Clearly fit seems to matter to posters. Otherwise why tour a place or talk to current students? So I think students do some self-selection because they think a particular vibe fits them, but there is also acculturation that occurs once a student arrives on campus. Students can do a lot of growing and changing between 18 and 22. And their peers influence them (as well as faculty, coursework, geographic location). For some, their college campus may be the most diverse place that they have ever been. They might be exposed to ideas that they have never considered. It is not just that the students arrive already fitting = the culture. Nor is it that the particular culture adapts to its current study body. Both the students and the institution change and adapt to each other. I consider that a good thing. Otherwise stagnation.
The vast majority of the analysis controlled for tests scores. They found among applicants with the same scores, top 1% kids were notably more likely to be admitted than high income kids. One key contributing factor to why top 1% kids were more likely to be admitted than high income kids with the same scores, was non-academic factors, such as non-academic ratings, legacy, and athlete. In contrast, they found that among applicants with the same test scores, there was less variation in academic rating by income. Academic rating appeared to be more correlated with test scores than non-academic rating, which does not seem surprising to me. However, this does not mean than non-academic factors were more influential in concentration of wealthy than academic factors, such as test scores.
The study found the following degree of raw overrepresentation and underrepresentation at Ivy+ colleges, without controlling for test scores. By overrepresentation/underrepresentation, I mean relative to the distribution of the US population. For example, if 20% of Ivy+ students are top 1% income, that is a 20x overrepresentation. The higher you go in income, the more overerpresented at Ivy+ colleges, up to 29x overrepresented for top 0.1% income.
Degree of Overrpresentation by Parents Income (no control for test scores)
0-20th Percentile Income – 0.16x
20-40th Percentile Income – 0.24x
40-60th Percentile Income – 0.4x
60-80th Percentile Income – 0.7x
80-90th Percentile Income – 1.5x
90-95th Percentile Income – 3x
95th Percentile Income – 4x
96th Percentile Income – 5x
97th Percentile Income – 7x
98th Percentile Income – 9x
99th Percentile Income – 16x
99.9th Percentile Income – 29x
In raw numbers, top 1% was 16x overrepresented, and top 0.1x was 29x overrepresented. When controlling for scores, the distribution changes to
Degree of Overrpresentation by Parents Income (with control for test scores)
0-20th Percentile Income – 1x
20-40th Percentile Income – 0.9x
40-60th Percentile Income – 0.8x
60-80th Percentile Income – 0.7x
80-90th Percentile Income – 0.7x
90-95th Percentile Income – 0.8x
95th Percentile Income – 0.9x
96th Percentile Income – 1.0x
97th Percentile Income – 1.2x
98th Percentile Income – 1.3x
99th Percentile Income – 1.8x
99.9th Percentile Income – 2.6x
The 16x overrperesntation for top 1% dropped to 1.8x after controlling for scores. The 29x overrpresentation dropped to 2.6x after controlling for scores. This suggests the vast majority of the wealthy overrpresentation is correlated with wealthy kids averaging higher average test scores (includes aspects of academics that are well correlated with test scores), rather than differences in non-academic factors.
Consistent with this, a comparison of the portion of students with high (1400+) test scores and portion with high non-academic rations, by income is summarized below. Kids from top 0.1% income families and to a lesser extent top 1% had a statistically significant difference in portion with higher non-academic ratings, but the difference in non-academic rating distribution is far smaller than the large test score difference, with top 1% kids being 130x more likely to score 1400+ than bottom quintile income kids, or 180x more likely for top 0.1% income kids.
Distribution of High Non-Academic Rating by Parents Income
0-20th Percentile Income – 15% high rating
20-40th Percentile Income – 14% high rating
40-60th Percentile Income – 14% high rating
60-80th Percentile Income – 16% high rating
80-90th Percentile Income – 15% high rating
90-95th Percentile Income – 15% high rating
95th Percentile Income – 16% high rating
96th Percentile Income – 16% high rating
97th Percentile Income – 17% high rating
98th Percentile Income – 17% high rating
99th Percentile Income – 18% high rating
99.9th Percentile Income – 21% high rating
Distribution of 1400+ Scores by Parents Income
0-20th Percentile Income – 0.1% of kids get 1400+ score
20-40th Percentile Income – 0.3% of kids get 1400+ score
40-60th Percentile Income – 0.5% of kids get 1400+ score
60-80th Percentile Income – 1.3% of kids get 1400+ score
80-90th Percentile Income – 3.0% of kids get 1400+ score
90-95th Percentile Income – 5.5% of kids get 1400+ score
95th Percentile Income – 7.5% of kids get 1400+ score
96th Percentile Income – 8.4% of kids get 1400+ score
97th Percentile Income – 9.8% of kids get 1400+ score
98th Percentile Income – 11.2% of kids get 1400+ score
99th Percentile Income – 13.5% of kids get 1400+ score
99.9th Percentile Income – 18.4% of kids get 1400+ score
Living up to your username here . . . excellent analysis that I wish could be sent to every pundit who has confidently proclaimed that test scores are the least-unfair element for poor kids.
