White Flight From Admissions-Test High Schools

I’ve written this approximately 5 billion times on Twitter, but Razib Khan’s review of Charles Murray’s new book, coupled with my recent failure to create any articles at all, combined to convince me to put it in essay form.  The point has nothing to do with the book, which I haven’t read yet but is certainly excellent, or the review, which is perceptive, particularly the last bit.

But this part here is often repeated and quite misleading:

In New York City, the elite Stuyvesant public high school illustrates the discrepancy between the quality of our discussions of systemic racism and how race, class, and education actually interact in the real world. Admission to the school is based on a standardized test. The current student body is 73 percent Asian American and 19 percent white. New York City is about 15 percent Asian American and 30 percent white. In other words, white kids are under-represented in the student body (though far less so than blacks and Latinos).

John Podhoretz mentioned the same stat a few months ago:

If you add whites, blacks, and Latinos together, they will constitute around 37 percent of the kids at these eight schools. Now take a wild stab at the ethnic origins of the absolute majority of admits—a stunning 53.7 percent in all. You guessed it. Asian.

This is the case even though Asians make up a little less than 12 percent of New York City’s population. Black people make up 26 percent. White people make up around 26 percent. Latinos make up around 26 percent. And Asians? Around 12 percent.

The stats are all true, but there’s one big problem: both Khan and Podhoretz use the wrong stat. It doesn’t matter how many whites and Asians are in New York City, but how many of them are in public schools. 

White kids are twice the absolute number of Asian kids in the New York City general population, but public school population is a different matter:

    • 40.6 percent Hispanic
    • 25.5 percent black
    • 16.2 percent Asian
    • 15.1 percent white

These stats have been mostly consistent for a decade or so; the NY Times correctly lists the populations back in 2010, when whites were 15% and Asians14%.

So Razib and Pod use the fact that whites are twice as populous as Asians to demonstrate the shocking disparity in the specialized school population. But in fact, whites and Asians are equally represented in the public school population, so the disparity is already half as bad as they say. 

But there’s still more data that is rarely mentioned in the white/Asian disparity debate–Podhoretz does bring it up, but as usual misses the relevance. Given that whites and Asians represent equal populations in public schools, it’s worth looking at the attempts and admit rates:

NYCDOE on 2020 SHSAT results

The dark blue line at the bottom shows the acceptance rate of testers by race. Notice that whites and Asians have basically the same admissions rate (caveat: the thousand multi-racial or unknown race have a very high acceptance rate). Asians probably have higher scores on average than whites, as more of them get into top-ranked Stuyvesant. Given the much higher, one might say obsessive, Asian dedication to test prep, this isn’t surprising. New York City test scores show 3rd through 8th grade white and Asian ELA proficiency rates the same in English, although 77% of Asians test proficient in math and “only” 66% of whites do the same. I couldn’t find granular data for 8th grade in NYC only, but suffice it to guess that they test at roughly parity in ELA and Asians do better in math.

Point being that whites and Asians both have high proficiency levels on state tests and identical admissions rates to the specialized high schools.

However, take a look at the top line of the graph, showing overall testers by race.  Despite their overall population parity, whites and Asians have very different interest level in the schools. Twice as many Asians took the test as whites. 

This reporting tool reveals that in 2019, at least, there were 12,769 Asian and 11,147 white 8th graders.  So about 64% of Asian and 39% of white 8th graders took the SHSAT.

Well, that’s…..weird. 

It’s almost like whites aren’t nearly as interested in attending the “best public schools in New York City”. 

This thoroughly screws with the traditional conservative take on merit-based high schools, which goes like this: “Asians work hard, study hard, care about school, and that’s why they dominate admissions in the best schools.” 

How come whites aren’t testing–that is, applying–to the “best” schools? 

In an article that almost certainly couldn’t be written today, the Atlantic goes through the history that explains why whites are less interested. First, a very interesting graph on Stuyvesant student population over a 20 year period:

Atlantic also did populations for Bronx Science and Brooklyn Tech

Notice that black and Hispanic population has been cut in half, but white population has dropped by more than half. If Stuyvesant admissions are truly test-based, as I’m sure they are, that shouldn’t happen. Only at Brooklyn Tech did white population increase, and only slightly.

