Just to be clear, I have no doubt that there are lots of poor ethnic Asians with great SAT scores, at Berkeley and elsewhere. The hypothesis that SAT scores are generally correlated with income hardly precludes the possibility that there are many, many people with low income and high scores, and it would not at all be surprising to find that lots of them wind up at Berkeley or other UC campuses.
Everyone is right to criticize Guinier for simplistic application of the larger statistics to Harvard. Harvard (and its ilk) are full of people who are not particularly well captured by statistics. And I think that’s largely going to be true of Berkeley as well – you are still looking at a very thin layer of the very best students, and they’re not the result of some simple formula involving family income and reading hours (although both of those probably have an effect). Nonetheless, it remains true, one way or another, that Harvard’s student body is very disproportionately rich – and way beyond the top 10% of wealth and income.
(Everyone is also right to criticize Guinier for what is really implicit in the passage we have been addressing: the implied premise that Asian is just a subset of White.)
And? This is a bad thing? Maybe its only a problem for those who think no one should be allowed to be that way. That interpretation explains a lot, and its one I arrive at all the time these days.
Of course, to be fair, simplistic application of numbers or concepts is SOP in legal publishing. What passes for many peer reviewed law review articles is shocking, to say the least.
“Yes, because college adcoms can’t possibly figure out that the kid from New Trier has the means to afford the expensive tennis coach / lessons and the access to the science lab for an internship that the kid in the South Side of Chicago doesn’t have. Come on now.”
And you’re saying that college adcoms can’t possibly figure out that the kid from New Trier also has the means to afford the expensive private SAT tutoring that the kid in the South Side of Chicago can’t? Come on now.
Huh??? I’m saying exactly what you are. Of course college adcoms get that New Trier kid has more opportunities than South Side kid. I’m expressing incredulity at the assumption that they couldn’t figure it out.
I am rather late to the party, so apologies for taking this back to some of the earliest points that were being made. Some were pointing out various examples of Guinier’s poor choice of illustrations and data selection and saying that makes the entire article and her conclusions dubious. Others responded that even if they agreed that those were poor examples, a couple of weak arguments doesn’t mean the entire piece and its conclusions are wrong. To me it is a matter of degree, but certainly when an author makes certain fundamental errors in choice and logic in areas with which I am familiar, it causes me to be very cautious in accepting anything they else say that I am not as familiar with, and dissuades me from taking the time and effort to check their sources and overall reasoning path.
So one thing that really struck me as outrageously ridiculous and just so sophomorically wrong that it is hard to take anything else she says seriously is the following. I am a little surprised no one has mentioned it yet, although I can see why people might say it is not directly on point to her article. But again, if someone can have this fundamental a misunderstanding… Well, here it is.
What person, of any decent educational level, doesn’t understand the difference between a phenomenal memory and intelligence/reasoning skills? Did it seriously not occur to her that the LSAT mostly tests the ability to do logical inferences and not memorization? That Enrique might be of just average intelligence (or less) and that is the reason for his low scores as opposed to whatever verbal references he might have missed? In fact, with such a great memory I would think that it is highly unlikely to be due to the latter. But even putting that aside, who doesn’t know that a person with a photographic (or eidetic) memory can seem brilliant in everyday conversation, get great grades (especially if they major in something that rewards memorization) but not actually be that intelligent, if we define intelligence as the ability to synthesize existing information into new information and/or logical conclusions? Why would she think that knowing what is on page 384 would lead to a high LSAT score? The absurdity of this seemingly tangential paragraph in her piece is enough to make me doubt her intelligence and thus her whole piece. For others, while that alone might not be enough, when added to some of the other examples pointed out by xiggi and others it sure seems to demean her academic chops.
You might want to re-read the paragragh you extracted…
Test anxiety due to lack of confidence is a plausible cause of underperformance on standardized tests - the same phenomenon happens in highly competitive sports.
Guinier never mentioned anything about missing verbal references in this example…
Are you suggesting that whatever the LSAT measures is more useful to society than whatever winning a Rhodes Scholarship measures? If yes, then you should base your argument on that premise, not what is on page 384.
Based on this statement it appears that you are passing judgement on her reasoning skills without checking her overall reasoning path. Does that seem logical to you?
