Unz Ivy Stats Flashback

This news story reminded me of some threads from a few years ago about Ron Unz, the political activist who wrote a statistics-filled article a few years ago claiming that Harvard and other Ivy League colleges discriminate against Asian-Americans and in favor of Jews in undergraduate admissions. It turned out that some of his numbers were off by factors of 2 or 4 or more, and then, amusingly, or horrifyingly, Unz offhandedly remarked that one of his more high-profile statistical claims (cited in a notorious New York Times column by David Brooks) had been based on “five minutes of cursory surname analysis.”

Unz really pulled the rug out from Brooks on that one! The world-famous Ted Talk speaker and humility expert was too proud to retract his column, so Unz had left him to twist slowly in the wind. (Hence the image above; if we’re gonna mix metaphors I might as well go whole hog.)

In all seriousness, I remain upset by the way that false claims can stick in the public discourse. Nothing new here, of course (insert Twain-attributed quote here) but particularly sad in this case in that I expect the numbers in question were not lies but rather mere mistakes.

At some point I have more to say about this case as an example of how certain intuitions can lead us astray—believe it or not, I see some connections between the reasoning of Unz and Brooks and that of authors of various unreplicated papers in social psychology—but for now, I’d just like to help out anyone who’s coming to this particular himmicane in the middle (for example, after reading the above-linked article in today’s paper) by pointing you to my previous post on the topic, from a couple years ago, which contains links to several earlier posts and some long discussions on the details of the case.

15 thoughts on “Unz Ivy Stats Flashback

  1. Andrew,

    You are, of course, absolutely correct about Unz’s mistakes when it comes to the issue of Jewish admissions. But what are your thoughts on discrimination against Asian-Americans? Unz and others believe that at least 10% more of the students you teach at Columbia would be Asian American if your university did not discriminate against them in admissions. Do you think that they are right?

    • David:

      That I don’t know, I haven’t looked at it. Perhaps as with the Jewish admissions, people will email me with additional information. I didn’t do any original research for that either, I just did my best to make sense of everything I read.

      • Andrew,

        I am somewhat bothered by your admission that you haven’t bothered to look into whether Asians are discriminated against in admissions in elite institutions, like Columbia.

        The discrimination against Asians in university admissions is actually THE primary message of the Unz piece. Yet somehow the discussion point that gets picked up by the media and blogosphere is instead “who gets favored?” I don’t think you need to be Asian to see the unfairness of this change in focus. As long as the question is about whether specific subgroups receive preferential treatment, there is little discussion of “Is there bias against Asians and whether/how one should do anything about it?”

        Now hey, this is your blog and you can write about whatever you feel like. But this seems quite representative of the criticisms that people have levied on post-publication peer review regarding some notion of “fairness”. I think often those are made for self-serving reasons, but there’s often some truth there as well. The fact that you never said, “I don’t agree on the overrepresentation of Jewish students, but the overall point of systematic bias against Asians seem important and worth studying” is telling.

        P.S. I wonder if you think that Unz is using Asian discrimination as a trojan horse to pick fights against certain groups. I have no idea what Unz’s intentions were, but these things seems to happen often. But somehow I doubt that since you don’t seem the type (at least from this blog) to hold back from these views.

        • Tul:

          I really have no idea. I only found out about the flaws in Unz’s other calculations because some people emailed me about it. This is not an area where I’ve ever done any original research.

          Also, you fault me for not writing, “I don’t agree on the overrepresentation of Jewish students, but the overall point of systematic bias against Asians seem important and worth studying,” but I actually did write this on my blog: “let me emphasize that Unz’s statistical mistakes do not necessarily mean that all of his ideas are wrong or meaningless. There certainly have been large demographic changes in the United States in recent decades, and the result is increasing academic competition. These things are worth studying.” So I think your criticism is misplaced.

  2. Understood. The issue of discrimination against Asian-Americans in admissions at places like Columbia is an interesting one. It is also a topic that you might be uniquely positioned to talk about. Perhaps start with this graphic?

    I suspect that there will be many more news stories on this topic in the year ahead. You writing about it (and soliciting comments from your excellent readers) would be very useful, just as your analysis of Case/Deaton helped the public conversation last month.

  3. Andrew:

    I taught high school for many years in California, and my Asian American students who applied to Berkeley were disproportionately interested in Engineering, and hence disproportionately applied to admission to Berkeley’s School of Engineering, which is more selective than Berkeley’s School of Letters and Sciences. I don’t know whether the Ivy League universities that Unz looked at handle admissions in a similar fashion, but if they do, then analysis of admissions rates would have to control for students’ intended major / school to which they apply. Given what you have said about Unz’s analysis, I can’t imagine that he did that.

