Tuesday, March 19, 2013

Transparency and holistic admissions

All my Facebook lefties have been linking this article about Fisher v. Texas, the bulk of which is the usual accusations of racism and image-manipulation against opponents of affirmative action, but the chewy center (and what all my FB linkers have attended to) are these numbers: 
Even among those students, Fisher did not particularly stand out. Court records show her grade point average (3.59) and SAT scores (1180 out of 1600) were good but not great for the highly selective flagship university. The school's rejection rate that year for the remaining 841 openings was higher than the turn-down rate for students trying to get into Harvard. 
As a result, university officials claim in court filings that even if Fisher received points for her race and every other personal achievement factor, the letter she received in the mail still would have said no. 
It's true that the university, for whatever reason, offered provisional admission to some students with lower test scores and grades than Fisher. Five of those students were black or Latino. Forty-two were white. [MSI note: this is on. pp.15-17 of the brief.]
Neither Fisher nor Blum mentioned those 42 applicants in interviews. Nor did they acknowledge the 168 black and Latino students with grades as good as or better than Fisher's who were also denied entry into the university that year. 
See, so racial discrimination against Fisher was, if not impossible, at least unlikely, because 42 white students with worse overall admissions scores (as determined by the admissions committee, and including both academic and personal qualities) than Fisher were admitted, and 168 minorities with better ones denied. What interests me about this is that, under the "holistic review" regime, this is what it takes to disprove discrimination - the university must divulge the profiles of all applicants, rejections, and "provisional admissions," by race and admissions score for the given year. And I'd think this disclosure does not go quite far enough to prove that in Fisher's case, race was not the decisive factor. How many white students with equal or better admissions scores than Fisher were denied that year? Was it proportional to their application numbers? Was the minority rejection rate similarly proportional? The brief submitted by the University of Texas omits not only these details, but also the factors that went into the admissions scores in the first place - that is, what were the data behind these admissions scores, broken down by race? And this is only Fisher's single situation; think of how many Abigail Fishers there are in the country who apply to universities whose admissions profiles they more or less fit, even if at the low end, and are rejected. Can each one demand such exhaustive evidence from a school that her rejection was free of discrimination?

The goal of holistic admissions was of course to avoid this very outcome, in which every application datum would have to be quantified and plugged into a formula like the one that the medical school in Bakke used. Universities would henceforth not count race alone as a plus factor, but also such various pluses as violin aptitude, demonstrations of humanitarian concern, and being from the sticks. Everyone can get pluses, and from practically anything! The only catch is that these schools would not have to say how much of a plus any of these pluses was worth in the whole plussy scheme of getting in, because of course individualized evaluation varies by individual. So no reasonably qualified applicant can ever know why he was rejected, only that he was evaluated in his wholeness by professionals, and found wanting. There is a parallel anxiety from acceptance as well, the quiet hope that it was real qualification and not some accidental trait that determined one's admission, but this anxiety generates fewer public complaints, just as few students complain to TAs that their grades are too high. So although the opacity runs both ways - you can never know why you were either accepted or rejected - it's the opacity of rejection that is the source of public ire.

Ironically, under the very regime intended to avoid simplistic comparisons, the only persuasive defense against accusations of racial discrimination is near-total transparency in university admissions. It's not enough to insist that the school is X percent white/black/Hispanic/Asian to demonstrate nondiscrimination, since we can't know what percent it should be without knowing who was rejected and why. Harvard is 17 percent Asian, explains the Crimson, which is three times the percentage of Asians in the population, so we can know "that affirmative action is not resulting in far fewer Asians being admitted than there would be otherwise." But, of course, this - how many of the total the Asian applicants should have been admitted - is precisely what we cannot know simply from the number who were admitted.

And when it's some place like Harvard or Princeton, we can all pile on the sore losers who whine that their rejections were "not fair" because, with the overall acceptance rate so low, there is apparently no fairness or desert worth considering, so the opacity of admissions is actually an asset for the most selective schools, a kind of metaphorical analog to the statistical reality. Thus sayeth the ever-humble Crimson in the same editorial:
Making such a claim in the first place smacks of hubris—it is difficult to argue that any applicant definitely deserved to have gotten into Harvard, no matter how good his or her SAT scores or impressive his or her extracurriculars. Amazingly qualified applicants get rejected from Harvard all the time, making it almost impossible for this student to prove that it was his race, instead of any other factor, which resulted in his not being admitted.
It's because we Harvard students are all so amazing that you can't ever prove that you deserved to dwell among us. No one deserves such favor; it is a pure act of divine sovereign grace to be admitted. You should be grateful that the gods of Harvard even bothered to consider your application, little worm. This view obviously does not apply to state universities, access to which is viewed as a matter of desert and even right rather than divine election, subject only to some basic quality control. But the problem of proof, should an institution choose to undertake proof, is the same for both.

