Meritocracy: Admitting My Mistakes
In publishing a 30,000 word article covering such a broad range of complex and controversial topics, I was certain that my work would necessarily contain at least a few factual errors or omissions. The hundreds of individuals examining my material over the last three months have located several, and being from an academic background, I am happy to recognize these:
- On p. 20, I carelessly described America’s yearly Math Olympiad teams as having 5 members, and a sharp-eyed former Olympian noted this was incorrect. Indeed, over the last 40-odd years, the teams have ranged from 6 to 8 members. My actual calculations used the correct figures and remain unchanged.
- A central core of my analysis relied upon state NMS semifinalist lists, of which I managed to locate 43 on the Internet. The legion of commentators who reviewed my findings subsequently managed to locate a 44th, a Massachusetts list that was published in the Boston Globe during 2008. I have analyzed this additional list and appended the results as an addendum to my Appendix E. Incorporating this additional data produces no significant change in any of my national estimates.
- On p. 22, I described the five most selective UC Campuses as being Berkeley, UCLA, San Diego, Davis and Irvine, and analyzed their aggregate ethnic distribution. Commenters have persuasively argued that the Santa Barbara campus is at least as selective as Davis and Irvine, so it should have been included in my analysis. Doing so produces no significant change in my overall results.
- Near the beginning of my article I had noted that although complaints about official corruption of every sort are a leading topic on the Chinese Internet and also in Western media coverage, I had never once heard such a claim about admissions to elite Chinese universities. This led me to conclude that the process was entirely meritocratic, and a couple of individuals with good knowledge of China confirmed this. However, during one of my recent Yale Law events, a student from China stated that he and his friends were firmly convinced that any of China’s 350 Central Committee members could easily obtain an admissions slot for his friends or relatives, so my claim was incorrect. This conflicting evidence may be reconciled if the number of such corrupt admissions each year is so tiny—perhaps a few hundred out of over eight million—that it is completely invisible to the general public. I should note that the New York Times just ran another major story on colleges in China, emphasizing every possible unfair aspect of the system, but nonetheless indicating that admissions were entirely meritocratic and objective.
I fully acknowledge all these unavoidable errors in my work. But the recent, widely-distributed criticism presented in the main post and lengthy comment-thread of Prof. Andrew Gelman, a prominent Ivy League statistics professor, falls into an entirely different category.
On pp. 26-27, 30 of my article, I had noted strong evidence of a sharp decline in Jewish academic achievement at the very high end, as indicated by the large drop in the number of likely Jewish names among the winners of top math and science competitions during the last six decades, and this is correct. As an example, I mentioned that during the thirteen years since 2000, just two of the 78 names of Math Olympiad winners appear to be Jewish, and this is also correct. However, the extensive personal and biographical research of Prof. Janet Mertz, one of my sharpest critics, has determined that “Daniel Kane” is actually a full-Jew, “Brian Lawrence” and “Alison Miller” are half-Jews, and one or two other corrections. As a result, the decline in actual Jewish performance was from over 40% during the 1970s to 12% during 2000-2012, rather than to the 3% figure I suggested.
Obviously, large statistical errors are unavoidable when simple surname analysis is applied to such tiny sets of names. If Prof. Mertz were willing to extend her exhaustive research methods from the several dozen individuals she investigated to rigorously determining the precise ethnic background of the more than 4,000 winners of Olympiad, Putnam, and STS competitions from the 1930s to the 2000s whose surnames I examined, I would be very interested in seeing her findings.
In my opinion this one small point, namely the precise number of Jews or part-Jews in the 2000-2012 Math Olympiad lists, seems to be about the only substantial and verifiable charge made against my analysis in the entire lengthy critique. A crucial part of the critique consists of the claims of an anonymous individual calling himself “NB,” which are based upon his private analysis of non-public data and cannot be externally verified.
The angry criticism of Prof. Mertz and “NB” had been floating around the Internet for some time, and had been widely ignored or dismissed. For unaccountable reasons, Mertz and “NB” had made no effort to publish their critique anywhere, even if merely on a personal blogsite. However, last week their charges were excerpted and presented by Prof. Gelman’s on his own blogsite, thereby lending them strong credibility. One additional source of confusion was that since Gelman was providing a platform for the claims of an anonymous individual, it was sometimes not entirely clear whether he was personally confirming or merely reporting some of the critical remarks in question.
I directly responded to some of these charges in the resulting comment-thread, which grew very lengthy and sometimes heated. Nearly all the other participants remained anonymous, which hardly improved the tone of the debate. With well over 100 often lengthy and multiply-nested comments, the discussion eventually became completely unreadable.
Therefore, as a means of resolving my disagreement with Prof. Gelman, I sent him a private email presenting my own framework of estimated results and suggesting that he do the same, thereby allowing us to determine the range of our agreement and disagreement. He responded and suggested that he publish my note as a new posting, followed by his own response, and I agreed: our exchange appears here. Readers should note that Gelman’s response emphasizes that he had not himself studied any of the matters in question, nor had he formed any opinions about what the correct figures might be, but that he was merely passing along the critical claims of his anonymous Internet correspondent. I find this exchange highly enlightening, and urge individuals to read it. Among other aspects, I believe my own note provides a very useful framework of discussion for anyone who wishes to quantitatively challenge my findings.
Gelman’s revised position led one of his agitated commenters to become concerned, urging him not to”run away” from the “fight.”
