Thursday, February 02, 2012

Transparency in college admissions

Daniel Golden (Bloomberg) reports that the Department of Education is now investigating both Princeton and Harvard regarding discrimination against Asian-American applicants.

Below is an excerpt from an op-ed Bloomberg asked me to write on this topic.

What Harvard Owes Its Top Asian-American Applicants: Stephen Hsu

It’s a common belief among Asian- American families that their children are held to higher academic standards than college applicants from other ethnic groups. Such practices were openly acknowledged after investigations at universities like Berkeley and Stanford in the 1980s and 1990s.

Have they been corrected?

The U.S. Education Department is investigating complaints that Harvard University and Princeton University discriminated against Asian-Americans in undergraduate admissions.

Statistics seem to support the claim of bias across most of elite higher education. For example, in comprehensive data compiled as part of Duke University’s Campus Life and Learning project (as reported in a recent analysis by Duke economist Peter Arcidiacono and collaborators), Asian-Americans who enrolled at the school in 2001 averaged 1457 out of 1600 on the math and reading part of the SAT, compared with 1416 for whites, 1347 for Hispanics and 1275 for blacks.

Holistic Admissions

There is every reason to believe that a similar pattern holds at nearly all elite universities in the U.S., with notable exceptions such as the California Institute of Technology. In fact, Duke may be one of the mildest offenders when it comes to Asian-American admissions: With the goal of increasing its overall student quality, Duke has reportedly been more friendly recently to Asian-American applicants than traditional powers such as Harvard and Princeton.

Schools like Harvard and Princeton brag that each year they reject numerous applicants such as Jian Li (who filed a complaint against Princeton) who score a perfect 2400 on the SAT. How would we feel if it were revealed that almost all of these rejected top scorers, year after year, were Asian- Americans? I challenge Harvard and Princeton to refute this possibility.

To be fair, most elite universities practice what is known as holistic admissions: Each candidate is evaluated on a variety of measures, including athletic and leadership activities in addition to academic performance. It is possible that the gap in academic average between Asian-American and white admitted students is compensated by gaps in the opposite direction on these other variables. Looking again at internal evaluations by Duke’s admissions office, we find Asian-Americans had higher averages than whites in the following categories: achievement, curriculum (each about one-third of a standard deviation) and letters of recommendation, while trailing very slightly (less than one-tenth of a standard deviation) in personal qualities.

Lacking data on factors such as legacy and recruited athlete status, we can’t make a complete determination of the fairness of the process, and in fact the appropriate weight of the various factors in a holistic admissions process will be subject to vigorous debate. ...


dwbudd said...


good luck with that.  I don't disagree with or oppose what you're saying in the least.  Quite the contrary.

But on what ethical, moral, or legal leg would one rest to argue that data comparing Asian and white applicants should be released in toto, but that similar data for black and Hispanic applicants should continue to be shielded?  And you and I (and presumably, everyone else reading this) knows what the end-game of that scenario would be.  If the gap (on average) between white and Asian applicants on the math/reading portions of the SAT is 41 points (1457 vs. 1416), and the the gaps between white and Hispanic and white and black applicants are, respectively 69 and 159 (1416-1347, 1416-1257), what would such data mean for affirmative action?  Essentially, the average distance between white and black students is four times the gap between white and Asians.  Such an analysis would almost surely expose to sunlight the "game" schools like Harvard have been playing for decades.

You actually think that a Democratic administration that MUST get massive majorities from these two ethnic groups to keep power are going to hand over the weapons with which to eliminate them?

Call me crazy, but I don't see Eric Holder leaping off the couch to help.

Bobdisqus said...

Isn’t “holistic admissions” the same pig just with lipstick (
I didn’t see it mentioned in your op-ed discussion of “holistic admissions”,  however what roll does probability of future donations play in that formula and do you accept it as a reasonable objective on their part?

steve hsu said...

>data comparing Asian and white applicants should be released in toto, but that similar data for black and Hispanic applicants should continue to be shielded<
I'd like to see *all* of it released, but I agree with you there are reasons why DoE won't push too hard for this.

RKU1 said...

Betcha this comment thread will be a *lot* longer than the previous one...

RKU1 said...

How would we feel if it were revealed that almost all of these rejected top scorers, year after year, were Asian- Americans?
That's a very clever point, which I'd never myself considered.  It represents a very simple, clear, and self-contained question, which would be difficult for the top Ivies to avoid answering.

Based on that long public list of National Merit PSAT scorers from CA which was floating around a year or two ago, my guess would be that 90-95% of the 1600ers which Harvard and Princeton reject are Asian, though I suppose the true figure could be 100%...

steve hsu said...

Maybe someone at the Dept. of Education will check on this question for us ... :-)

LondonYoung said...

Think about that when you cast your vote this fall.  

LondonYoung said...

Sad, isn't it, that your comment can have a smiley emoticon at the end?  Imagine if this were bias against blacks instead of Asians. Your smiley might be the end of your career ....

Yan Shen said...

 "Call me crazy, but I don't see Eric Holder leaping off the couch to help."

Call me crazy also, but I haven't seen much enthusiasm from right-wing white Americans either in terms of ending the blatant discrimination in the college admissions system against Asian Americans. As I've stated before, I suspect that most right-wing white Americans who are against affirmative action for blacks and Hispanics vis-a-vis whites are motivated primarily by perceived ethnic self-interest rather than principled universalism.

Yan Shen said...

Because East Asians are highly skewed away from V and towards M, wouldn't this have drastically reduced the number of East Asians who scored a 1600 on the SAT?

RKU1 said...

I greatly fear that the possible impact of distance may have given "LondonYoung" a severe misunderstanding of the political dynamics in today's America...

Virtually all mainstream conservatives/Republicans strongly support racial AA, though many of them sometimes pretend that they don't.

In fact, in many respects Republican/conservative AA is much more extreme and ridiculous than the Democratic/liberal kind.  This is partly because they feel they have more to prove ideologically and partly because their available talent-pool is far shallower.  For example, roughly 90% of politically-active blacks lean Democratic, and this skew is perhaps even more pronounced at the upper-ability ranges.  Therefore, when Republicans need to fill their racial quotas, they often are forced to dig very, very deep, thereby turning themselves into total laughingstocks, the more so because they do this while endlessly proclaiming their adherence to color-blind merit.  

Numerous examples from the last decade easily come to mind, but here's just one.  When Obama was running for the Senate and his white Republican opponent dropped out due to a sex-scandal, the silly local Republicans selected as replacement a preacher-type rightwing black activist from an entirely different part of the country who'd never held a real job and whose only career qualification was that (I think) he'd once been Bill Kristol's college roommate.  Strangely enough, Obama won what was expected to be a very tight race in a huge landslide...

RKU1 said...

That's a good point.  But I had the impression that the "recentering" back in the 1990s had greatly increased the number of 800s in both categories, so I think it would have been more of a factor before that.

Maybe someone should start a website or a Facebook group called "I got 1600 but was rejected by Harvard/Princeton" and we can see the names of the angry students who sign up.

Yan Shen said...

Because of the East Asian math/verbal split, I'm actually not sure that East Asian Americans were over-represented relative to white Americans amongst perfect 1600 SAT scorers, or if they were, the degree of over-representation probably wasn't significant. The main problem with the SAT as far as most East Asians are concerned is that the math section has a very low ceiling.

By the way, if there's ever a Facebook group called "I didn't get a 1600 on my SAT and was rejected by Harvard and Princeton", you can count me in. ;)

Bobdisqus said...