The Atlantic points out that the drop began in 2002, when mayor Michael Bloomberg instituted school choice:

The white population at Stuyvesant hovered around 40 percent from the late 1980s until the early 2000s, according to the National Center for Education Statistics. Around 2003, when Bloomberg became mayor, the number of white kids at these schools dropped as the number of schools that screen for academic criteria like grades or exams, or require an audition or interview, more than doubled.** This selectivity increased the pool of schools that were considered “good,” which diverted many white students away from the specialized schools and into these newly prestigious schools.

So when Bloomberg allowed all public schools to screen for academic criteriam, suddenly, whites weren’t interested in Stuyvesant and the other specialized high schools. (Note: while elite private schools do suck up a lot of bright black and Hispanic kids, a lot of the others attend these “choice” high schools.)

White NYC public school students aren’t failing in their efforts to get into the specialized high schools. They aren’t interested. They aren’t applying. They have schools that are just as good.

While the national status of the specialized high schools is unquestioned, it’s pretty clear that the local status is much diminished. Asian students are overrepresented in these schools not because the schools have high academic standards and desirable status, but because bright white, black, and Hispanic kids have equally good options that they find more attractive. Why bother with obsessive test prep and a school culture that Americans find unpleasant when you can set a perfectly solid standard that keeps out the low-skilled and unengaged? 

Look throughout the country at the test-based schools and you’ll find a similar pattern, although the SHSAT is the only test whose results are readily available online. But in every area with a mostly Asian test-based public high school, look for  an equivalently ranked school that’s mostly white, that also has more blacks and Hispanics. I don’t know enough about Virginia area schools to state this with a certainty, (I’m trying to write more, research picayune yet interesting details less) but this 2019 reporting reveals that whites comprised only 25% of the admissions to Thomas Jefferson High School, while they are surely more than 25% of the population.  In other areas, like San Francisco and Lowell, whites rich enough to live there pay for private school or move south or east in the megalopolis.

It’s simply inaccurate to refer to the specialized high schools as “New York City’s best schools”.  It’s totally wrong to argue, as most conservatives do, that Asian kids are “dominating” academics, or that the lack of black and Hispanic kids at these schools is a mark of a “failed education system” or even “black culture not valuing academia”.

The simple fact is that everyone involved in this debate has no idea what they are talking about. They see the specialized high schools as pure meritocracies. In fact, they are an expensive service NYC provides Asian immigrants who want to get their kids away from black and Hispanic students.

Now, some of this is my interpretation, so I’ll try and break it down.

Fact: white student interest in the NYC specialized schools is far lower than Asian interest, despite similar representation in public school population and similar acceptance rates to the eight schools.

Fact: when Bloomberg allowed public schools to restrict admissions, a large number of prestigious local schools set standards and this led to a drop in white student interest in the eight specialized high schools.

Opinion: whites are uninterested in the NYC specialized high schools because of the Asian immigrant culture.

Opinion: immigrant interest in any “merit-based” institution, coupled with the ability to build less competitive, but still exclusive, institutions leads to reduced white interest. 

Opinion: what the media refers to as “best schools” is not what American parents consider “best schools” if the schools are overwhelmingly populated by Asian immigrants and contain the competitive, grade-obsessed Asian culture.

The white flight away from test schools to public schools with test-in standards explains why white parents were freaked out when di Blasio threatened to end this option. DiBlasio can’t change the specialized high school requirements but he can end white parents’ ability to sculpt their own schools. If diBlasio did end these programs, I’d expect white participation in the specialized high school application process to double.  Alternatively, whites may decide it’s time to leave the city entirely, although that’s a different decision in New York City than it is when moving from San Francisco to, say, Palo Alto.

So the primary rationale for ending or changing test-based high schools is to make public schools more attractive to American (read white) parents.  The real driver isn’t equity. That’s just cover. District leaders might talk about the importance of black and Hispanic achievement, but  (as many have noted) ending test-based admissions always leads to more white students. Most of the parents screaming about the end to test-based schools are Asian parents–most of them immigrants, many making demands that might possibly be considered a tad ungrateful. Given a choice between a school that pleases newly-arrived Asian immigrants and one that pleases their tax base of mostly white but all high-income parents–and also improves black and Hispanic participation–which option is the best political strategy?

