolate signal from noise when you use similar populations; when Card calls it "nonsensical" to do so, he's being mendacious.

If the analysis by Arcidiacono had been done on the whole applicant pool, the obvious criticism to make would be that statistical analysis of the whole is invalid because it can't tell the difference between forbidden racial discrimination and allowed legacy and donor preferences. Given that Arcidiacono took steps to isolate those factors, the criticism being made is that it's invalid because it isn't looking at the whole pool of applicants. No possible statistical analysis of possible discrimination would be invulnerable to this sort of post-facto cherry-picking.

This might be one of those cases where it's good to be smart, but not too smart.

The mega-brain answer here seems to be that you can decode racism even as Asians are drastically over-represented relative to the general population.

Can you, really?

I say their total percentages are too close to those of the UC system, or even cherry picking Berkeley.

"An applicant’s race, they [Harvard admission officials] said, can help, but not hurt, his or her chances of admission." Why would I want to attend a college that thinks this is a logical statement?

Because it's freaking Harvard, that's why. Kind of a golden ticket.

marplot [mahr-plot], noun
a person who mars or defeats a plot, design, or project by meddling.

This issue is more complicated than who is in the sample.

Trouble is, tests are not the goal, but a proxy for the goal. The goal is to produce well-educated, productive, well-connected people who go do interesting things and give the school a good name and make money and give the school some of it so it can continue to buy stars and dazzle people and admit the next batch. Part of this goal must be to choose students so that the school will have a successful bottom quarter. If a school only admits total academic stars, some of them will be disappointed and bummed that they end up in the bottom of the class. The athletes and legacy admits can avoid this outcome. They already feel happy with their place in the world and are not insulted or broken-hearted to be in the bottom quarter.

The Asian students test better than they are. The tests are just proxies, they are not the goal. But when a test is introduced, students will aim very hard at the test because it is a metric that is used. Other schools have kept numbers for a long time on who-did-what later, and knows this. My guess is that hvd does this too, but does not want to talk about it.

One thing that has happened in the Bay Area (because we have a large Chinese population, which has stimulated all these homework and tutoring shops -- really there is one on every block) is that because the Chinese study so hard, then the others must too, and they resent it. It is an inefficiently high standard. The test is not good enough at measurement to be worked on as hard as it is worked on. Yes, I am saying that the level of studying is inefficiently high.

The high school suicides due to school pressure around here make the national news. There have been a dozen or so over the last few years. There are now full-time attendants stationed at places where the road crosses the train tracks to inhibit these suicides. What does not make that news, but is discussed in the Bay Area Mandarin newspapers, is that all of the suicides were Asian kids. Not some, but all.

Excellent post.

#5

since it's behind a GDPR wall - I wonder what countries law is applied, when a murder happens in Antartica?

Article 8 of the Antartica Treaty states that each country has jurisdiction over their citizens. Does that mean this murder is treated like it happened in Russia, even if the victim was an US citizen?

Comments for this post are closed