I think it safe to assume we will find some Ashkenazim on the rejected list. 

I find it hard to feel much empathy with these rejected perfect scorers of any race or ethnicity.  What are the odds these people are going to go on to lives of failure and despair because of this rejection? Just what is the fall back school they will have to suffer through?

Guy_Brodude said...

Hopefully this gets some momentum. It seems to me, Prof. Hsu, that you are the national "face" of this quasi-movement? I'd like to think that your opponents (to the extent that there is serious opposition to what you are proposing) would not be so base as to slander you with cheap accusations of racism...but we shall see.

The issue, to me, is quite simple: Harvard and Princeton and every other private university should have the right to conduct admissions in a way they feel is best-suited to their own interests. But, if that admissions policy includes, either implicitly or explicitly, discrimination against Asian-Americans (or any other ethnic group), then it's not something I want to support with my tax dollars, whether through student loans, research grants or any other form of government aid. And everything I've seen indicates that the majority of Americans agree with me on this.

sineruse said...

The US Asian verbal SAT distribution is higher than that of whites, about 2x higher representation at 750+ in the late 1990's and presumably more than that since.   For total M+V SAT, in Espenshade's study of three Ivy League schools with Fall 1997, about 38 percent of the admission (if done by SAT alone) would have been Asian with the cutoff somewhere above 1500.  I don't know if that was 1997 applicants or a mixture with earlier years, in which case the late 90's Asian representation would be higher.  Of that, mostly East Asian, I assume.   Assuming the Asian proportion goes up higher in the scale, 50 percent today may be an underestimate.

Yan Shen said...

It's possible that 1) high Asian SAT-V scorers are disproportionately Indian American as opposed to East Asian and 2) selective immigration leads to a higher proportion of East Asians with high SAT-V scores than would be the case if East Asian Americans were a perfectly representative sample of East Asians overall.

I think if East Asian Americans were a perfectly representative sample of East Asians overall, you would find that they score slightly lower than white Americans on the SAT verbal, slightly higher on the SAT writing, and significantly higher on the SAT math.

sineruse said...

A math-verbal split does not mean lower average "V" distribution in the US E.Asian test-taker population.  It can mean slightly higher verbal and much higher math and a degree of upper-tail overrepresentation that depends on the weighting of "M" and "V" type factors.  In particular, if you take a pool selected on M+V and the later selections are on, say, 2M + 3V that give higher relative weight to V, you will see a groups with a M > V split will show high absolute (i.e., relative to population) representation, but a drop in representation compared to the first selection (i.e., underperformance).     I don't take "M" or "V", much less an M-V split, as axiomatic constructs the way you seem to, but they do fit -- or predict -- that aspect of the data.

sineruse said...

I don't think Indian numbers were high in 1997 (or 1993 or 1982-ish, which were the other years of Espenshade's data).  College board disaggregated "ethnic/sex data" tables with verbal scores also stopped around 1995-97. Select immigrant EA kids age 17-18 only start to show up en masse in the late 1990's, if even then.    I think US Indians outperform US E. Asians on verbal today but not all that disproportionately; you probably overestimate ability differences (if they exist) and underestimate coaching differences (which surely exist).

I also don't think there's a basis to judge how the average East Asian would do in English compared to whites if raised with it as a first language and culture.  Could be higher, lower, the same, or a mixture depending on sub-abilities considered.  For highly selected immigrant kids clearly it should be at least a bit higher, though with some disadvantage from immigration (gen-1) or a slowdown from bilingualism (gen-2).

jaim klein said...

May your wish be granted, so I am informed, is a Chinese curse. In yiddisch we say: Zol es im onkumn vos du vintsh im..

Steve, we Jews have been there and done that. The current American admission system is a product of a process that started by not admitting Jews, then the French Revolution's Equalité arrived to America and it resulted in an avalanche of Jews applying to the top universities and taking over the top places. Next an obfuscatory system was devised to limit the number of Jews to what was felt bearable. If not for those unmeritocratic criterions, Harvard would have been a branch of the Yeshive University. Even so, it almost is.

What I am saying is that absolutely objective meritocratic system of admissions will produce an ethnically unbalanced student body. You can work out if it will be 70% Chinese, 20% Ashkenazi Jew and the rest Indian, or some other mix. You cannot hide that and it will surely anger America's 150 million Whites and 55 million Africans and God-nows how many Latin mestizos. It will make them very very mad.

The pogroms of Malaysia, Indonesia etc. will look a picnic what may happen in America. Malays are  considered a kind, non-violent race; Germans and Anglosaxons are not. Be well.


Ju Hyung Ahn said...

Perhaps Asians should list African American Studies as their desired major and transfer later to different schools within the university.  I wonder how such procedure would improve the odds of being admitted.

Guy_Brodude said...

Pogroms? Honestly, I think the average American doesn't really care about Harvard; it's totally abstract to your typical middle-class, flyover-country type. 

dwbudd said...

Yan -  

I wouldn't call you "crazy" for trying to thread-jack what I posted with an irrelevant side-bar.  I would not want to speak for "right-wing white American" bogeymen, but I would agree that they (like virtually EVERY group of people) are far more likely to speak out against "unfairness" they perceive being targeted at them, and are likely to remain silent when it affects others.  Real altruism is rare.  

I would suspect that your complaints, for example, about the obvious discrimination that institutions like Harvard demonstrate against Asian applicants is motivated by your own ethnic tribalism rather than some high-principled "universalism."

dwbudd said...


I think you have to disentangle conservative support for Affirmative Action (I believe what you mean by "racial AA") from the sort that the GoP practise when picking candidates.  The two have surface similarities, but are in fact, radically different.  

The Illinois senate race (the one that propelled Mr Obama to national prominence) is a good example...of a specific pathos.  The Republicans understand that they typically will not get more than 10 per cent of the black vote.   The Democrats equally understand that without 90 per cent of the black vote, they simply cannot win an election.  Additionally, the national media, with obvious exceptions Murdoch's empire basically handmaidens of the Democratic party, continue to spread the propaganda that the Republicans are anti-black, anti-women, klansmen, whilst the Democrats are essentially the only sensible choice for blacks (and now, as well, Hispanics).

So, both understanding this calculus, will in races of this sort, opt for the candidates that give them the best shot at winning.  It's plainly pandering, and the sort of candidates that the Republicans are forced to field in part to fend off the ridiculous smears of racism are not terrific (with a few exceptions).  

This, at least in my opinion, is a far cry from explicitly pushing policies that discriminate in the common trades.  That the GoP will pick a nincompoop like Michael Steele to be their chair is NOT the same thing as passing laws or appointing judges that codify who gets into what school or will be hired for what job.  I'm not planning, for example, to apply to be the head of the RNC, and I would suspect few reading this blog are, either.  I WOULD suspect that most have kids who will be applying to UC Berkeley, or Harvard, or some other school where Democratic policies will most certainly have effects.  And I would further suggest, those effects will not be positive.

dwbudd said...

Steve, worse than that (not pushing too hard), I suspect that the DoE will put on a theatre of pretending to investigate.  IMHO, a phony "investigation" which on the surface exculpates the universities is worse than doing nothing.

steve hsu said...

I tried to word the op-ed carefully to make clear that anti-Asian discrimination, while certainly something that happened in the past, may have now be fixed. But I do think we (or at least I) need a closer look at the data to see what is really happening. I also think transparency about admissions will have other positive effects. You'll note in my last paragraph I again endorse color blind admissions.

