"Race" in College Applications FAQ & Discussion 12

Applying to College
1541

While Berkeley is one of the nation’s finest universities, I explicitly limited my comparison to colleges that have ample endowments and that meet full need, so that family finances were less of a determining factor. This approach might be imperfect, but I did give it a bit of thought. For example, I was wondering whether or not to include Johns Hopkins since its endowment was considerably smaller than others, and therefore was less likely to be as generous with financial aid.

Could you point me to your source for CalTech’s graduation rate?

1542

@hebegebe - a few years old but https://iro.caltech.edu/gradrates

1543

[quote] @SatchelSF It's well known in the literature that if you match URM, whites and (Northeast) Asians for intelligence, URM will subsequently underperform and Asians will outperform the baseline white performance on subsequent tasks.

One might also consider - egads! - that the institutions in their style of teaching and performance evaluation might just be biased against males, and that the differences in grades and certain other academic metrics observed throughout life confirm that the whole society is hopelessly biased.

[/quote]

So if URMs have lower grad rates it’s because they’re less intelligent (or rather, members of less intelligent groups) but if men have lower grad rates, it’s because society just isn’t fair to them.

But society is perfectly fair to URM groups so we don’t need to consider that because it’s all genetic.

Laughable but predictable.

1544

@apprenticeprof - “If people of African descent were biologically inferior in intellectual capacities…”

All I said was “different.” I’m not always successful, but I do try to be careful not to use normative language, preferring to note that we are talking about differences, often involving both relative strengths as well as weaknesses, between groups; many (including myself) believe these differences reflect at least 100,000 years of largely separate evolutionary tracks under very different environmental conditions. Obviously, it is hard to avoid using terms like “lower” or “higher” when comparing numerical quantities like test scores or GPA, for instances, and these descriptive terms sometimes carry unintended, normative connotations.

“it would stand to reason that African-Americans – who, on average, are genetically about 25 % Caucasian – would outperform more recent arrivals from Africa.”

Depends very much on the character of the arrivals of course. If there is positive selection - if only the most capable from a particular area emigrate - that the new arrivals outperform certain native subgroups really says very little. It might be interesting, for instance, to test a seemingly negatively selected group - say refugee and economic migrants from Africa making their way into Europe recently - and compare their ability with the legal African immigrant cohorts in both Europe and the US. In any event, any strengths of African immigrants and their progeny are already baked into all the data we have, including scoring data, just as in the case of Asian immigrants and their progeny. It should be noted that there has not been any appreciable narrowing of any black-white scoring and achievement gaps for decades, although admittedly the numbers of African immigrants are relatively small I’d think. (There has been some apparent narrowing of school achievement test scores, but this has largely been a mathematical artifact from easing of standards generally - for an interesting introduction that is nevertheless accessible to the non-mathematician, see http://www.lagriffedulion.f2s.com/gap.htm.)

“Instead, discussions of Affirmative Action often note that a disproportionate number of its beneficiaries are immigrants or the children of immigrants, who have far less, if any, Caucasian ancestry.”

Obviously true, and calls into question what affirmative action as practiced is really all about. Steele, as noted above, makes a poignant argument that it has much more to do with affluent whites feeling good about themselves - and “virtue signalling” in modern parlance - then about addressing the actual conditions faced by African Americans historically (see http://www.jstor.org/stable/41211829).

1545

A little OT, but we’ve kicked around the idea that females are better students than males extensively in this thread - better GPA, higher graduation rates, etc. MIT has been mentioned particularly.

Well, despite admitting male students who are obviously less qualified than female students, MIT managed somehow to pull off something truly remarkable in the most recent Putnam competition, which of course it is now prominently touting in the circles that follow these sorts of things:

http://news.mit.edu/2018/mit-students-take-back-putnam-top-honors-0222

For people who are not math geeks, it’s hard to convey just what the Putnam means in the college math world - think the Super Bowl and World Series all wrapped up in one. The competition attracts just about every truly talented undergraduate math student in the United States and Canada.

MIT took 1st in the 2017 team competition. And in the 2017 individual competition, MIT took 4 out of the top 5, 10 out of the top 15, and 17 out of the top 25.

https://www.maa.org/sites/default/files/pdf/Putnam/AnnouncementOfWinners2017.pdf

Oh, not a single female in that top 25, and once again MIT when faced with the choice of picking its team picked 3 males, now for at least the 16th or 17th year in a row.

Just think what MIT could have accomplished had it tightened up male admissions requirements and admitted more females!!

Seriously, congrats to anyone from MIT reading this.

1546

I previously listed the most recent grad rate reported to the government, but considering Caltech’s small enrollment and extremely small percentage of black students, it’s more appropriate to instead average over a large number of years. The average graduate rate over the past 10 years as listed in IPEDS (https://nces.ed.gov/ipeds/datacenter/Data.aspx ) is below.

Caltech – 2.25 (27% black, 12% white)

Note that there is still a large gap in graduation rate, even though Caltech has a reputation for holding URMs to the same admission standards as other races. Part of this relates the bulk of variation in graduation rate not being explained by admission stats, particularly among black male students. For example, the study at https://digitalcommons.ilr.cornell.edu/cgi/viewcontent.cgi?article=1014&context=student looked at why graduation rate differed by race at an Ivy League school (probably Cornell). The author was able to explain the vast majority of the difference in grads rates among hispanic students by controlling for a variety of performance measures, poverty rate, and various other things; but could not explain a huge portion of the disparity in graduation rate among black students. The author writes,

Ignoring studies, I’'d expect it to be challenging for black students at Caltech for various non-academic and non-financial reasons. It is common for fewer than 2% of the student body to be black at Caltech. You’d usually be the only black student in your classes, including TAs and professors. You’d also hardly ever see black students outside of classes, in dorms and such. Many at the school, would likely have preconceived notions about you because of race. I’d expect many black students would feel like they didn’t fit in and belong there. I’ve worked at engineering companies in which female engineers are almost as rare as black students at Caltech. The few female engineers generally do last long in this type of environment and soon choose to work elsewhere, regardless of how skilled they are.

1547

That’s not how small data makes statistics erratic. Small data makes the calculated drop out rates unreliable. No one has any idea what constitutes a drop out and the impact of socio-economics on those rates. The likely driver of the higher drop out rates are the disadvantaged students across races. If the disadvantaged students are a larger % of the AA student population (as they are in the general population) you’d expect higher drop out rates.

There are a multitude of disruptive factors (like homicide and incarceration) that disproportionately affect the AA community. Unless you truly believe that universities are capable of accurately tracking all the details and that they’re careful enough to make sure its all recorded correctly, you have to bring a certain amount of skepticism.

Besides, what makes you think these universities have the same priorities as the ones you’re focusing on. Think of it like this. They want some # of athletes, art history majors, anthropologists, and of course STEM. They also don’t want a homogeneous student population. There are regional weights, rural/suburban/urban, under represented, disadvantaged. And clearly it’s not strict, because Asians and Jews are both heavily over represented. But both get limited by their geographic and suburban concentrations. STEM has clearly become a bottleneck.

1548

I took a little flak for posting yesterday (post #1542) that the academic literature shows that testing generally tends to be biased in favor of URM, in that scores tend to overpredict subsequent performance.

It is perhaps ironic that the very next academic study cited by another poster in this very thread confirms it (see study cited in post #1549):

“[Bowen and Bok] also note that SAT scores tend to overpredict academic success for blacks. In other words, black students actually perform worse academically than their SAT scores would predict. In addition, blacks on average have lower GPAs than whites [footnote citations omitted].”

By the way, the author of that article is black, and was a Cornell senior at the time he wrote it. As I said, this is very well-trodden ground and is not a controversial statement even for undergraduates who have done a bit of reading in the field.

The Cornell study is interesting, largely for what it leaves out. After demonstrating that blacks at Cornell have the lowest average GPAs (you can see this by the regression coefficient estimates in the simplest Model 1 in Table 4 on p.10), the author does not explore whether and to what extent those black students (largely male) who fail to graduate found themselves at the very bottom of the class. Professor Sander, in the context of law school dropouts, did explore this avenue:

“Nearly 90% of black students in the LSAC-BPS data who…failed to graduate…placed in the bottom 10% of their classes. The median class rank of black students leaving law school between the first and third year was between the second and third percentile.” (p. 440 in the Sander article)

It is difficult to be in an academic environment in which one is at the bottom of the class, and suspect that fellow students are aware of that. (It’s not something that can be hidden, honestly.) Again, affluent, white supporters of affirmative action generally have no concerns about this real phenomenon, which could easily be addressed by reducing race preferences and allowing students to thrive in environments to which they are better “matched.”

1549

Clearly the answer is that there is a problem with the culture of the Putnam, right? It must be “failing our women” in some way. Also, nice cherry-pick. The vast majority of students at the most selective colleges never enter the Putnam, or even think about doing so.

1550

@OHMomof2 - Yeah, the Putnam might be a bit of a cherry pick; it’s just that we’ve been talking about MIT and female students so much, and MIT just announced the latest results which are truly impressive…

My point has always been that especially with the most selective STEM schools, institutions appear willing to overlook the greater variance in male behavior, as well as their apparent lower conscientiousness, in order to increase the odds of truly extraordinary achievement at the extreme right tail of the intelligence distribution, and, honestly, notable achievement at all points greater than about +2 standard deviations of the mean. That is because while males are only a little better on average than females in mathematics (and that’s innate, not cultural), the distribution of intelligence of males exhibits higher variance, making both the left (unfortunately) as well as the right tails “fatter.” While there are literally only a few hundred undergrads at any given time who have legitimately even a chance of scoring a single point or two in the competition (and it ranges up to 120 available points), +2 standard deviations is squarely in the middle at a place like MIT; in fact, it is going to be on the low side for the non-preference admits.

LaGriffe is always a fun read on these subjects. S/he (true identity was never revealed) made some predictions about female performance in the Putnam over ten years ago, as well as some predictions about the possibility of a female winning the Fields medal (basically, the Nobel Prize of the math world), based entirely off the only comprehensive IQ testing of kids that we have: the 1960 Project Talent database. So far, they are holding true. I’ve posted it before, and well worth reading: http://www.lagriffedulion.f2s.com/math.htm.

1551

How do you know it’s “innate and not cultural”? I’m impressed that you alone have resolved this question of nature vs nurture when the rest of academia is still arguing about the degrees to which one or the other is true.

1552

About variance in male intelligence being greater than female, it’s not a controversial position. Anyone can find reams of data confirming it.

Interestingly, it shows up all over the place, even in tests that people scream up and down have nothing to do with intelligence. For instance, the latest 2017 SAT data, which show the familiar variance ratios (and slightly higher male mean on SATM) just about identical to what was discovered in the 1960 IQ test database of kids: https://collegereadiness.collegeboard.org/pdf/sat-percentile-ranks-gender-race-ethnicity.pdf.

(Unfortunately, the SAT data are not particularly fine-grained. The aggregate female-male variance ratios incorporate all ethnic groups, including blacks, non-white Hispanics and Asians, which all tend to have higher kurtosis/less variance in their distributions, but if you squint a little you’ll see that the 1960 data - when there were many fewer Asians in the data, and fewer Hispanics as well - and the 2017 data basically agree.)

1553

That doesn’t explain how you know it’s “innate and not cultural”., @SatchelSF .

1554

@OHMomof2 - There is a concept well known to people who look at data all day, known as “consilience.” Sometimes, even when exact causal mechanisms cannot be pinpointed, there are nevertheless so many strands of confirming or at least consistent data points and evidence that the conclusion can be inferred. In that sense, the data are consilient. In fact, that is the way most inferential science is conducted :slight_smile:

As evidence comes in, we should always be willing to change our opinions, and I like to think I do from time to time.

In examining evidence, we should also be careful not to commit what Hume termed the “moralistic fallacy,” the belief that because we wish something to be so that it must be so. Obviously, many of the discussions on this thread touch on this concern.

About sex differences in mathematics, La Griffe gave a nice, nonscientific encapsulation of perhaps the probable origins:

"In early hominid societies women gathered and men hunted. The division of labor exerted distinctly different adaptive pressures on men and women, leaving them with different cognitive strengths. The legacy of the hunter is visuospatial proficiency including the ability to mentally rotate and invert objects in 3-space…also valuable in mathematics. Women’s better ability to identify objects and recall their location is the legacy of the forager.

Related cognitive sex differences may be found in other species, often with similar origins."

1555

I don’t think anyone is disputing that there are observed differences in IQ distribution. The question is to what extent they are the innate products of evolutionary development, and to what extent they are related to current and more recent historical circumstances.

To (stereotypically, I fear) bring it down to an even less scientific level, we’re dealing with something of a chicken and egg problem. Do certain groups gravitate to certain fields because they are measurably better in them, or are they measurably better in them because they have been disproportionately encouraged to pursue them?

I have no experience with math competitions, so I’ll defer to others on that, but I do have some experience with competitive trivia, which is heavily male-dominated in pretty much any setting you can imagine, from high school and college quiz bowl to pub trivia.

There are two obvious biological conclusions that could be drawn from this:

  • Men are naturally better at retaining large numbers of facts on a variety of topics than women, even if women might be better in certain tasks that require synthesis of facts.

-Because of testosterone, men are more aggressive. This correlates to speed (which is a large factor in many trivia contests), willingness to venture a guess when one isn’t certain, and perhaps even the desire to engage in competitive trivia in the first place.

I’m not going to condemn people from considering those possibilities (I’m also not someone who thinks Larry Summers should have been fired). However, there’s also non-biological explanations for the same results:

  • Men have traditionally been socialized to be more aggressive than women, which correlates with speed, willingness to guess, and desire to engage in competitive trivia. This creates a field likely to skew male.
  • The effects of the above are then compounded because questions likely to have been disproportionately written by men might disproportionately favor questions that lean toward stereotypically male than stereotypically female interests (i.e, more questions on science and sports than literature and art --this of course leads to the question of why some interests are coded to one gender or another, but we'll leave that aside for now). Even a writer who worked hard to avoid such disparities in terms of number of questions might find it difficult to write equivalent kinds of questions for a field that he knew a lot about and a field that he didn't.
  • Women tend to select against or drop out of trivia as an activity because it is predominantly male.

Given that, it strikes me as arrogant to assert that the science on such disparities is settled. And just intuitively, there’s a lot more basis for thinking there might be inherent differences in intelligence between men and women (who do display significant biological and hormonal differences that could, in theory, correlate to other kinds of difference) than that there might be inherent differences between races.

In either case, after centuries of disparate treatment and socialization, good luck on separating out the biological from the social and cultural.

1556

I couldn’t agree more.

@SatchelSF your responses haven’t answered my question - how you know the differences you describe are “innate and not cultural” when this issue is studied and debated and argued about worldwide, all the time, and is emphatically NOT settled.

It seems argumentative to keep insisting you justify an assertion that is really not justifiable, so I’ll let this go.

1557

The other thing is that if intellectual achievement is purely innate/genetic and not influenced by cultural/environmental factors, then the value of attending the “best” high school, college, etc. should be minimal.

So if there are any parents who believe that, they could save a lot of money by living where housing is cheap (because the K-12 schools there are seen as “bad”) knowing that their kid’s innate/genetic intelligence will show no matter what, and then send the kid to a lower cost in-state public or automatic-full-ride merit scholarship college, knowing that their kid’s innate/genetic intelligence will show no matter what.

1558

This seems to be evidence that standardized testing, including IQ, are far from the end all in predicting graduation and other measures of academic success. Instead the study found that test scores were not what the author called “significant predictors” of graduation rate among black students (particularly black males). Other studies have found similar results – that the bulk of the variance in graduation rate depends on other factors. I’d expect it’s more that these important “other factors” are correlated in race and were not included in the model. When you include other key factors for estimating graduation rate instead of just test scores, then the model works better and pverpredicts less. In theory, with a detailed enough model that looks beyond just test scores, there would be no overprediction of graduation rate.

The study suggests the difference in graduation rate is especially notable among black males, not as much black females. Consistent with this, if you look at colleges that have a high female enrollment, then the graduation rate looks a lot more similar. Some examples are below, looking at the average 6-year graduation rate across the past 10 years.

Wellesley College – 90.6% White, 90.1% Black
Vassar College – 92% White, 90% Black
Emory (60% female, no football) – 90% White, 88% Black

While the selective colleges with the largest gaps, don’t show as obvious a pattern – maybe a correlation with having a lot of engineering students and/or having admissions that focuses more on stats and less on other holistic criteria, although this doesn’t explain Carleton and Haverford.

University of California-Berkeley – 90% White, 74% Black
Caltech – 89% White, 73% Black
Harvey Mudd – 89% White, 74% Black
Carleton College – 94% White, 83% Black
Haverford College – 94% White, 83% Black
MIT – 94% White, 83% Black

1559

@apprenticeprof - You seem very interested in all these ideas, and are obviously willing to read. We have learned much in recent decades, especially since the full sequencing of the human genome, and we have come a very long way since the Harvard school of Gould, Kamin and others held sway. About nature versus nurture, a general book that I enjoyed very much (and I’ve recommended it before, so apologies for the repetition) is Cochran and Harpending’s The 10,000 Year Explosion: How Civilization Accelerated Human Evolution:

“[There is] the notion that modern humans everywhere have essentially the same abilities… If humans have not undergone a significant amount of biological change since the expansion out of Africa, then people everywhere would have essentially the same potentials…. But as we never tire of pointing out, there has been significant biological change during that period – tremendous amounts of change, particularly in those populations that have practiced agriculture for a long time. Therefore, the biological equality of human races and ethnic groups is not inevitable: In fact it’s about as likely as a fistful of silver dollars all landing on edge when dropped.”

I find that reasoning compelling, and I believe the burden has really shifted onto those who would posit that all races and ethnicities (and both sexes, by extension) must be the same, despite the massive differences we observe at all times, all places, and through all recorded history.

As I wrote above, I am always willing to listen to evidence that in fact we are all exactly the same in intellectual character, capacity and inclinations, despite the obvious biological differences, which include things like brain structure and even neurobiological pathways. I don’t think that the people who believe in the equality of everyone understand the societal implications that necessarily follow: namely, that if it is one’s position that we would all be the same “if only,” then for every disparate outcome we must engage in a seemingly never-ending cycle of blame and recrimination on the basis that one group (or sex) is always and everywhere “keeping down” the other. My modest proposal, as always, is to treat people as individuals, not as members of victim or oppressor groups.

As for the suggestion that perhaps male inclination towards violence might also be just cultural, I assume that was made in jest! There is too much evidence in the mammalian world (including every species of primate of which I am aware) of differential aggression by sex to really entertain the question seriously. It’s nature, not nurture.

1560

The problem with this logic is US college admissions isn’t an intellectual meritocracy. Other than Caltech, Harvey Mudd, and to a lesser extent MIT, elite US colleges don’t admit just smart kids until they run out of slots.