As usual, conservatives–and, for that matter, black opponents like Glenn Loury, John McWhorter, and Wilfred Reilly–get it all wrong. They are wailing about the injustice to the Asian immigrants and the threat to merit. They complain that black students aren’t encouraged to compete or achieve. But this is all bullshit. There are smart black kids. They have other options that they find much more attractive.

Asian dominance of specialized public schools has nothing to do with academics or high standards. It’s white flight. Districts are taking action to end white flight. This is unsurprising and something that conservatives would normally support except their loathing of public schools interferes with their ability to notice the obvious. 

I  shouldn’t have to point this out. But there’s one thing that I’ve noticed elites of either party are categorically reluctant to discuss, and it’s something that a (barely) upper middle class white living in extreme diversity is uniquely suited to observe. White people living around huge populations of Asian immigrants aren’t enamored with them. When you read that Asian participation in an institution or activity has soared, start by wondering whether absolute white participation has significantly declined. Then find out why.

Naturally, when elites do discuss this behavior, it is tagged as racism by conservative and progressive alike. Plus ca change, and all that.


There are many reasons my writing has fallen off a cliff this year, not least of which that remote teaching is far more work. But another key reason that I just figured out recently involves the huge push over the last year to end SAT and other test based admissions policy, an issue that would normally be right in my wheelhouse. But my opinions on all this have grown more complicated in the past five years. 

There are people who think I dislike Asians, and that’s just not true. I do have concerns about the huge influx in Asian immigrants since the 90s. (there’s a clear distinction between them and those who have been in the country for decades and multiple generations.) South and East Asian academic culture is rife with fraud and competition, and I don’t welcome its influence here in America. I grit my teeth and scowl at Americans* who talk about the Asian work ethic, Asian two-parent families, and all the other “positives” that they are just sure lead to Asian success (recently on display in Andrew Sullivan’s podcast with Wesley Yang). It’s all bullshit. Asian immigrants are people, and like all people they range from utter scum to totally amazing. But culturally, they have attitudes and values that are in total conflict with American ones, and Americans should not allow these values to override what makes American education great–and yes, American education is great. Nor has our education system ever been overly obsessed with grades, competition or endless study. We should seek to acculturate Asian immigrant education values, not reward them.

Leaving aside all the value-judgment (theirs and mine), I am also disturbed at the constant confusion between grades and achievement, as well as the utter ignorance of the real problems with the SHSAT scoring that would reward an Asian immigrant (or anyone else) scoring in the high 90s in math and the 40s in verbal while rejecting a black or Hispanic kid (or anyone else) scoring in the high 80s on both.

But none of my distaste leads to approval for college admissions offices discriminating against Asians by using enough of them to boost their metrics and dismissing the rest. Nor does it mean we can ban test prep. And using grades as a proxy for achievement, which is the actual means by which we are eliminating test-based admissions at both college and high school, is utter madness. Not only are grades basically fraud, but a grade-based admission process will encourage segregation, particularly for blacks and Hispanics. Black kids at integrated high schools have better scores than those at majority minority high schools, but lower grades. This is the Big Lie that runs all through the “top students at each school” admissions process. Grades favor low achievers. Tests don’t. 

However, I’m far more skeptical of the value of test scores past a cutoff mark than I was before around 2016. That is, an SAT score of 600 verbal or math has meaning. I’m not sure a 2400 SAT score does anymore. Well, it does, but not reliably. That’s a topic for another article. 

So tests, yes. But cut scores and a lottery, not ranking by score. That’s a topic for another day as well, but I wanted to be clear that I am unhappy–very, very unhappy–with the current move away from test-based college admissions. I believe we are destroying the value of a college diploma beyond redemption. It’s very worrying.

Once again, parts of this are more disjointed but focus is on writing to get past this perfection block, so bear with me.

*I mentioned this throughout, but unlike Asians, who often say American to mean “white”, I mean American as all races. If I mean American whites I usually say so.