Regarding Duke specifically, the researchers using the CLL data don't think that there is evidence for Asians underperforming their admissions strength, but I am awaiting further clarification form them. A technical point: they do not think senior grades are useful, due to compression of range (i.e., grade inflation), and think junior year class rank is more indicative of performance at Duke. They find that Asians are more prevalent in the harder STEM majors and have similar average grades to whites in those majors, in line with higher application strength. But, again, I await further analysis on their part.

dwbudd said...

This post (and Yan's following) illustrates perhaps the most toxic side effects of the more or less black-box process by which Harvard, Princeton, et al decide their admissions criteria.  In the absence of light, fungal plumes freely grow.  

Most people who are candidates for a place like Harvard (obviously, a self-selected and non-representative cohort) are more or less aware that the test scores and class ranks of those admitted versus those not admitted are not exactly "random" when distributed amongst racial categories.  Admissions officers say that there is no bias and no favouritism, but refuse to release the numbers that would back them up.  So, what we're then left with is essentially an argument from them to "trust us; the process is 'fair.'"

It's sort of like the old Marx Brother's line, "Whom are you going to believe, me or your own eyes?"

So suspicion and mistrust grow to the point that no one actually believes that there is merit in who gets "in" and who does not.  It undermines faith that the system is fair, and leads to all sort of speculation about star chambers and the like.  

As an aside, I've seen the Jian Li case - and find that the most interesting bit is not that he got rejected by Harvard (getting in I truly believe is an almost random, luck-based process unless your surname is "Kennedy" - I was first in my HS class and got an 800 on the Maths SAT and 770 on the Verbal, and this was 20 years ago before it was "re-centred," and I did not get in to Harvard.  No matter; later on, when I was going for my PhD in mathematics, I WAS accepted to Harvard, and had the pleasure of writing back to the department that I would not be coming using more or less the same condescending language they used when rejecting me as an undergrad, and informing them I would be off to Stanford instead) is that he also was rejected by Princeton, Stanford, MIT, and a few other schools.  I wonder what was NOT in his transcript?  

BTW, surely you must see the inherent fallacy of your FaceBook survey, no?  

Edwin said...

"You cannot hide that and it will surely anger America's 150 million
Whites and 55 million Africans and God-nows how many Latin mestizos. It
will make them very very mad."
African Americans,not Africans.Africans are admitted in the same pool as any other internationals,with no quotas.

Guy_Brodude said...

White gentiles, blacks and Hispanics are already under-represented at Harvard, Yale et al. relative to their overall percentage of the population. I honestly don't think the average (middle-class, middle-of-the-country) American really cares too much about Harvard; it's totally abstract for him.

I'm honestly not looking to start a fight or a flame war, but my back-of-the-envelope calculations indicate that Jews are a bit over-represented at the aforementioned schools (controlling only for intelligence of course). Most likely this is just a case of geography; Jews are disproportionately found on the East Coast and in the largest metropolitan areas, are thus more aware of the Ivy League and are more likely to apply than white gentiles of similar intelligence. I'm certainly willing to be corrected on this, though, if anybody has any better numbers.

Yan Shen said...

 "I would suspect that your complaints, for example, about the obvious
discrimination that institutions like Harvard demonstrate against Asian
applicants is motivated by your own ethnic tribalism rather than some
high-principled "universalism.""

That's funny David, because I've often been a vehement critic of affirmative action for blacks and Hispanics at the expense of whites. But then again, what would I know?

dwbudd said...


you use a lot of words essentially to say that universities like Harvard and Yale use non-academic criteria (e.g., athletic skills, who is a legatee) when deciding whom to admit.  That's not really a revelation.

This begs the question about whether Yale or Harvard, which are not country clubs or minor league athletic leagues are really "meritocracies," should use such criteria?  I don't argue that they should or they shouldn't.  

The most interesting conjecture you offer is about geographic (a proxy for SES) clustering.  Obviously, if more of group "X" live in say, Palo Alto versus East San Jose, and schools are trying to balance economically challenged vs. economically privileged areas, that will have a proxy effect.  If one looks at the data from where I spent most of my life (the San Francsico Bay Area), ethnic Chinese are clustered disproportionately in places like Saratoga, Cupertino, and to a lesser extent, Palo Alto.  There are relatively few Chinese in places like downtown San Jose or Redwood City, which are not as financially well-off.  I would like tot see the data not blended among "Asians," but instead, separated by Chinese vs. Vietnamese, for example, who DO cluster in economically poorer areas like central and East San Jose.  I would like to see how Vietnamese (who are not less "Asian" than Chinese people) are proportionally represented at the UCs and Stanford, and what their academic numbers look like. One of the challenges in the US is that people use a term "Asian," which encompasses a massive, diverse group; one the actually captures a significant number of Caucasians (South Asians), Chinese, Japanese, Vietnamese, and Pacific Islanders.  I would further suspect that those who are REALLY getting the short of end of the stick are southeast Asians (Viets, Cambodians, Laotians) who are economically a lot poorer, but nonetheless get pooled with ethnic Chinese who typically have far more resources at their disposal.

Finally, I find the whole "underperforms" argument a bit odd.  I presume by "underperforming," you mean students who are at the higher end of the test score/class rank distribution upon entry, but somewhat lower on the GPA/class rank distribution upon graduation.  Which do you suspect is less likely to be affected by bias - how one tests on the SAT, or what grades a professor assigns in a class?  I've seen this logic applied when arguing that the SAT is biased against girls, who tend to do BETTER in high school according to grading than boys, but do WORSE on the SAT.  The argument one must swallow is that a test, scored by a computer is less likely to be biased than a letter grade handed out by an all-too human teacher.

I find such reasoning suspicious.

Yan Shen said...

 Do you think that Sineruse has a point about the Duke Engineering school having more stringent admissions standards and East Asians disproportionately gravitating towards STEM? If a school says that they'll hold all STEM applicants to a higher admissions standard than they do liberal arts applicants because STEM fields are more rigorous and East Asians just so happen to be disproportionately concentrated amongst STEM applicants, that would certainly produce a negative effect against East Asian Americans without it necessarily being racial discrimination.

dwbudd said...

Yan, we've never met, and only you know your true motivations.  I don't spend a lot of time reading the food fights between various folks here, but have seen some of them.  I've not seen any posts here arguing against AA for black or Hispanic applicants "at the expense of whites," but that of course doesn't mean they do not exist.  A sampling here doesn't support that contention.

All of the arguments I've seen here are that Asians are at a disadvantage compared to white applicants.  

It's a blog written by Steve Hsu, so that's unsurprising.  

Personally, I consider myself a very, very "right-wing" person, and I would align myself with Professor Hsu, that the admissions criteria should be exposed, and if Harvard or Princeton want to say that they are admitting "the best" students, and all of those students happen to be ethnic Chinese, that's the way the ball bounces.  It's not in the best interests of our country to deny leadership opportunities to our best and brightest.  And whining about how "Asians work too hard" would ultimately put is in the same mess as France or Britain - declining, former world powers that are managing their own gradual demise.

BTW, I find it incredible that there exist groups like "Chinese for Affirmative Action" in San Francisco, who actually file amicus briefs against actions to dismantle laws that so obviously hurt them.  It's not quite the same as "Jews for Fascism," but it's not far off.

Yan Shen said...