    PS: I found it odd, but sadly not surprising, that the Times article did not mention that Harvard is already free for students whose make under $65,000 https://college.harvard.edu/financial-aid/how-aid-works/harvard-financial-aid-initiative

    • Quite right, we need to look at within-department numbers. The smoking gun would be evidence that Asians in engineering departments come in with higher SATs than white students in those same departments.

      Otherwise, the “stagnating” Asian enrollment reflects nothing more than colleges’ wholly legitimate interest in maintaining thriving departments in the humanities, which Asians are so far less interested in.

  4. University admission requirements can be arcane. My university has for many years had a “top x% admissions policy,” but the interpretation is not straightforward, and seems to morph over time.

    A number of years ago I encountered the following case: A student in Engineering really wanted to be a math major. He had moved from an Asian country to the U.S. when he was in middle school and knew no English, so his grades in English and History were poor enough to keep him out of the top x%. However (at that time) the School of Engineering had more applicants in the top x% than they had slots for. This meant that they were excused from automatic admissions, and could use any admission requirements they chose (perhaps subject to some rule such as that at least 75% of their admissions had to be in the top x%). This kid had been top performer in the state math competitions, so Engineering gladly admitted him — but since the college of Natural Sciences was subject the The Rule, he could not have been admitted by Natural Sciences (which includes math).

    Fortunately, once admitted to the University, students could transfer to any other major provided that major would accept him — but this was not widely announced. When I told him he could transfer to math and how to go about it, he was delighted.

    So admission can sometimes be a matter of knowing the loopholes and how to use them. Thus one needs to consider the details of policies, and how widely the details are known, as well as raw numbers in considering discrimination.

    • Why are the requirements so arcane? Bug or feature?

      Is the complexity driven by the desire to be subjective without appearing to be so? e.g. To squeeze in donor’s children?

      • Rahul: I think the reality is more like the old joke, “A camel is a horse put together by a committee.” Similarly, many bills passed by congress have arcane aspects.

    • Eli:

      I’ve thought (and posted) a lot about this. I expect Unz is like many of the well-intentioned researchers whom we’ve criticized, that he feels he has discovered the truth and that statistical criticisms, a factor of 2 here and a factor of 5 there, are just technicalities, a smokescreen created by his political enemies to distract from his big point. Marc Hauser probably felt that way too—that the small-minded haters couldn’t disprove his big theories so they were just yapping at his heels on procedural grounds. David Brooks and Jon Entine too. Even Weggy, for that matter. Or Stephen Jay Gould. It all becomes a matter of politics and then they can close their ears. It’s soooo frustrating. It’s the way the world works, but it still frustrates me a lot.

  5. Calling him a “political activist” in the initial sentence seems unfair, since it is obviously meant to bias readers in the direction of disagreeing with his findings.

    I realize that you can reply “but its true that he is a political activist”, but its also true that descriptions which are technically correct can also be used in misleading rhetorical ways. Consider, for example, an article that discussed your disagreements with Unz that started with “Andrew Gelman, a Jewish professor at an elite ivy League institution….”. Is it true? Yes. Is it a fair way to start an article? Obviously not.

    • Steven:

      This came up in an earlier thread and here’s what I wrote at the time:

      it’s not an insult at all to call someone a political activist. That’s what Unz does! He’s run for office, he’s funded political campaigns, he bought a political magazine. There’s nothing wrong with being a political activist. It’s a noble calling.

      And here’s the first sentences of Unz’s wikipedia article:

      Ron Keeva Unz (born September 20, 1961) is a former businessman and political activist . . .

      I suppose I could’ve added “former businessman” to Unz’s description but I don’t think it would’ve changed anything. “Former businessman” didn’t seem so relevant here (although it would’ve been relevant if I’d been commenting on Unz’s opinion on the corporate tax rate or whatever). I called Unz a political activist (and I could’ve called him a former businessman) because that’s what he does, it’s not meant to bias anyone.

      By the way, here’s the first sentence of my wikipedia article:

      Andrew Gelman (born February 11, 1965) is an American statistician, professor of statistics and political science . . .

      I’d have no problem if you were to refer to me as an American statistician, or as a professor of statistics, or a professor of political science. And, for that matter, in the present discussion my employment and ethnic background are arguably relevant too, so feel free to bring them up. Whatever. It doesn’t really matter. The numbers are the numbers.

Leave a Reply to Martha Cancel reply

Your email address will not be published. Required fields are marked *