About educationrealist

26 responses to “White Flight From Admissions-Test High Schools

  • dlr

    You should move to substack. Substack’s like button (heart) is right at the top, where it is easy to find! WordPress’s like button is at the bottom of the article, but hard to find, since it is followed by all the comments, etc.

  • White Flight From Admissions-Test High Schools – Nonpartisan Education Group

    […] White Flight From Admissions-Test High Schools — educationrealist […]

  • Atavisionary

    >Black kids at integrated high schools have better scores than those at majority minority high schools, but lower grades.

    Not white people’s responsibility to carry them.

  • Johan Larson

    We’ve been discussing this post over on Data Secrets Lox.

    • educationrealist

      I didn’t want to create an ID, but I think some of the readers are thinking that these selective schools are a step down from the 8 specialized high schools, and my understanding is they aren’t.

    • Johan Larson

      >It’s not as if colleges are terribly impressed with 80-90% Asian schools.

      Really? Why not? They seem terribly fond of elaborately prepped and tutored youngsters from other backgrounds. That sort of thing is pretty standard in well-to-do families these days. The stereotypical Asian upbringing is doing the same things, just harder.

      • educationrealist

        “Really? Why not? ”

        You’ve perhaps heard that they are unimpressed with Asian test scores and basically require a 300 point premium on the SAT? There’s a whole lawsuit about it.

  • theodorewhitfield

    “In fact, they are an expensive service NYC provides Asian immigrants who want to get their kids away from black and Hispanic students.”

    This is a cheap shot, and mars an otherwise thoughtful article.

  • Jay

    I’m certainly not going to argue that cheating isn’t rampant among some groups of Asians. I’ve known a lot of Asians over the years and I’ve seen some of them do it. But I’ll also note that Asian cultures* are more hierarchical than the broader American culture and places much more importance on rankings. I suspect that a significant part of the demographics of test-driven schools is that Asians are more likely to think that going across town to attend the #1 rated school is worth it and others are more likely to prefer a good (but not #1 rated) school that is attractive for other reasons (e.g. location).

    *I’m mostly talking about Chinese, Japanese, and Korean cultures with which I am semi-familiar. Other Asian cultures vary considerably.

    • educationrealist

      I agree with all that. But did I mention cheating? I might have mentioned it as an unattractive part of the pressure cooker but I searched the article and couldn’t find it.

      • Jay

        You used the word “fraud”. If you replaced “cheating” with “fraud” in my comment, it would retain the intended meaning.

      • educationrealist

        Well, I was talking about Asian culture back in those countries, and I think that’s a fair statement. But no worries. I wasn’t pushing back, just wondering what you were responding to.

  • Anonymous

    I enjoyed your return to posting, but your figures seem to be broken: the tags have no srcs.

  • Avocado

    The current SAT scoring system doesn’t distinguish between very smart and very-very smart kids because test takers at the 99.5 percentile and above (1-out-of-200 or better, +2.58SD) all get the same 1600 V+M score (2019 SAT mean and SD). There are no extra points available to separate 1/200 students from 1/500 or 1/1000 (+3SD). Indeed, it is sort of bogus even to try to choose between supposedly 1/100 (1550 V+M) and 1/200 (1600 V+M) scorers individually because of expected small (10-30 points) essentially-random variations in personal scores.

    This has an effect on perceptions of college “bias against asians” especially by middle-aged and older folks like appellate-court judges. People who took (or their acquaintances took) the SAT before 1995 and think a 1600 score means ~1/1000 achievement, not just ~1/200, are tempted to think that all those 1600-scoring asian applicants are super-geniuses.

  • Joel

    I’m getting here late but, like ER, I’m catching up. It may be that ER is the only person who will read this, but that’s reason enough to make the effort.

    First, thanks for this essay. No need for disclaimers about “disjointed” technique; the style is fine, and readers come here for the unique insights. For instance, explaining how the much-discussed high proportions of Asian students at the test-in public high schools are largely attributable to white students choosing alternative selective schools. For all the coverage elsewhere, only ER explains this. It’s a revelation for anyone who has been following this in mainstream reports.