 "I would further suspect that those who are REALLY getting the short of
end of the stick are southeast Asians (Viets, Cambodians, Laotians) who
are economically a lot poorer, but nonetheless get pooled with ethnic
Chinese who typically have far more resources at their disposal."

The Espenshade study stated that poor Asians benefited relative to whites in the admissions process from AA. Perhaps those poor Asians were disproportionately Southeast Asian and Pacific Islander as opposed to East Asian and South Asian and at least some colleges have already realized that they should separate out those ethnic groups accordingly.

steve hsu said...

Yes, it's a definite possibility. The whole issue is rather subtle, which is why I favor transparency.

It's entirely possible that older folks like me are conditioned by the bad old days 20 or 25 years ago when bad things were definitely happening, whereas today most of the negative effects on Asian applicants are not due to racial bias.

dwbudd said...

Interesting.  I am not familiar with Professor Espenshade's study, but maybe the schools are wising up to the fact that "Asian" is a somewhat inapt statistical grouping.  I recall the withering battle back in California surrounding Proposition 209, now nearly 20 years ago, which effectively banned race-conscious admission.  At least when done explicitly.  At the time, an analysis done by Jerry Cook at UCSD found that Vietnamese applicants to the UC med schools had a significant (about two-thirds) higher likelihood of rejection than did groups favoured by AA (namely, black and Hispanic students). 
Things may have changed.  

Anonymous_IV said...

DWBudd writes he "had the pleasure of writing back to the department that I would not be coming using more or less the same condescending language they used when rejecting me as an undergrad,..."

You probably realized that this kind of "retaliation", however psychologically satsifying for you, was far from original.  You may not have realized (but must realize by now if you've stayed in academe) that you were barking up the wrong tree: individual academic departments are almost entirely in control of their own graduate admissions, and have almost no say in undergraduate admissions which are handled by a different part of the university.

Guy_Brodude said...

sineruse, do you have a link to that Yale data?

Richard Seiter said...

dwbudd as an observer I would like to thank you for your civilized style of debate and cogent statements.  I have trouble finding sources on the far right which make arguments I find compelling (e.g. I wonder, as I think Yan Shen does, how many far right folks would advocate your principled stance on meritocracy when they are not the party benefiting.  I mostly agree with your alignment paragraph but do wonder if there is room for some AA.  I think diversity of viewpoint can add benefit over a meritocracy defined over a too limited set of criteria.).  Can you recommend any blogs or periodicals for me?

About your last statement it would be interesting to know their motivations.  Could it be a form of appeasement/camouflage?  An analogy would be the observation that closeted gays are sometimes the most strident anti-gay crusaders (I would welcome a better and less flame war prone analogy, but could not think of one offhand).

jmct said...


Your piece made the 'top stories' section on the Bloomberg terminal. You're big time! Don't let it go to your head!

steve hsu said...

What about the $cha-ching! part? When will that happen? :-)

I'm getting lots of emails from finance people about the piece. Almost all are favorable. A few are mad at me that I don't address Affirmative Action, but I tell them to go back and read the final paragraph more carefully, and then they are happy.

dwbudd said...


thanks for the comments; rest assured, I can be as polemical as the next guy, but have found it to be of low utility when trying to get actual information.  Can be fun sometimes to engage in an all-out flame-fest for amusement.  

Don't get me wrong, I would not say that I am any more principled than Yan or any of the aforementioned "White Nationalists."  I advocate for a more meritocratic process because I think to do otherwise puts our *country* at a competitive disadvantage, and thus I am far less concerned with how policies might affect my specific demographic sliver than with the bigger picture.  I may be somewhat cynical, but I suspect that if you take virtually anyone and really boil their arguments and motivations, there is at the centre, self-interest.  Maybe Yan Shen is a saint - he might very well be - but I think if he is, he is in very, very exclusive company.  

As you may have heard, Bernard Shaw once said something to the effect that the power of accurate observation is often called cynicism by those who lack it.

With respect to whether there is "room for some AA," I would ask the question "cui bono?"  Who benefits? Diversity of opinion and experience surely can have positive effects on learning, on creation, and on management.  One need look no further than the scandal revealed in Japan during the whole Olympus fiasco to see the corrosive effects of solipsistic, horizontal management.  Adding differing viewpoints allows perspectives and depth in the decision-making process.

The problem is, this is not the sort of "diversity" a place like Harvard seeks, which is much more a phenotypic diversity.  I am sure that the admissions folks at Harvard would be far happier to have a student body whose ethnic make-up exactly mirrored US census data by race and geography, even if every single one of their students came from a house-hold with a father who was a lawyer and a mother who was an architect, who voted a straight Democratic-ticket, attended essentially the same upper-middle class high school, and had the same income.  They presume race as a proxy for everything else.

Can't recommend any real "right wing" blogs - I don't really read any particular blog or magazine regularly, aside from Professor Hsu's of course.  I try to read a somewhat wide variety of sources, since no one has an exclusive licence on "the truth."  

Finally, with respect to the last statement, I presume that "Chinese for Affirmative Action" (an outfit in the San Francisco Bay Area) exist largely for political reasons.  In a region where Willie Brown (former California Assembly speaker and mayor of SF) is attacked as being a right-wing stooge, not being aligned with the Democratic party and its goals assures one of an almost complete lack of access to political power.  Self-loathing or closeted gay people are (I think) motivated by other, more personal goals.

dwbudd said...

Of course.  I realise it now; I realised it then.  It was not "original," but then, as a mathematician, I've been accused on many occasions of lacking imagination.

All of which is beside the points that 

a) Stanford offered a better programme
b) Palo Alto, Calif vs. Cambridge, Mass is a pretty easy QoL choice
c) The psychological satisfaction was immediate, and powerful.

We all know that giving the finger to the guy on the Bayshore Freeway driving 30 in the number one lane is not particularly effective, but it sure is satisfying, even if briefly.

5371 said...

Ever heard of running amok? Ever wondered where the word comes from?

dwbudd said...

Ha ha ha.... Never thought of that.  

RKU1 said...

Well, I seem to recall that just a few years ago, the Bush Admin came out very strongly in favor of university admissions policies which ensure the same full "racial diversity" as by explicit quotas, and this was a major factor in swaying the Supreme Court in a landmark decision.  But perhaps Bush and all his supporters weren't "real Republicans."

Similarly, when Pete Wilson decided to make national opposition to racial AA his ticket to the White House in 1995, he put one of his black cronies in charge of the effort, an individual whose entire livelihood for decades had been heavily derived from minority-set aside contracts and that sort of thing.  Unsurprisingly, that individual has spent the subsequent 15 years siphoning off an astonishing fraction of all the cash he's raised for this efforts into his personal bank account, although the totally incompetent national media only broke the story in the last few months.

For a good half-century, a very substantial fraction of the American public, certainly including most Democrats, has been greatly unhappy with most aspects of racial AA, which, after all, was largely begun by the Nixon Administration over strong Democratic opposition.  However, for the obvious reasons I cite, conservatives have failed to capture much public support on this issue.  I suspect that many AA opponents may be less well educated, but they're not total imbeciles.

RKU1 said...

Impressively detailed analysis.  It does seem plausible that there's a very large athlete/non-athlete gap, and this could account for most of the reported Asian/white gap.  Also, the idea that a threshold-selection effect plus the different shapes of the white and Asian achievement curves might be a factor is interesting.  Obviously, getting access to the underlying data would be crucial in analyzing these possible effects.

RKU1 said...