Private universities may be under less pressure to prove that they do not discriminate, but in either case, the only real demonstration of nondiscrimination under a holistic review system is transparency. Names might be withheld, but to show that a holistic review of an applicant did not discriminate in any partial aspect of his wholeness, the rest must be brought out into the light - how much did his violin playing compensate for his lower verbal SAT, and was that enough to merit admitting him over a minority with a higher verbal SAT score but lower class rank? Of course, this is precisely what universities least want to reveal, so much so that they won't release their admissions data even for academic research.

And their institutional caginess is understandable - the people who work in admissions are full of biases like the rest of us, and while few of them are likely acting on a grand racial agenda to either deny or admit all the minority applicants, they are given access to very personal information about each applicant, and they can't help forming judgments about them. This is where Unz's argument (adapted from The Gatekeepers) pointed to what I thought was the most likely cause of bias in admissions: people like people who are like them. So if the people doing admissions have never encountered 4-H clubs (and neither have I!), or think that good students who have few extracurriculars and little evidence of out-of-school engagement are boring drones, then subjecting such people's "holistic" admissions decisions to public scrutiny will reveal that yes, schools have been discriminating against some white (rural) students, or against Asians and other recent immigrants, but it won't reveal that this was not racially motivated, but rather the natural outcome of asking people with all kinds of pre-existing cultural and political interests to select a set of "interesting" candidates from a large pool. That is, it's the inevitable outcome of "holistic admissions." This is not necessarily because admissions committees are comprised of under-educated morons, as Unz alleges, or because membership in 4-H clubs and singular focus on academics are "regarded with considerable disfavor by the sort of people employed in admissions," but simply because they're unlike the priorities that these people hold. I suspect that universities know that transparency will make them look pretty bad and they'll be largely unable to justify their decisions as not racist, but having committed to this approach, they now have no way around this situation except to hide it.

So perhaps you can now see why I'm so intrigued by this defense of UT's admissions practices from the pro-affirmative action left. It's almost persuasive against Fisher's individual claim, but at what cost to the broader cause? Is every side now convinced that transparency will only vindicate their own suspicions? Can everyone now sue for - or simply demand - such explanations for their rejections as Fisher has gotten? (If so, perhaps Miss Self-Important will file some briefs of her own... Why you no love me 10 years ago, Wellesley College? No statute of limitations excuses - I demand answers!)


Phoebe said...

Responding on WWPD as well, but one small and unrelated point re: admissions officers and familiarity - isn't it possible that, when confronted with the unfamiliar, they'd celebrate it in the name of diversity-broadly-speaking? Not good students without extracurriculars - that would be bad-different - but something farm-y sounding. It seems like, after reading however many applications from kids in the Northeast who've all done lacrosse and tutored inner-city youth, 4H (which sounds agricultural? am I way off?) would stand out.

Miss Self-Important said...

Yes maybe, or at least getting a plus for being from the sticks may compensate for a bias against farmy sounding things.

But when I read the Unz article, I did think that, faced with an application claiming a state prize in cattle breeding or something like that, and another claiming a state prize in chess or debate, I'd likely assume the latter was more competitive and difficult to achieve b/c I know nothing about what it takes to breed cattle or who one would compete against in this activity. Or, if someone applied with a lot of local church volunteering vs. someone with a lot of inner city youth tutoring, I'd think the local church stuff was less interesting b/c it sounds easier and less useful, but really b/c I don't have a clue what doing that is like, and what its value really is beyond feel-good platitudes.

All of which is ironic b/c, for one thing, I don't think most of this kind of thing should matter for college admission in the first place, and second, I'm generally more politically sympathetic to precisely the kinds of people whose applications I would find culturally alien. I want them to let them in, but if I'm forced to evaluate them on all kinds of qualities beyond their academic achievements, I would look at their applications and think, "You bred cows and collected canned food at church, while this kid from Westchester bred recombinant DNA and started a program to teach kids in the Bronx calculus." Whom would you pick?

Phoebe said...