For whatever reason, Gelman soon posted a follow-up column mostly attacking David Brooks, but also now once again dismissing me as a “political activist” who used “sloppy counting.” Presumably, someone who was less “sloppy” would have immediately recognized that “Daniel Kane,” “Brian Lawrence,” and “Alison Miller” are obviously Jewish names. I find it highly intriguing that although Gelman chose not to substantially engage the 1,000 word framework of my statistical analysis that I offered him for constructive mutual dialogue, he instead preferred to return to the use of insults, while focusing on a single phrase in my 30,000 word analysis, a phrase characterizing a sample of negligible size. Furthermore, the phrase in question was accurate, unless Gelman actually believes that names such as Daniel Kane and Brian Lawrence “appear to be Jewish.”
Individuals who become emotionally involved with a particular position of ideological or ethnic advocacy may lose their ability to dispassionately analyze data, and this intellectual failing may sometimes even apply to award-winning Ivy League statistics professors. Meanwhile, litigators who choose to completely ignore the overwhelming volume of the facts in a case but spend all their time angrily pounding the desk on an insignificant one hardly demonstrate the strength of their position.
As it happens, I just returned from a speaking engagement at the University of Chicago Law School, at which I argued that the strong evidence I have demonstrated for the existence of Asian Quotas in the Ivy League may indicate that the landmark 1978 Bakke verdict was decided based on fraud. The response from the large audience—which included blacks, Hispanics, whites, and Asians—seemed overwhelmingly positive. However, one or two of the audience members afterward came over and argued with me at considerable length about my Jewish findings, even though these had not even been mentioned in my talk. Highly-emotional ethnocentric zealotry should be distinguished from neutral scientific inquiry.
On a different matter, I recently read Education and Politics at Harvard, written in 1975 by renowned sociologists Seymour Martin Lipset and David Riesman, and although most of the material was not new to me, I came across one remarkable item. Conservatives often ascribe any substantial decline of average student quality at top universities to the role of affirmative action in admitting underqualified minorities, and this belief is widely—if quietly—shared in certain quarters. Therefore, it is always helpful to locate an independent check of its validity.
On p. 319 of the Lipset/Riesman book, Riesman casually notes that each year Harvard enrolls some 700 National Merit Scholars (NMS), who constitute America’s highest performing high school students. But as I mentioned in my article, Harvard’s NMS total had fallen to 396 in 2002 and just 248 in 2011. Thus, over three decades Harvard’s number of such top students has dropped by almost two-thirds even while the volume of Harvard applicants has grown enormously. This is certainly consistent with the recent claim of a former Harvard Senior Admissions Officer that only 5% of students are admitted these days purely based on academic merit.
Could a major part of this decline be due to the impact of affirmative action policies? Nearly all NMS are either white or Asian, and in 1975 Hispanics, blacks, and foreign students probably totaled about 10% of each class, while today the figure is over 25%. So perhaps that explains at least some of the general decline in NMS numbers.
However, we must also consider a different factor. Asian enrollment was completely negligible in the 1970s, but over the last decade has probably averaged nearly one-third of the combined white plus Asian total. Also, there is overwhelming evidence that in recent years, Harvard has imposed sharp restrictions on its Asian numbers, requiring Asians to achieve much stronger academic results to gain entrance. Therefore, it seems likely Harvard’s NMS are disproportionately Asian, perhaps half or more of the total rather than merely one-third.
Thus, even if we assume that every remaining NMS is white, the trajectory of white student decline is quite remarkable. During the 1970s, Harvard enrolled 1400 white Americans, of whom 700 or 50% ranked as NMS. By 2002, the number of whites had dropped to around 850 with just 25% being NMS. And as of 2011, only 15% of Harvard 800 whites seem to have achieved that national academic distinction. Obviously, none of this striking decline in the average quality of Harvard’s white students has any connection to “affirmative action” in the usual sense of that word.
Finally, in my companion sidebar I argued that Harvard had gradually become a hedge fund with a football team and a few professors, and suggested that since college tuition constituted a negligible portion of annual revenue, such charges should be eliminated. Although it makes perfect sense for Harvard, Yale, Princeton and other wealthy “hedgefundiversities” to abolish their tuition, the actual American trend seems in the other direction.
For example, earlier this week the New York Times revealed that Cooper Union, which has provided New Yorkers a free, top-quality education for the past 153 years, had made a decision to begin charging tuition. The problem was that income from its valuable property holdings no longer covered expenses, and the school’s annual loss had reached the unsustainable level of $12 million per year.
What were those expenses? Well, during the 2000s the university had taken out a $175 million dollar loan and invested the cash in the stock market just before that the financial collapse; repayment on that loan now runs $10 million per year. In addition, Cooper Union had recently spent $177 million dollars to erect a hideously ugly but “audacious” modernist-style building as part of its major campus expansion. So future Cooper Union graduates may be forced into permanent debt-peonage, but they will have received their education in a provocatively-designed building.
We must consider the differing interests of a university’s students and its top administrators. Students might prefer an education that is free rather than one that costs $10,000 or $20,000 per year, but why should a university president care? Indeed, the larger the total budget and more expensive the cost to attend, the stronger the case for raising the salary of the president. And while the direct educational benefits of a $177 million avant-garde architectural centerpiece are probably nil, the enhancements to the cocktail party chatter of the individual who made the bold decision to build it might be enormous.
Such is the pattern when the interests of rulers diverge from those of the ruled, and the former show no concern for the latter.