I am saying that when an article or paper is littered with enough examples of weak reasoning which I can spot without doing any further research, I feel quite free to dismiss the quality of the article in total and confident in that decision. You are free to not take that approach, but I would bet most people do exactly as I do. Intelligent people perform those kinds of judgement calls every day. It maximizes our most valuable resource, time.
I saw the reference to being too tense, but not only am I skeptical of that explanation, it has nothing to do with my point that linking prodigious memory with intelligence is a wildly false premise, LSAT example or not. I am simply pointing out another explanation that is at least equally valid, and I would argue the simpler and thus more likely explanation (Occam’s Razor). Again, just my opinion. It isn’t right or wrong, since we cannot experiment on this person, not do we have enough other evidence of his performance over the years to bring to bear on it. Except I would point out that if he performed well enough on various exams throughout the years to have the grades to win a Rhodes Scholarship, I think it brings a lot of doubt to the explanation of test anxiety. Besides, aren’t the GRE’s required when applying for a lot of those scholarships? If so, he must have done pretty well on that. In any case, I just think that a person that has made it all the way through undergraduate with that sterling of a record to be claiming (or having claimed for them) test stress is a pretty weak argument. Hence my skepticism.
I am making no judgement about what an LSAT measures vs. what winning a Rhodes Scholarship measures, the latter of which is nearly impossible to say anyway, or how that relates to society. You are arguing about things that have nothing to do with my point. Which is really quite simple. To me, anyone that writes an article that has as one of its premises that a person is a genius because they have an amazing memory is very suspect in their overall intelligence and reasoning skills.
Mastadon Although fallenchemist and I have our share of disagreements, I think your mistaken in this case.
First, “stereotype threat” is a theory that doesn’t seem to pass muster. In the example of Enrique, why if stereotype threat is at work, why was Enrique able to pass tests in his standard coursework with flying colors as indicated by his selection as Rhodes scholar but not do well on the LSAT for which he took the practice test over 30 times?
In order for “stereotype threat” theory to be legitimate, Enrique would have had to demonstrate that he had a command of logic and critical thinking that would mimic the type of abilities that are on the LSAT. We have no such data. The only evidence that Prof. Guinier provided was that Enrique had a fantastic memory and that he received a Rhodes scholarship. As stated by fallenchemist, having a fantastic memory is not the same as being able to take information and synthesize it.
Although winning a Rhodes scholarships is impressive, winning this award does not show that one has a grasp of logic and critical thinking that is required for high LSAT scores. It only indicates that Enrique had the qualification that the Rhodes scholarship committee was looking for their in their scholarship recipients. As an anecdotal note, I had a friend in college who won the Rhodes scholarship as an Economics major, I had to tutor him in Calculus on nearly a daily basis for him to do well to maintain his high GPA. He managed a B+ in Calculus, although he averaged an “A” in everything else.
Fallenchemist never stated that the LSAT measured something more “more useful to society than whatever winning a Rhodes Scholarship measures”, he only stated the obvious that Enrique did not possess the necessary cognitive ability to do well on the LSAT because he did not have the logic and analytic requirements that the LSAT required to do well on it.
I do agree with his views and assessments of Prof. Guinier.
Espenshade used SAT scores to study the advantages given to different groups by elite universities during the admission process. SAT scores are fairly stable over time ( the verbal score was recentered by 50 points in 1995). The number of applications sent out by top students has grown during the past 30 years so the number of applicants that an elite school receives has risen, the number of matriculants has stayed the same and the % of applicants admitted has dropped. The following data from Yale for matriculated students illustrates this:
year …percent admitted …50% verbal SAT score…50% Math SAT score
1986 …18.1% …670 …700
“The problem is that the data that is culled from very generic sources hardly fits the points she raises about … schools such as Harvard. She took the data and ran with it!”
Right. I think it’s possible that she’s right about much of this, but she isn’t providing much (any?) evidence that supports that conclusion. I also don’t appreciate the clickbait headline implying that Harvard and Yale rely more on the SAT than other schools do. I think the opposite is true – many public schools have bright-line in or out policies based on scores alone, while H & Y have the resources and staff to read the applications individually and consider the students’ circumstances. Besides, those bright-line schools often have an accessibility aspect to their missions that H & Y don’t share. They are explicitly dedicated to seeking outlier applicants.
Hanna I agree that H&Y looks at other factors in its admission decision probably more than most schools, but when the average profile of its freshman has a SAT composite of 2300/2400 it is hard to argue that H&Y doesn’t rely on the SAT any less than other schools.
Prof Guinier uses this observation to her advantage in making her unsubstantiated claims. She is making persuasive argument, not logical argument.
Few schools don’t rely heavily on the SAT so it seems like it would be hard to determine one school that relies more heavily on it than others.
Many schools offer free tuition, stipends, room and board just to kids who attend with very high exam scores. Harvard doesn’t do that for a kid with a 2400 so who’s more glamored (term from True Blood) by high test scores?
I really can’t tell. Would a state school reject these kids with 2400s because it’s focused on some other criteria? Most kids with high scores assume the state school will take them, the odds are assumed to be 100% because they scored so well whereas Harvard and Yale are still crapshoots. So who really is all about the score?
@xiggi wrote:
“The SAT is obviously taken in its appropriate context but so are the multiple elements of the applications.”
Espenshade uses the SAT because of all the variables that are available on a college application the SAT is the single best predictor of admission rates at elite colleges. Other variables provide weaker correlation for predicting admission.
“obfuscate the inherent lack of control…”
Espenshade obfuscated nothing. The descriptions of his methods are straightforward.
“have no other option to use blunt force…”
No. His statistical analysis consists of a series of logistic regression models
Many of the elite holistic colleges we frequently discuss on this forum follow a different pattern, particularly in more recent years than the Epenshade study. For example, the acceptance rate and SAT scores at Stanford over the past 10 years are below. The admit rate changed by approximately the same factor as Yale did in your example above, yet SAT scores had relatively little change.
year …percent admitted …25th/75th verbal SAT score…25th/75th Math SAT scoree
2004 CDS…13.0% …680/770 …690/780
2014 CDS…5.1%…680/780…700/790
I’d expect the applicant pool is as strong or stronger today than it was a decade ago, and Stanford wins more cross admits among the most sought out applicants than they did a decade ago, yet test scores had little change. This suggests that Stanford is not making decisions by looking for the highest scores among the applicant pool and instead focusing on other criteria. If you look at actual admissions decisions, they show a similar pattern and cannot be predicted well by test scores at Stanford and at most other highly selective holistic colleges, even though they can be predicted fairly well by stats at most colleges with similar admit rates to the Epenshade study (averaged ~25% admit rate in study), which tend to be less holistic than HYPSM… However, there are other selective colleges with similar admit rate to ivies that do appear to focus more on scores and admit the overwhelming majority of applicants with top scores. Vanderbilt is a good example, and their CDS shows a very different pattern from Stanford’s as quoted below. Note that Vanderbilt’s scores have overtaken Stanford’s in spite of the far higher admit rate among what is likely a lower average score applicant pool and are on pace to soon overtake all the ivies as well.
year …percent admitted …25th/75th verbal SAT score…25th/75th Math SAT score
2004 CDS…38.3% …620/710 …650/730
2014 CDS…13.1%…710/780 …720/800
Data10 You might want to compare apples to apples vs apples to oranges. The Yale data compare 1986 to 2012. Your compares 2004 Stanford to 2014 Stanford. If you go back and compare the same periods I’m sure you will see a similar pattern for Stanford and Vanderbilt as to Yale. 1986 did not have superscoring as they do now.
As I mentioned in prior posts, in 1986, you would be hard pressed to find anyone who took the SAT more than twice because the super majority took the SAT just once.
It sounds like you missed the point of my post. Changes in superscoring and number who take the test multiple times is all the more reason the older Epenshade data is less relevant today. That said, I’m not familiar with admission policies of the listed colleges in the 80s, but if was anything similar to current policies, then you will see different patterns for different colleges, much like the Stanford vs Vanderbilt example.
Data10 Not sure how you come to the conclusion that the Espenshade data is less relevant today. Please explain. It seems to me the opposite, that the data is just as relevant today.