Yeah, I'm not sure I'd be the best example, because I might wallow in native-New-Yorker UMC self-hatred and pick that which seemed least familiar, most agricultural. And because, like you, I don't think these things should enter into admissions, and would be tempted to try and ignore them. Because what does any of it mean? Kids who've achieved great things, up to a certain age, maybe did, or maybe their parents signed them up to do something productive with their summers.

It would seem, though, that if schools were aware of this, and still set on factoring in extracurriculars, they could look more straightforwardly at leadership, initiative, etc. Did the rural applicant just collect cans, or organize the drive? This would of course leave the question of whether leadership at 15 is indicative of anything particular (and I sure hope not), but it's better than nothing.

Andrew Stevens said...

Of course, Phoebe is on to the real reason for holistic admissions.

Because what does any of it mean? Kids who've achieved great things, up to a certain age, maybe did, or maybe their parents signed them up to do something productive with their summers.

What it means of course is the people who came up with this system can ensure their own kids qualify for the most selective universities. I.e. the whole point of holistic admissions is to bias it towards the upper middle class, since poor kids don't have anybody telling them how to game the system. If you went strictly on academic and intellectual merit, you'd have way too many riff-raff.

Phoebe said...

Andrew, there's much to that, but it's not the whole point of holistic. There is an aspect of this that's affirmative action. An aspect that's a holdover from the days of keeping Jews out. A really huge aspect that's about appeasing rejected applicants and their neurotic parents - assuring them that they were not quickly dismissed by a computer, but hand-assessed by a concerned professional. Also, probably, as I believe MSI suggests, an aspect that's about contributing to the prestige of the schools themselves.

Miss Self-Important said...

Phoebe: Well, someone like you must be on the admissions committee at Harvard, b/c they do find Arkansans etc to admit each year (and I don't think it's b/c their staff is Arkansan). One other thing that occurs to me specifically w/r/t White Rural 4-H-ers is that I suspect (based on no data, but this was also one of my objections to Unz's finding that whites were being discriminated against in elite admissions) that they don't apply to places like Harvard in the same numbers as Boring Asian Nerds, so the opportunity to discriminate against them for their lame-sounding extracurriculars may be smaller than the opportunity to exclude Boring Asian Nerds who aren't sufficiently "involved" in out of school activities. If elite schools need some token Rural White 4H-ers and the pool of options is small, they can more easily swallow their confusion about what cow breeding is and take them, while there is no shortage of Asian applicants among whom to prefer the athletic or artistic or socially-involved ones.

Andrew: Sort of, but you're conflating two things - one is why places like Harvard adopted holistic admission (in the 1970s), and then why less selective and public schools followed suit (in the 1980s and '90s). In the latter case, it's mostly b/c the Supreme Court told them they had to if they wanted to continue to factor race into admissions (and they did want to). In what I imagine is the vast majority of college admission decisions in the US, holism hardly comes into play. The University of Illinois, for exampled, posted GPA and ACT cut-offs, and if you met them and didn't have a criminal record, your admission was basically assured. For more marginal students (as in the Fisher case) "holistic" factors can come into play, but that's the small minority of admissions cases at UIUC every year (well, plus the state government corruption admits).

The Harvard situation is more complicated, and seems to be the result of both genuine efforts to diversify the student body so that it was not comprised entirely of Bostonians and New England prep school graduates, and also ham-handed attempts to keep it from getting too broad. But the bias in favor of children of UMC professionals is still, I think, largely the result of perception and not a concrete policy of discrimination. The people on the admissions committees are children of UMC professionals, and they are drawn to applications that look like their own aspirations. If poor or rural kids knew better how to appeal to the sentiments of these types of people, to game the system, as you say, they'd probably have a better chance, but wouldn't that be true in any such situation where holistic evaluation is demanded? If the admissions committee were a bunch of Iowa farmers tasked with providing "holistic evaluations" of its applicants, wouldn't you expect to see a lot more church-volunteering, FFA Midwesterners on campus the following year?

Andrew Stevens said...

MSI: I was only really talking about the elite colleges here. I agree with you that the lower tiers followed suit for the reasons you gave and that they don't really matter since they are clearly open to the lower classes (though obviously still biased toward the wealthier, but to a certain extent that really can't be helped). But the top Ivy League has always adopted the various policies they have from 1800 all the way to today in order to attract wealthy applicants and/or applicants who are clearly going to be wealthy (i.e. the children of the UMC). That's why they were excluding Jews in the 1920s, why they were excluding Asians in the 1980s, and why they have excluded other poor people always.

It also pleased them to support the various civil rights movements so they adopted some tokenism as well, including geographic tokenism. (This is why they always accept a couple of Arkansans every year, but never more than a couple.) This was done, I am sure, at least partly out of a sense that it was the right thing to do and partly out of a sense that it was the economically prudent thing to do. Once their donor base had reached a critical mass favoring such an approach, they had to adopt it to keep the donations flowing. But they also couldn't afford to go too far in either the diversity or merit direction or else the children of their alumni wouldn't stand a chance of admission.

Of course, I also grant that there are pressures on their admissions policies to never lose sight of merit as well, namely competitive pressures. If they don't get their fair share of the really top students, their reputation might decline and it might be Yale, Princeton, and Brown, or something.

Miss Self-Important said...

So this is a question that, as I mentioned, I had about Unz's original claim that admissions committees were discriminating against non-Jewish white applicants. You seem to assert the same thing when you say they won't take more than two Arkansans a year. But here is my question - how many Arkansans are actually applying to Harvard and its peers? Are poor and rural whites underrepresented b/c admissions committees are against them, or because there aren't even enough of them applying for the admissions committees to hate?

Again, I have no data on this, but my hunch is that a much smaller percentage of the top students from Southern and Western states apply to Ivy League schools than similarly qualified students from Northern and Eastern states. In part, this is b/c high-scoring poor students on the whole apply to top schools at much lower rates, and in part because of the regional prestige of certain Southern colleges (SMU, Ole Miss come to mind) with which these good students would actually be more familiar than with places like Swarthmore (of which they wouldn't hear) or even Harvard (which they've never been within 500 miles of). Duke and Vanderbilt are traditionally Southern schools whose prestige has gone national, so to speak, but there are many schools that most Northerners have never heard of that are nonetheless impressive to Southerners and would seem to draw down the numbers of top Southern applicants to the Ivy League.

But, if Harvard gets 20 Arkansan applications each year and admits two of those students, that's a higher rate of acceptance than for the general pool. So is there any way to know whether that's happening?

Andrew Stevens said...

I agree that Southerners are not going to apply at the rates of New Englanders. (Though California seems to be well represented at Harvard, in spite of Stanford. It's only the South and the Midwest which are woefully underrepresented.)

See this article.

Particularly this quote:

"Why does each class appear statistically like the one ahead of it, if 'quotas' are not used? There seem to be two answers. First, each man on the admissions committee feels in his own mind that a 'good mix' is a necessary thing for Harvard College. If the admissions committee has just okayed nine consecutive students from a small town in Oregon, it will become wary of admitting more. Perhaps, as Whitla suggests, the advocate himself will not be able to find it in him to argue a tenth case enthusiastically. More important, there is something of a quota built into the admissions process. This is the docket system. Applicants are divided into 22 dockets, according to the secondary school the student attended. Far from Harvard, the docket divisions are large geographical units; Docket B, for example, is The Rockies, and G is Ohio and Kentucky. But further East, Docket K is Cambridge, Docket I is entitled Andover and Exeter, Docket P is Boston Public Latin, and New York City's public schools have a docket of their own, separate from the metropolitan area's private schools. Peterson calls it 'a New Englander's map of the United States.'

"After all the numerical parameters are available, a computer examines the evidence and announces an approximate quota for each docket, based on how many were admitted from this docket last year and how the numerical evaluations of this year's students in the docket compare with this year's applicants overall. Thus, if California's applicants suddenly become twice as good as last year (by the numbers) relative to the whole group, the computer will allot more places to California-but not twice as many."

I have no doubt the docket system exists because of the quantity of applications from each region, but it's also going to organize the thinking regardless of how many applications come in from that region on a year-to-year basis.

Miss Self-Important said...

I'm not sure the dockets look like that anymore, if they still exist. But I remain confused about what you hope to see happen - more applicants from underrepresented areas, or more acceptances for those who do apply?

In the first scenario, everyone should aspire to Harvard, and Harvard should select the best academic prospects among all high school seniors, not only the best among its actual applicants. This would still result in more Northeasterners, Californians, and Upper Midwesterners than anyone else b/c they have better schools, but not necessarily proportionally more of these, since these areas also have the largest populations.

But why would we want that scenario where Harvard is a proportional map of America? All the smart students in the country should not go to Harvard, and Southerners' current bias towards their own schools is, on the whole, a good thing for the decentralization of opportunity so long as they persist in seeing places like Sewanee as good schools. For the same reason, I don't think it's a big problem for Harvard to give over a disproportionate number of spots to Bostonians and New Englanders, because Harvard is a Boston and New England school, and it owes more to its immediate neighbors than to Oregonians.

In the second scenario, we're dealing only with the people who already choose to apply to Harvard and the claim that those from poor or rural backgrounds are discriminated against. This is the area where we need further evidence, because as of now, we don't know how many people from underrepresented areas are applying or how many are being rejected, and what their qualifications were. I would obviously agree with you here that bias against those already applying is bad, but I'm more skeptical of the arguments that more good students ought to be encouraged to apply to Harvard if they are otherwise disinclined.

Andrew Stevens said...

MSI: I honestly couldn't care less what Harvard's admissions look like. They're a private school so as far as I'm concerned they can admit whoever they like however they like. My objection is only to any perception that their admissions policy is some sort of model of meritocracy or social justice, when it has always actually been aimed at making money, merit and social justice be damned. I am not arguing that their admission policy should be a model of merit or social justice. A great many people, however, are very confused about this, though I realize that you're not one of them.

Anonymous said...

"it's not enough to insist that the school is X percent white/black/Hispanic/Asian to demonstrate nondiscrimination, since we can't know what percent it should be without knowing who was rejected and why"/"how many of the total the Asian applicants should have been admitted - is precisely what we cannot know simply from the number who were admitted""

This is a non-sequitur though because there is no number that "should" of been admitted, not when there is more applicants than there are seats.

In Canada I would be hard pressed to find someone even someone who was accepted to a top university that was required to list their extracurriculars, sometimes there is a optional essay component which can be used to list accomplishments but still more program specific reasons why one would make a good candidate not really a life story thing.

Miss Self-Important said...

Anonymous: But then the Crimson's claim that affirmative action is not "resulting in fewer Asians than there would be otherwise" must also be a non-sequitur, since there is no "otherwise" in a situation where there are 20 times the number of applicants that the school can admit, there is only the fact that everyone is equally undeserving of the divine grace that is admission. What this Atlantic article claims must also be a non-sequitur - we can never know if Fisher would have been admitted, no matter how many ways we compare her application to the others that year. But that's not really true - human beings sit on admissions committees, they make the decisions, and they at least claim not to make them arbitrarily. So we can at least know why people were not admitted, no?

Yes, Canadian universities are different. Better, worse, I don't know. I don't think the holistic paradigm has done American universities much good, and there are appealing aspects of the Canadian school system that demonstrate that holism is unnecessary for finding good students. But there are many ways that the American system differs structurally, not just at the level of admissions policies, the most notable of which is probably the importance of private colleges. I don't know enough about Canada to know whether we can simply adopt its policies without changing our structure.

Anonymous said...

I actually agree that there should be more transparency when it comes to admissions policies as practiced today, especially considering that when it comes to affirmative action one of the most contentious issue of our time, most people have no idea what it is and how it works I myself only have a foggy idea i.e making up for past discrimination by making sure African Americans are more represented at any given school but apparently the Atlanta article and others like it would indicate that's not quite right either.
What I was saying was more about your other topic that's in there, that there is this general anxiety in the public over who is unqualified and getting in and who is being unfairly crowded out, that is not actually a issue to be worried about or write articles about, because when obviously wrong and unfair things like racial discrimination are accounted for (Which is why Fisher's case was such a big deal/whether not affirmative action is discrimination) who gets in and does not is up to the discretion of institution, if someone meets all the criteria and then is rejected (barring discrimination/being a farmer or from Indiana is a unconvincing allegation in my book) not in the least bit unfair because private universities are businesses and it is well within their right, or sovereignty if you will to reject who ever they want, for what ever reason or for no reason at all
On the contrary it seems to me that there is this perception floating about that is the cause of so much irritation, that people seem to believe if one does everything their told to do by said institution than they now possess this divine right endowed by the creator and solidified by the countries founders to be let into said institution when they do not one is not wronged or harmed by not being let into a certain university, nobody with excellent credentials was ever turned away from Harvard and then unable to go some place else, this perception that it is a tragedy of tragedies, one which will never be rectified in all your years to be rejected from top schools is so prevalent it is what made Fisher decide to sue not really whether or not she cares if white people (other than herself) are being discriminated against.
I don't know why this perception is so prevalent I sure hope it's not because America's post-secondary is a lot like their public and private high schools..