Well, that's certainly possible.  But I don't recall seeing any enormous density of Jewish names among the complete public list of high-scoring PSAT students for CA a year or two ago.  And CA contains a pretty good slice of America's total Jewish population.

Bear in mind that over the last generation or more there's been a total demographic collapse in America's high-end Jewish population, and a very disproportionate share of births have derived from the more religious strata, which tend to be academically undistinguished.  Anyway, even given religious fecundity, I think the overall age-distribution curve is among the older-skewed for all American ethnic groups, though I haven't looked at the figures recently.  Teenagers can't get 1600s if they don't exist.

dwbudd said...


first, thanks for the reply.
On your first paragraph, I'm most assuredly *not* a Republican, so I don't want to wade into the "who is a real Republican" food fight.  The idea that George W Bush - who presided over an enormous expansion of federal power and spending - is any sort of "conservative" is risible, but that's another argument.  But if you're talking about the Grutter case (which upheld the University of Michigan's ability to use race in deciding admission), the Bush justice department filed an amicus brief on behalf of the plaintiff, who sought to have it overturned.

The facts of the case are that the decision that essentially allows overt racism to be used in making admission decisions was a 5-4 vote, with the four "liberal" justices, unsurprisingly, supporting the University, the four "conservatives" siding with the plaintiff, and the swing vote - Sandra Day O'Connor casting the deciding vote in a legally dubious argument that in my view boils down to, "I know that this is wrong, but we need to continue doing it for a while longer."

Similarly with Pete Wilson, that Ward Connerly has enriched himself personally, perhaps illegally, hardly makes the case that one ought to vote Democrat, in this case in spite of the fact that you KNOW they will push laws and judges who will hurt you, but the "other guy," well, a single, well-connected individual may siphon off some of the money.

BTW, I'd love to hear the counter-argument when considering many of those pushing for AA (e.g., Jesse Jackson) have become millionaires essentially extorting money.  T.J. Rogers, the CEO of Cypress Semiconductor once famously called Rev. Jackson "a seagull- he flies in, sh*ts on everything, and then flies out."  (Mr Rogers tends to be a bit more plain-spoken than I am, so apologies for the coarseness).

I think your history about AA is misguided.  President Nixon surely helped entrench AA, but the idea that it was "largely begun" by the Nixon Administration over "strong Democratic opposition," is misleading.  AA began officially in 1961 (JFK) and greatly expanded by LBJ in 1965 (both under executive orders, not votes).  The significant legislation (the 1964 Civil Rights Act) famously attracted a quote by Hubert Humphrey, later the Democratic candidate for President that if it ever led to quotas, he would "eat the paper it's printed on."

Of course, less than a decade later, it had done just that.

Many of the Democrats who opposed it flocked to the GoP after 1968 and Mr Nixon's infamous "Southern Strategy."  

Nixon made things worse, as he did with virtually everything he touched.  

If we fast-forward to the 1990s and recent days, I think most people (as you say) have been unhappy with aspects of AA.  It's been exposed for what it is, and that is going to turn off most people who are not nakedly looking to exploit it for their own gain.  It's been reformed, and things are better.  But plainly, the data that Professor Hsu offer indicate that more needs to be done.

With a straight face, which candidate would you say is MOST likely to repeal AA?  

a) Barack Obama
b) Mitt Romney (or whichever of the guys in the clown car race ultimately wins)
c) Ron Paul

A bit to the contrary to your final sentence, though, I think that while conservatives have not been able to capture much support, the GoP have captured a lot of support on this issue.  The problem is, they've failed to actually do much, instead using it as a stalking horse to grab power and then work to benefit corporate task masters, who do not want AA dismantled.

Bobdisqus said...

For “right wing” blogs I highly recommend my favorite old cold warrior curmudgeon.

Guy_Brodude said...

Yale may be an exception to the general rule here. They are acutely self-aware of their weak reputation in STEM and are making an effort to improve it, including recruiting talented high schoolers with a demonstrated interest in STEM subjects.

I'd wager that the chair of the Physics Department at Harvard doesn't make calls to prospective undergraduates. Wasn't Yale the only school to accept Jian Li?

Ju Hyung Ahn said...

Thank you for sharing your insights.
Since RKU1's brief of George Bush's position on affirmative action is very misleading, I'll add couple sources to bring light on this matter:

I. George Bush was openly against affirmative action as the universities practice it now.

[Indent] Q: Ari, on affirmative action, because the
administration is not asking the court to overrule the Bakke decision,
is the government not conceding that it is okay for race to be a factor
when selecting students for college admission?
What the President is saying is he, as President, is setting a vision
and a goal for the country, and that is that diversity on our college
campuses is an important goal to achieve. He is saying the manner in
which the University of Michigan, by giving students 20 points on the
basis of the color of their skin, and only 12 points, for example, on
having a perfect SAT score is the incorrect way to achieve the goal of
The President, as you look at his record in the state
of Texas, adopted different approaches to dealing with this difficult
issue about how to promote diversity on campus and to do so in a way
that is race-neutral. And the President is urging and is pushing through
this brief and through his statement yesterday the wheels of
universities in the direction of increase your diversity and do so in a
way that is race-neutral. And that's the message that the President is
sending. [End Indent]Source: George W. Bush: Press Briefing by Ari Fleischer

II. His administration also filed a friend-of-the-court brief against UM's affirmative action program as dwbudd noted.

Brief of Amici Curiae United States, Grutter v. Bollinger, 123 S. Ct. 617 (2002) (No. 02-241).


Looking at the list of amicus brief filers from the above source, it is clear that opponents of AA come from Republican and Libertarian background and proponents from Democratic.  Where are all these supposed democrats condemning AA?

Frankly, I find it irrelevant whether many independently-minded,
politically-left intellectuals oppose AA or not when the elected
officials of the Democratic Party are reluctant to make changes.  It seems
to me that the Bush administration and conservative judges, despite
their shortcomings, made concrete efforts to remedy the problem.  This gives me all but one more reason to support "crazy" Republicans.

RKU1 said...

President Nixon surely helped entrench AA, but the idea that it was "largely begun" by the Nixon Administration over "strong Democratic opposition," is misleading.

Actually, Nixon's "Philadelphia Plan" is usually regarded as the landmark step behind the creation of the modern AA racial quota system, and was highly controversial at the time.  Perhaps the tentative Kennedy/Johnson efforts might have anyway eventually evolved into something like this, but perhaps not.

As for comparing the relative attractivenss of the present Democratic and Republican positions on AA, I think it just boils down to whether you prefer your racial quotas served straight, or with a tangy flavoring of dishonesty and hypocrisy on top.

And obviously Ron Paul is the only current candidate who might---just might---do something about AA.  Which is something like 137th reason he's regarded as a lunatic-fringe figure by respectable elite opinion, certainly including respectable conservative/Republican elite opinion.

RKU1 said...

I really don't think my analysis of Bush position on Grutter was misleading.  Frankly, I've been following this issue closely for decades and was certainly paying attention at the time.

Conservative anti-AA activists had spent a full generation attempting to lay the legal and ideological groundwork for a Supreme Court decision intended to finally overturn Bakke and ban AA.  Lower court decisions had gradually worked their way up to the Appellate Courts and finally reached the Supremes.  Years of bitter confirmation battles had also gone into shifting the necessary personnel.  Everyone knew the Supremes were closely balanced, but it was widely believed that the balance would tip in the anti-AA direction.