    Gary Rubenstein, who has been a math teacher at Stuyvesant for 20 years, just wrote a series of 14 posts describing the issues in great detail. It’s well worth the time to read to get a deeper understanding, but his account lacks ER’s insight into White Flight. Like ER, GR has the first-hand knowledge of a practicing math teacher, but unlike ER, his identity is known so any public statements have the potential to put his career at risk. Such is the tragic plight of the public school teacher. ER’s pseudonym is an essential and necessary shield without which he’d lack the freedom to state inconvenient facts.


    In news reports about education, only “nonwhites” are victims of Disparate Impacts. This overlooks the disparate impacts on non-nonwhites, causing White Flight. All parents are going to do what they think is best for their own children. Clausewitz famously said “War is the continuation of politics by other means.” So is gaming the system.

    Competition for scarce spots in good schools has been based on merit; the new competition is based on gaming the system.

    Parents pay for test tutoring, teachers teach to the test, administrators exert pressure on teachers to dispense unearned passing grades, politicians boost graduation rates by downgrading standards. Boards of Education relax admission requirements for elite STEM high schools (Lowell in San Francisco, Jefferson in Virginia).

    High school Exit Exams are eliminated, community college remedial courses are eliminated, college admission SATs are eliminated.

    Underfinanced colleges admit unprepared students, propped up by nondischargeable student loans. Politicians get votes by promising to erase the debts. Like universities across America, the University of California depends on “out of state” tuition of students from China (26,912 in 2019).

    The problem with pretending is that it doesn’t change reality. Players lose credibility and diplomas lose value. Freddie deBoer, a self-described socialist, makes the point that “human beings are not equal in their abilities” in a recent post. He repeats the theme from last year’s book The Cult of Smart.



    ER readers already know most of this. This comment is partly to support ER’s post, and to make a few connections for the benefit of anyone belatedly getting to the last comment in a weeks-old essay. What’s happening in American public schools right now should be a concern to those of us who already benefited or who care about the next generations. From pre-K to PhD, things are rapidly deteriorating, and these institutions are the foundation of our country. We need good schools for everybody, and special schools for some, whether remedial or advanced.

  • Calvin Hobbes

    Something at least a bit related to the whites versus Asians situation in selective NYC schools is the situation with MathCounts. MathCounts is the premier math contest for middle schoolers. Many East Asians are fanatics about MathCounts, sort of the way some dot-Indians are fanatics about the spelling bee.

    Each of the 56 states sends a 4-person team to National MathCounts. Here are the 2021 teams:

    Click to access 2021%20National%20Competitors%20Final.pdf

    Here are the top 56 participants:

    Click to access 2021_national_results_for_we.pdf

    Based on the names of the top 56, it looks like at most 5 of them have both parents who were born in the US, probably more like 2 or 3. Some Indians. Lots of Chinese!

    I don’t know whether it would have been hard to cheat on the national test this time, since it was online. It would have been hard to cheat on the national test before it went online, but it would have been ridiculously easy to cheat on the state tests in most states. The same state test was used in all states, but the state competitions were on different weekends! The grading of state tests was done by parent volunteers, so Uncle Bob in one state could have passed the test on to nephew Jim in another state. I have not heard about MathCounts cheating scandals, but it would be surprising if there weren’t some cheating, given how easy it would be and how seriously some people take the contest (and certain aspects of certain cultures).

    One of the MathCounts feel-good stories is Melanie Wood:



    MH: When did you first realize that you had a special talent for mathematics?
    MW: When I was in seventh grade, a teacher invited me to be on the school’s MathCounts team, and I went to the competition without any preparation or any idea what it would be like. I was shocked to win first place in my city, then my state, and then 40th in the nation. Bob Fischer, the coach of the Indiana MathCounts team and a middle school teacher, sent me problems throughout the following school year to help me prepare for the next year’s competition—in which I placed 10th in the

    Melanie Wood had great success in MathCounts with little preparation at first, because she’s REALLY smart. Just being really smart is not going to cut it anymore, since the Asians especially train for it like crazy.

  • Murray/Sailer on Powerline Podcast | educationrealist

    […] more, long enough for the Wall Street Journal to notice it back in 2005. I wrote recently about the decline in white applications to the eight NYC specialized high schools. Whites and Asians are both about 15% of the NYC public […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: