The confounding effects aren’t handled in the simple analysis that @coase described. That’s why coase referenced more sophisticated “fixed effects” models.
Back in the old days, one professor used to assign this exact problem in one of his advanced courses as the class project. This was nice because we were able to get some high quality work done for free. Stripping out the fancy math, the basic idea behind a fixed effects model is very simple. For example, you look for classes that an electrical engineering major and an english major take in common and see how they both do in order to get a calibration. Ideally, you want the overlapping classes to be on “neutral territory” - i.e, a class that doesn’t favor the aptitudes of the electrical engineering major over the english major, or vice-versa. A more sophisticated model infers clusters of aptitudes and handles this. Anyway, you figure out which of the overlapping classes favor which majors and you correct for this bias too. Once you have a calibration of the students’ respective abilities, you can figure out which majors tend to grade harder or easier.
The data was all anonymized, but it wasn’t that hard for the faculty to figure out what majors the codes referred to. And the results showed exactly what you might suspect -
The majors at this college that were stereotyped as easy majors were, in fact, easy.
The majors that were stereotyped as hard majors were, in fact, hard.
To me, the lesson was this - the students may not have all been geniuses, but they were all geniuses when it came to knowing what the hard and easy classes were. By hanging out at a fraternity for a couple hours, even the weakest students could tell you the results of a sophisticated, multi-variate fixed effects model with multiple aptitude clusters estimated on almost 200,000 data points
Dr. Isler attended a tiny Christian high school. She is very smart and hard-working, because anyone who gets a physics degree from an Ivy has to have loads of raw, seething brainpower What “preparation” do you think she needed that she wouldn’t have gotten at Yale, if she’d been there as an undergraduate? Doesn’t Yale have calculus classes, intro physics classes and intro chemistry classes?
I am sure that if you asked Yale’s admissions department, they would tell you that they want to admit very smart, very hard-working students from tiny Christian high schools and other schools that don’t have good science and math classes. I am certain they do not say, “That kid is incredibly smart and hard-working, but we don’t take students like that if they went to high schools that didn’t have good science classes.”
I think she in particular would have been fine at a top undergraduate school.
Re: switching patterns in majors. I do think we tend to want to lean on info that appears absolute. We like to think the valedictorian is truly superior to the kid who placed fifth. Or that the engineering student dealing with the most complex concepts is somehow smarter than the gender studies kid. I’m not sure if it’s some human need or some flaw in the scope of the analytical thinking. In my training, no study was any better than the basic assumptions allowed it to be- or constrained it to be. How the initial framework is considered is crucial.
I guess an issue I have with the arguments about holistic is that one side tries to find absolutes. You want to think the val has superior chances (and the point examined is usually, again, hierarchical- eg, who gets a better college gpa. So you end up defending hierarchical thinking with hierarchical proofs.) Or that the kid chosen despite slightly lower stats can somehow be labelled inferior.
Different thinking, mindsets, skills and sorts of sweat go into different fields. You can be brilliant in the social sciences and never calculate the thrust needed to lift an airplane. You may have thought you were fascinated by engineering when in hs, despite never experiencing it, then found your own better path, once exposed to the “more” at college.
More than that, post-college success, as individuals and as contributing members of society, who go on to influence others in positive ways, is not as simple as numbers, whether or not you stuck to your 17 year old choice of major, how a class is graded, or your income.
If you can’t see that, I guess you can’t understand what “more” holistic is looking for.
Note that switching patterns between majors may be affected by whether a major has long prerequisite sequences. Switching into such a major may require graduating later than switching to a major with shorter prerequisite sequences, so that may deter switches to long sequence majors.
Also, some majors at some schools may not have spare capacity to take in many or any students switching in.
According to Yale’s Common Data Sets, Yale’s entering class in 2007 was 17.7% URM (Hispanic or black non-Hispanic). For that class, Yale had a freshman retention rate of 99% and a 6-year graduation rate of 98%. This suggests that very few enrolled students, URM or otherwise, are “floundering” at Yale, further suggesting that there’s little, if any, mismatch.
Hunt’s right, the most selective schools accept only students they expect to succeed. They may dip slightly lower into the SAT pool to get the URMs and first-gens and geographic diversity they want, but not so far as to shoot themselves in the foot by admitting large numbers of students who are likely to fail. They don’t need to, and their admissions staff are just not that incompetent.
The problem is that she probably didn’t receive the preparation at her high school that would have allowed her to succeed at Yale. I suspect that she got that preparation at Norfolk State, though. She may have been a huge outlier at Norfolk State, but at that school today only 1% of freshmen scored above 600 on the math or reading portion of the SAT, and none scored above 700 on either portion. That suggests a level of math preparation that, in my opinion, would make it pretty hard for a student to succeed as an astrophysics major at Yale. Yale and Norfolk State have different missions. It appears, from her example, that Norfolk State can work with a student who has a lot of brainpower but weak preparation, and help the student succeed. That isn’t really what Yale does, because it doesn’t really accept students like that.
UT has a lot lower graduation rate than an Ivy, but is there any reason to believe they are taking students with a lot of Bs and some Cs in the holistic admissions as affirmative action admits? Reading the case materials, I’m getting the strong notion that the 7% rule picks up the top students from weak schools-- students that at least know how to play the school game by getting As, even if their courses are easy-- and the holistic admissions pick up students from the second decile in good schools-- so they have good preparation, even if they didn’t do the admissions game as fervently as the top students.
I’d love to see the regression coefficients for the little app that tells who graduates.
And if I’m computing graduation rates of URMs, I’d like to see the rate omitting recruited athletes. We don’t want to lump URMs brought in for diversity reasons with illiterate athletes brought in to win football games.
To echo what Hunt said, if the author felt she was unprepared going into Norfolk State University, she clearly would have been way under-prepared going into Yale, where the student next to her could very well have been an IMO gold medalist, or an Intel Science fair winner. The author could be equally bright as those students, but her peers have refined their math and science skills to military grade levels.
This is not a reflection on her, but to the differing environments that they come from. For example, in my upper middle class town, there are lots of kids that enter the Johns Hopkins CTY program, and as a result many take the SAT before age 13. I know several kids that got above a 750+ on the math section by age 11 or 12. Fast forward a couple of years, and some of these same kids have won the state MathCounts and are competing in national MathCounts, or the USA JMO, or science fairs. These kids may or may not eventually get into Yale or {insert selected prestigious college here}, but they have developed serious skills and a work ethic that will serve them wherever they go.
Unless the author is Terrence Tao brilliant, it would simply be unfair to expect her to be able to compete at Yale as a freshman. Fortunately for her, she found a very good match college, that pushed her to develop her skills so that she could enter and complete the Yale PhD program.
I am a strong proponent of tilting the scales based on SES, because the lack of preparation and opportunity could happen just as easily to the poor white or Cambodian student as with the Black student. But the tilt level has to be realistic enough so that the students can actually succeed.
Good point. University of Texas black football players reportedly have a graduation rate of 44%, well below the university’s overall 6-year graduation rate of 80%.
So you are saying that Yale shouldn’t admit the best students from rural schools and tiny Christian schools because they can’t possibly succeed? If that were true, it would be a damning indictment of Yale. But it’s certainly not what Yale says about itself.
ETA: I know that there are other Ivy graduates in this thread. I ask you, is what hebegebe says what you want of your alma mater-- that it only admit students from top programs, because other bright, talented students went to worse schools? That is certainly not what I want of Brown. I’d be embarrassed if Brown announced that it wasn’t taking students from schools with weaker programs, no matter how bright, talented and promising the students, because the students didn’t have the same resources that students from private schools and top suburban public schools have.
I want Brown to admit that bright girl who wants to be an astrophysicist and who worked hard in her weak school. When she steps into her intro physics class, I want the prof and the TAs to tell her, “We know you can succeed, and we’re going to help you do it.”
The student from the inadequate high school is I think, is the kind of student for whom SATs are most important, imperfect as they are. Of course Yale and Brown shouldn’t make a policy of barring students from weak high schools from attendance. In those cases, however, standardized test scores become a reasonable way of assessing whether or not the student is likely to succeed. If the kid from the weak high school has around a 650 on each section, that’s not so low that she can’t make up for deficiencies through hard work and some coaching. If she’s got a 500 on each section, that would suggest to me someone so woefully unprepared that it would be a disservice to admit her, even if she’s the valedictorian of her school and has overcome significant hardships.
The more interesting question is whether or not even the girl with the 650 might at times be better served by a less competitive school than Yale, especially if she wants to go into STEM. My gut instinct is that the answer is usually no, but can be “yes” in certain circumstances.
It is also important to remember that UT is not Yale, so the student getting a boost from AA there might actually be in a worse position for success. As Hunt said above, at Yale, even the student admitted with somewhat weaker academic qualifications is likely to be a pretty accomplished student, or at least strong enough that he or she isn’t likely to wind up in danger of flunking. At UT, that isn’t necessarily the case – although, as CF said several posts ago, my guess is that’s more of a problem among admits let in under the top 7 % plan than the holistic plan.
(I’d be “shocked, shocked” if anyone on cc thought there is a difference between top 7% and top 8%,but obviously the Texas State Legislators/UT Regents do.)
It would be interesting to see a breakdown of SAT scores for 7% rule admits v. holistic admits. My guess would be that most of the low SAT scores are among the 7% admits from poor schools who get in purely on class rank in an under-resourced and non-competitive academic environment, If that’s the case, it would support Cardinal Fang’s suspicion that there’s likely to be a greater mismatch under the 7% rule than under holistic admissions.
There’s no “suspicion” needed. UT has had to implement extensive coaching/remedial programs to address the mismatch of the under-prepared students admitted via the percentage plan. It’s all reported in that NYT article “Who gets to graduate” that I’ve linked several times.
Perhaps true, but that masks the need for prep. If someone dreams of UT and is comfortably in the top 5% of his/her HS, there is only the need to show up for the SAT/ACT. Filling in random bubbles is enough for admission. Such kids may have the ability to score a 650 Math, but if that takes practice, a 500 will do ‘jus’ fine, thank you.’
In contrast, those 8% or worse, hoping to get in during holistics, will bust their butts trying to improve their apps, including test prep.