However, strong internal voices for a Bush brief directly attacking AA (and its dishonest equivalents such as "diversity") lost out, and the Bush people essentially endorsed "pro-diversity measures" while opposing "rigid numerical quotas" or that sort of thing.  They basically said that diversity was an important value, but it should be implemented through the sort of "holoistic" approach used by e.g. Harvard rather than by e.g. giving blacks 20 extra points.  Most of the more sincere anti-AA conservatives were completely outraged, but the Bush position probably helped move O'Connor, the swing vote, into following elite opinion and effectively reaffirming Bakke for another generation.

I looked on the Wikipedia page for Grutter, but (unsurprisingly) it seemed at total variance with my own recollection, so I spent a couple of minutes Googling, and located a couple of rightwing discussions of the issue, much closer to my own opinion.  These particular sources are hardly the most reliable, but I do think they're completely correct in this particular case, and they provide lots of additional links.

The bottom line is whether or not racial AA is still around today.  If so, it survived because of the Grutter decision.  Frankly, facts like that seem more salient that the desperate spinning of a flack like Fleischer.

David Coughlin said...

For all the discussion, I get stuck focusing on "Maybe someone at the Dept. of Education will check on this question for us ..."  The Educational Establishment does not want insightful outsiders peering into their operations.

David Coughlin said...

Sometimes I wish you had subject lines because the subject of this line would be 'Forming, Storming and Norming'.  The putative goal at UltraMegaBigCo when you put together a team for a project is to get the right mix skills and personalities.  You frequently have to work across company bounds and in the context of three or four hundred people trying to get their jobs done.  Every player on the A-team doesn't have to be A-grade, but they do have to A-contributors on the project.  I wouldn't be surprised if the elite schools thought they had some secret sauce in their admissions philosophically aligned with HR verkakte like that.

Ju Hyung Ahn said...

Perhaps the situation we're facing could be best described as being between Scylla and Charybdis.  Facing a total destruction of our fleet by Scylla, do we have the luxury to question the motives behind Charybdis's offer of help?

highly_adequate said...

The real and deep problem I have with Steve's complaints about discrimination against East Asians has everything to do with their underperformance -- not so much in the academic setting, but in the real world.

Take a serious look at the proportion of East Asians who receive the most distinguished honors in any profession -- even in the STEM professions  which they tend to prefer -- and you will virtually never find a proportion that nearly matches the proportion of East Asians who already DO get accepted at elite institutions. That is, if, say, 30% of the elite STEM programs are now East Asian (roughly true, I'd guess), one would expect to see 30% of the most honored contributors in the STEM professions to be East Asian, but the numbers appear to be vastly below that. (And how many distinguished contributors outside of STEM fields who are East Asian can you think of? I can't imagine that they exceed even the proportion of East Asians in the larger population, not to speak of the much larger proportion admitted to elite institutions.)

If it were otherwise -- that is, the proportion of distinguished contributors were in excess of the proportion admitted to elite institutions (as was true for Jews, I believe, back when the issue of Jewish quotas was a live one) -- then I could see a compelling argument that the East Asian future leaders of America in various disciplines were not being given their due because of discrimination.

But it isn't so. Perhaps someday this may change, though I'll admit I'm skeptical. But, given the current evidence, it looks as if Harvard is getting it right when their admission policies don't accept more East Asians, insofar as Harvard is fundamentally interested in achievement of its students after graduation.

In the end, East Asian students have to demonstrate they can make a big splash in the real world. Until they do so, my guess is that Harvard and every other elite institution is going to find one way or another to adopt policies whose effect is to keep the number of East Asians admitted at a reasonable figure.

highly_adequate said...

Look, the data points on the number of East Asians in American Men and Women of Science is irrelevant to the point I'm making. The number of scientists listed in that publication is in the latest edition, 132,324. Do the math on that: assuming a career of 30 years, that's a little over 4,000 a year. If one restricts oneself to the top 10 elite STEM programs, that would be about 400 grads a year. In other words, just about every grad from an elite STEM program might qualify (though of course many will graduate from other institutions). I should think that the East Asian students at those institutions might achieve that much at least -- they have to go somewhere, and do something, right?

And honestly I don't think the names on the papers that you link to mean much -- God only knows how important the papers will ultimately prove out to be, or how important might be the contributions of East Asians among the numerous authors on the papers. I would think a more objective measure would be top medals and honors in a given profession. While there may be a time lag, I certainly have the impression that the high proportion of East Asians in elite STEM programs has been a fact for at least 20 years, and possibly 30 or more. Where are the achievements comparable in proportion? Science in particular tends not to have a long time lag in achievement and recognition -- 20 or 30 years is quite an ample time to come to great prominence, if one is likely to do so ever.

And Jews certainly were major achievers at minimum in Europe by the early twentieth century, and, so far as I can recollect from my history, were already becoming a major presence even in American science by roughly that period. Where can one point to comparable success for East Asians? It's only the Japanese who seem to have accomplished much on the world stage in science, and, oddly, they appear to have the least impressive SAT scores among East Asians in the US. Point is, SAT scores don't seem to be adequate unto themselves to make good predictions regarding ultimate success at the very highest levels. Perhaps the Tiger Mothering of the Chinese and Koreans actually does long term harm to the creativity of their children, however effective it seems to be in getting very high SATs and GPAs. Or perhaps the deficit is something more basic  -- I won't pretend to know.

And Silicon Valley demonstrates my point, I think, not yours. There are very few East Asians proportionately among, say, the leaders of the most successful companies in Silicon Valley -- though they may start a fair number of businesses. Again, it's big time success that's of relevance here, not simply throwing one's hat into the ring.

Yan Shen said...

 "There are very few East Asians proportionately among, say, the leaders
of the most successful companies in Silicon Valley -- though they may
start a fair number of businesses. Again, it's big time success that's
of relevance here, not simply throwing one's hat into the ring."

Really? Because I thought that companies like Yahoo, Nvidia, YouTube were considered fairly prominent.

highly_adequate said...

So Jian Li actually got into Yale, but still complained about discrimination against him at Princeton?

I know a student in my family who got perfect SAT scores AND perfect SAT achievement scores on the 3 or 4 such tests he took, and a near perfect GPA from a top HS, yet was not admitted either to Yale or Princeton -- though he was admitted to Harvard. He was of European background. It never occurred to him to complain about maltreatment at the hands of Yale and Princeton -- he understood the randomness of such decisions, and was quite happy things turned out as well for him as they did.

jaim klein said...

Did I say 55 million African-born Africans in the USA? What is your argument? 

RKU1 said...

Newt! Newt! Newt will save us!

On a more serious level, consider that our ideological and political landscape is filled with a large number of environmental niches.  To the extent that such niches are currently filled with highly ineffective groups or individuals, these prevent the emergence of more effective replacements.  Hence their existence is worse than useless.

LondonYoung said...

For the record, LY returned to the U.S. a few months ago after 7 years in London, but as he's living in NYC he never meets any other republicans ...

Guy_Brodude said...

According to Wikipedia, he got into Harvard eventually (via transfer).

I agree that it is a random process; probably there are a lot of people like your relative from all ethnic groups.

Reactionary_Konkvistador said...

One is hard pressed to describe Latin Americans as lacking blood-thirst when it comes to disposing of market dominant minorities. But any Chinese market dominant minority in the US may well in the long term forge and alliance with China and eventually pass into its protection (at least that is what I would be aiming for). 

nvnl said...

As a high school student in the US, I empathize with Dr. Hsu's call for more transparency in college admissions, but even though I want to know what the acceptance thresholds really are, it's dangerous to open up the process too much. The reason is, under too much scrutiny and litigation, the admission process may become formulaic and admission simply the expected result of meeting some set of public criteria. Given the intense competition to get into US colleges, this would certainly lead to perverse incentives for students to perfect a limited subset of prerequisites. If I knew with good certainty that I could get into Harvard with nearly 2400 SAT, a 4.0 average, an internship with mention in a publication, and a school sport; then I would quit my piano, pickup basketball, tutoring, and blog to focus on those goals. Certainly, these are impressive accomplishments, but much less so if I spent the entirety of my high school career (or longer) focusing on them alone.

This brings up another point: college admissions offices need the latitude to admit students that may have achieved less thus far, but had fewer resources to do so. After all, in real life and in real research, perfecting the prerequisites does not equate to being able to produce new value or insights. Who is more deserving of admission and has a higher potential to succeed: a student who studied 8 hours a week for a year, whose parents hired a private tutor, and who scored in the 95-th percentile; or one who had to work at his/her parent's shop 8 hours a week, could not afford tutoring, but still managed a respectable 90-th percentile? It's hard to say, especially because the first student would never admit to having had such an advantage in his/her application. Yet the numbers would show only their final results: a 95-th percentile candidate versus a 90-th percentile candidate. If admissions offices face pressure or find it politically easier simply to admit the higher achiever, then they very well may pick the lesser of the two candidates.

Moreover, in my experience, test scores like the SAT have been frequently emphasized by visiting admissions officers as a way to provide a cutoff for under-achievement, not as a measure of over-achievement. A student who comfortably achieves the 98-th percentile and focuses on other aspects of his/her high school career should not be penalized because another student studied harder and longer simply to ace the test. In fact, the first student seems wiser and better suited for the rigors of life and research. Considering all this, it seems like only with a grand sense of entitlement and quite a bit of naivety that someone would complain about not getting into his/her preferred college with perfect test scores.

So it is my opinion that a call for transparency in admissions may be needed, but it should not put undue pressure on college admissions offices to justify all their selections. There are right questions to ask and inquiries to make, but they don't include "why didn't so-and-so get into Harvard?".

nvnl said...

 I don't understand how this is necessarily relevant to discrimination. Yes, it's an accomplishment, but at the end of the day it's still just a standardized test. It seems like a petty thing to sweat so hard. And if the only good credentials of the applicant were his/her SATs, then it's a waste of intellectual effort, and the candidate should be rejected. Spend your efforts on real work, not perfecting a standardized test. Not saying this is necessarily the case, but the perfect SAT rejection issue is not sufficient.

zk7 said...

Once again, unmentioned in any of these discussions is that the most significant 'affirmative action'
discrimination takes place outside of college admissions.  So you have to go to Yale instead of Princeton, OMG, your life is over.  The reality is that, and it is not difficult to find out, Asian
Americans and Asian immigrants are together the biggest per capita
recipients of affirmative action where it really counts -  preferences in U.S. government contract
awarding through 8a and non-8a set asides and automatic bidding differentials and 'small business' financing / often forgivable loan set asides that are only for 'disadvantaged minorities', which includes all Asians, Asian immigrant, and all other non-white  (West of the Khyber Pass) minorities and immigrants (and non-8a set asides and bidding differentials includes Hasidic Jews)
and they also receive preferences in corporate supplier
set asides/preferences and in government job preferences.  The reason that most Asian
American advocacy groups, as opposed to individuals like Steve and Yan, are
mostly not joining in the 'outrage' about college admissions is because
they know the score, and it can only threaten the real and very large
preferences that Asians (both American and immigrants) continue to get.  The idea that Asians support affirmative action against their own interests is laughable, as is the college admissions focus of this blog.

zk7 said...

oops, should be 'East of the Khyber Pass', ie who the US government considers Asian vs White.  I guess it should be East of the Khyber Pass, south of Vladivostok, south of the Aswan dam, south of the Rio Grande.  Or some other nonsense.

Ju Hyung Ahn said...

The extracurricular activities should definitely be addressed, but this should be done in more systematic and even elitist way.  Playing instruments and sports is great, but we can't just take applicants' anecdotal evidence for it.  If applicants demonstrate proficiency in each area by participating/placing in competitive environment, many people will accept extracurricular activities as part of applicants' academic credential.  Tutoring and other community services can be addressed by means of letter of recommendation.  By the way, *shooting hoops for fun* shouldn't count as part of aforementioned credential.  There is obvious need to screen out more frivolous hobbies such as video gaming and playing cards.

"A student who comfortably achieves the 98-th percentile and focuses on
other aspects of his/her high school career should not be penalized
because another student studied harder and longer simply to ace the
test. In fact, the first student seems wiser and better suited for the
rigors of life and research."

Going from 98th percentile to absolute perfect score in SAT isn't something that can be done by whim.  According to Wikipedia, 98th percentile in current SAT is 2140.  SAT score of 2400 is 99.98th percentile.  If we crudely convert this to IQ, we get 153 vs. 131.  The cognitive profile of these two groups is markedly different.  Please don't take going from 98th to near 100th percentile as something like going from 80th to 82nd percentile.  These two processes are qualitatively different.

"nearly 2400 SAT, a 4.0 average, an internship with mention in a
publication, and a school sport; then I would quit my piano, pickup
basketball, tutoring, and blog to focus on those goals."

Contrary to popular beliefs that would like to label high SAT scorers as idiot savants, I generally find high scorers to be more polymathic (i.e., highly proficient in multiple areas they take interest in; not to be confused with dabbling into many different fields with little depth).

RKU1 said...

Well, it's obvious that a straight rank-ordering numerical admissions system would have huge negatives, and put even more ridiculous cram-pressure on the students, plus being quite bad from the perspective of "diversity" and national cohesion.

However, the current system---mysterious and secretative as it is---lends itself to the possibility of massive, hidden corruption and nepotism, and there seems huge evidence that this is a serious and growing problem.  Furthermore, corrupt systems to produce elites may tend to produce corrupt elites, and the disastrous calamities which our nation has recently experienced may be indicative of this.

Fortunately, I think there may be a third approach, which I'll have to try writing up one of these days.  The ancient Greeks were a very clever people about all sorts of things.

Sam H said...

sineruse: What do you do for a living? I'm just curious. 

Sam H said...

If there were NO racial and ethnic preferences at top schools there would be NO Blacks in many top schools maybe less than a handful of Hispanics. Preferences will never never take place. 

tractal said...

"Going from 98th percentile to absolute perfect score in SAT isn't something that can be done by whim.  According to Wikipedia, 98th percentile in current SAT is 2140.  SAT score of 2400 is 99.98th percentile.  If we crudely convert this to IQ, we get 153 vs. 131.  The cognitive profile of these two groups is markedly different."
You've made a major mistake here, because the SAT is less accurate as a measure of G the closer you get to the ceiling. The writing section probably doesn't measure much at all, unless you think that someone who loses 50 points for mistaking a few grammar rules and not having the essay section coached is really the 99.5 whereas someone who did study the arcane grammar and idiosyncratic essay is really the 99.98. On the verbal side, you don't think vocab study can turn a 750 into an 800? It really happens all the time. Its not unrealistic for a good tutor to add 150 points. In purely percentile terms it looks like turning the 98th into the 99.9, but the difference actually boils down to knowing how to ace the writing section, studying vocab, and learning how to not make careless mistakes on the math. 

I'm not saying that the 2400 kid isn't on average a lot smarter than the 2150 kid, just that training can have a large effect at the ceiling edge of a test. If both kids are easily capable of getting 9/10ths of the problems right, training them to not make careless mistakes/vocabulary boosting/essay "writing" formula is going to be more important  than it might seem. 

Ju Hyung Ahn said...

Notice the word *crude*.  It was obviously used to illustrate the dramatic difference of intelligence between the two population.  There aren't many tests that isn't susceptible to prep.  I suspect even most IQ tests are susceptible to it if people actually bothered to study each subsection.  Often, you see ultra-high IQ societies headed by psychologists and boast incredibly large memberships for a distinction (SD 4+) that they claim to have.  Harder tests simply require people who are appropriately high g (obviously with increasing return for even higher g people) to prep them.

Far too often, I see people make apologia for their "low" scores by promulgating that they have a life and others must be idiot savants who prep for tests 24/7.  However, I suspect the number of these idiot savants are miniscule (have miniscule effect on the admissions in the bigger picture and could be remedied with inclusion of extracurricular activities) and *most* people who scored higher will be simply smarter than people who scored lower.  Moreover, I already addressed extracurricular activities (with distinctions) should count.  If your extracurricular activities are spent on shooting hoops, playing piano, or blogging on casual level, I don't see how these extracurricular activities are extenuating circumstances for lower scores.
*EXCEPTIONS: Without loss of generality between V and M, A person *A* who scores 800 on M and less on V than a person *B* with 800 M and higher V may be smarter than *B* in M.  But, *B* will still be smarter than *A* in V in most cases.  This occurrence is especially common with international students.

By the way, 99th percentile in GRE/GMAT is more impressive than 99th percentile on SAT.  These two tests deal with different pool of test takers.  1480/1600 SAT scorers have expected GMAT score of lower than 760.

tractal said...

"By the way, 99th percentile in GRE/GMAT is more impressive than 99th percentile on SAT.  These two tests deal with different pool of test takers.  1480/1600 SAT scorers have expected GMAT score of lower than 760."
I think we're on the same page. My point was basically just that a rigorous test prep can make the difference between the 99 and the 99.95 on the SAT, whereas it is probably a lot harder to push the 99th into the 99.95 on something like the GMAT or LSAT where the ceiling is higher (and so where test taking techniques have less to do with top scores). In other words, the SAT would obviously be a lot better at distinguishing at the upper end if there were only "hard" questions (ala GMAT and LSAT, to an extent.) 
Because the SAT is so easy you should be careful before saying that the 2400 score is significantly different from the 2250 score in terms of IQ. So little of the test is dedicated to distinguishing at the upper end  that it is not unreasonable to think most of the  difference between an 800 M 750 V 700 W and a 2400 has to do with studying, at least in a significant number of cases. Again, the fact that the writing portion might not even be G loaded at all beyond some correlation with V should really give pause to the idea the SAT is really distinguishing the 99.98 etc. 


Ju Hyung Ahn said...

I agree with you on the limitations of the SAT.  However, there are also people who grind out their 2250 SAT score or even 2100.  Obviously, if you take these outliers too seriously the whole system is going to fall apart.

I just want to warn against these types of fear-mongering, often employed by political activists, that makes these extraordinary data-points as the norm.  It absolutely annoys me how people are so eager to accept that people who are more successful than them are just grinds.  I wouldn't mind if people used this rationale positively -- for self-motivation -- but far too often, people use this as a justification to put down intelligent people and serve their self-righteous purposes.  I want to conclude with the famous words of the legendary biologist James Watson, "I turned against the left wing
because they don't like genetics, because genetics implies that
sometimes in life we fail because we have bad genes. They want all
failure in life to be due to the evil system."

dwbudd said...


I would not say your analysis is "misleading;" for one thing, it imputes malign intent that may not be there, and for another, it's entirely possible for two people to look at the same history and come to an honest disagreement.  This is plainly illustrated by the fact that a quick search (one of Wikipedia, the other of Steve Sailer's "VDARE" site) reveals conflicting conclusions.
Race-based (or at the least, tinged) AA lives, proving harder to kill than a hockey-masked assassin.  A big "thanks" if you will, does go to the Grutter decision.  

One place I would differ, I think, from you is that I would tend to see "affirmative action" that is flavoured by socio-economic factors somewhat more sympathetically than I would AA based on race.  In that, I do think that society is not stable in the long term if it is perceived that the wealthy and the well-connected work to entrench privilege.  One of the best "predictors" of academic performance is whether mom and dad have university or advanced degrees.  I won't argue how much of that is environmental versus genetic - that is for another day.  But imagine a hypothetical where two applicants to UC Berkeley are more or less equivalent in terms of performance, but one is from the child of two executives in Palo Alto, and the other the child of two farmworkers in Salinas.  One would be perhaps the tenth admit from Paly to go to Berkeley, and the other perhaps the only kid from his high school to make the cut.

In such a case, I can see SES (socio-economic status) being the "tie breaker," and not having significant qualms about that.  What is practised nowadays treats a wealthy black kid from Palo Alto as being "diverse" compared to the other dozen kids from the same high school.  

It's a fanciful scenario, of course, and it is likely virtually never the case that two applicants are identical.  

BTW, in one of life's little ironies, perhaps the biggest beneficiaries of affirmative action have been white women.  Thus, in a sense, it's not a shock that Justice O'Connor voted as she did.

RKU1 said...

Actually, I've also always felt that SES-based AA would be perfectly fine, if implemented in a reasonable fashion.  My impression is also that such an approach would be pretty widely popular, including among many of the groups and individuals most hostile to racial AA.  Yet for some mysterious reason, the high-SES elites who tend to shape political and media discourse seem never to have given much consideration to the idea...

dwbudd said...

I wonder what that "some reason" might be?

nvnl said...

I don't think the conditional probability of going from 2140 to 2400, given plenty of studying and/or tutoring, in SATs is as low as you suggest. My experience is that most students who score something like 2140 already have math more or less "perfected" (an 800), and just need more practice with the other sections. But it's totally unwise to waste your time on perfection, UNLESS colleges eat it up and you get rewarded for it. This only reason someone with a 2140 would feel compelled to strive for a 2400.

This is not to say those that get a 2400 necessarily have put forth a significantly greater preparatory effort. But given the likelihood of misreading the preferences of colleges have for extremely high test scores, and perhaps being presured by their parents or culture; I strongly doubt their preparation effort, time or resources were the same for the perfect scorers.

Do I want college admissions to be about objective intelligence, but not at the expense of making it a war of attrition. And I believe people who claim higher test scores should always be rewarded more are the same people who are willing to invest more effort than the rest of the competition, as they don't mind the mindless drudgery and it would be an easy, relatively assured way for them to succeed. As I said, the SAT traditionally has and, I believe, should continue to be used as a cutoff, not as a mark of achievement. Other activities, perhaps frivolous to some, are where I'd rather spend my time and creative energies. It's precisely the unstructured nature of these activities that promotes intellectual discovery for me.

Please excuse me if I am presuming, but criticism of the shortsightedness of tests and metrics is not out of place in this debate or ever this blog. Dismissing "achievement" and "performance" seems to suit the author's point in these posts:

Also, I don't know why you'd ever compare SATs to intelligence, especially at
the high end of both spectra, as SATs are capped and intelligence
is not...

Blog Archive