My two cents is it is extremely unlikely we are coming to that, meaning it is very unlikely holistic review in general is going to end.
We’ve covered the basic reasons why:
(1) Our HS curriculum and evaluation system is far, far away from having the standardization necessary to make that truly feasible;
(2) Given (1), such a system would be even more loaded in favor of high socioeconomic status parents;
and (3), far too many stakeholders like holistic review, or simply don’t care about exactly what is happening at Ivy+ colleges.
Obviously there will always be some families who think their kids deserved to go to a higher-ranked college than was willing to admit them (or who anticipate feeling that way), and they can be relatively vocal about it. But just because the people who like the holistic system currently see no need to be vocal doesn’t mean they won’t defend it if it is actually under serious threat.
And then the vast majority of people simply have no reason to care much about which of the 3.2% most competitive college applicants are part of the the 0.8% who enroll at Ivy+ colleges, and which have to “settle” for elite LACs and honors colleges at public flagships and merit scholarships at other “T50” privates and so on. The families who argue it was terribly unfair that they had to “settle” for what like 97% of families would consider a dream outcome are not likely making many friends that way.
Now, that doesn’t mean specific things like legacy admissions policies are completely invulnerable. But their vulnerability stems in part from the fact they really are not particularly necessary for holistic admissions. They might well have certain financial benefits for private colleges, but finding a way to make up for losing such a financial benefit is far less of a change than completely changing what they see as the optimal mix of students in light of their institutional mission.
Anyway, we’ll see. But I think the winnable battles are likely going to be limited to things like reducing or eliminating legacy preferences at some private colleges, or CS majors becoming sufficiently oversubscribed at various “top” publics that they become restricted majors with heightened academic entry requirements (yay?). And indeed such battles are being “won”.
But forcing all of these schools to completely change their class composition and admissions policies, just to better suit a tiny percentage of families? Good luck.
Even if (1) were solved in the US, the small size of the most selective US colleges relative to the US population means that they would still be picking from an overflow of applicants at or near the ceilings of typically available high school achievements and credentials. Even colleges that focused entirely or almost entirely on academic strength would need to do some subjective review to distinguish between such applicants (like Caltech does now).
That’s right.
As usual, we have a good example available in the UK system. Again, the constituent colleges at Oxbridge add up to about the same percentage in their college system as the Ivy+ and some of the “elite” LACs add up to in our system. And at that selectivity level, even the far more focused and standardized A-level system is not sufficiently fine-grained.
So, they do a couple things. For some courses, they have their own testing. But more broadly, the faculty in the constituent college(s) considering you for admission will interview you. Of course that is not your typical friendly alumni interview in the US, but still, there is a lot of research about interviews and how they are not necessarily as objective and unbiased as the interviewers intend. And yet Oxbridge really have no choice in the sense they need this sort of supplement to distinguish between applicants at the necessary level of selectivity.
Incidentally, you can actually see the effects of this in the aforementioned average entry UCAS tariff points reported in the Guardian. Meaning Oxbridge courses do not always have the highest entry tariffs, even when those Oxbridge courses are otherwise regarded as the “best” in their field. The simplest explanation for that is just that once you introduce these other important factors, custom testing and the interview, you will predictably get more dispersion across the other factors as measured by UCAS tariff points.
But anyway, yes, it is going to be practically impossible to truly do this in a completely objective way.
No, it’s the weakness of the breadth and depth requirements when applied to geniuses.
But if you think the most important or even relevant criterion for an admission system is “would it work for Albert Einstein or Arthur Miller as they were then”…I don’t know what to say, really. Frankly, not sure why Miller would even need an elite university education (as it is now, wasn’t then) to write. Or why any university (not even just elite universities!) should drop their minimum algebra requirements because Miller failed his classes, or their minimum history requirements, because 100 years ago, a 16 year old Albert Einstein couldn’t be bothered.
And those are holistic requirements. Not “rack and stack”.
I would expect any non religious school that takes public funding in any form to stay away as far as possible from rhetoric like this.
@NiceUnparticularMan , your reasons as to why the dystopic world I envisaged won’t arrive any time soon are persuasive; but ideas do have consequences, and it is well to imagine the consequences of the one being advocated by several posters. For them there’s nothing nightmarish at all about those consequences; indeed, they’re very much desired - the rejection of distinctiveness and subjectivity in our educational institutions, perhaps in our lives more generally, root and branch, in order to chase the chimera of objectivity. Perhaps we will soon all be living in Canada.
@Tigerle , will colorful expressions also be banned in institutions that “take public funding?”
Here’re some facts:
-
The wealthy does enjoy, as anyone would expect, an advantage in those academic factors. According to this study, the top 1% of the income distribution would have contributed about 7% of students based on academics alone. This new study doesn’t include the percentages for individual schools, but the original Chetty study actually showed that that percentage was even lower for schools that put greater emphasis on academics. For examples, that percentage at MIT was 5.7%, and at Caltech, it was only 3%.
-
When non-academic factors (including various preferences) are included in admissions, the new study found that about 16% of Ivy+ students are from the top 1% of the income distribution, representing “about 2/3 of the ‘extra’ rich kids at Ivy-Plus colleges”, in the words of one of the authors of the paper (and that fraction would be much higher if one uses Caltech, the school that gives the least amount of consideration to non-academic factors, as the baseline).
Here’s something from the study that I didn’t fully expect:
Non-academic factors don’t predict success later in life (and only academic factors do), even though these students have all attended Ivy+ colleges.
Interesting. But I wonder how they tease out that “ambition” and the causality of it. Using their example, wouldn’t the kid who wants to run for office be the one most likely to choose the Ivy as the one most interested in “how things look”?
That’s a fair point. I do think it is right that regional, academic specialization, and other differences would still exist, but . . . would the Colleges That Change Lives? I know those only serve a few people, but they illustrate the sort of abundant variety that currently exists in the US system, and that it appears some would want to eliminate.
And actually, we do have an idea of what their system would look like in the US–the land grant US public university system. I know there is still too much holistic review for them at the most “national” public universities, although there are also various auto-admit policies (like the Texas Top 10% rule, including as modified by their flagship). I’m also holding aside the universities that started off independent and became state-affiliated (UVA, William & Mary, Pitt, and so on), because that is a bit of a different world. But if you focus on the land grant universities, that helps us understand what their entire system would likely look like.
And as predicted, there is quite a bit of similarity in that part of our college system. Not that all these schools are identical, there are definitely regional differences, there are different specializations, and so on. But, they also have some at least very common attributes as a group.
Including that they are, according to this study, a little less good at value added in top 1% placements, controlling for other factors.
And that is probably where this would head in practice should these people have their way–like most of the world, we would end up with just a lot of large public universities as our “top” universities. And they would in that sense destroy the very thing these families covet, namely a more “elite” alternative to the public universities that are already happy to admit their kids.
But I suppose if no one else was getting something different, they may be OK with that.
That’s sort of the point.
The non-academic factors are not supposed to predict individual relative success later in life. They are supposed to “predict” how well students will benefit each other as members of a college community.
So what you should expect if these schools are right about the benefits of putting together a college community like that is not that the individuals who came in scoring higher will do better later, but rather that their graduates as a group will do better later in a way their incoming characteristics alone cannot explain.
And the study confirms that appears to be true, albeit only in the tails–meaning controlling for a long list of other possible explanations, it appears there is at least some sort of value-added effect in the tails that is not explainable by individual characteristics.
Now, of course one is free to hypothesize there is some OTHER hidden explanation for that value-added effect. But I think it is fair to say the data is at least consistent with the hypothesis that putting together a mutually-beneficial community according to the models used by these schools, and not just selecting the individuals most likely to succeed because of pre-existing individual characteristics, is creating a value-added effect in the tails.
And I wouldn’t rule out in the middle too, as the comparator Plan B colleges also do holistic review.
It’s better said that non-academic factors, as defined in the study, had little bearing on future success. I’m of the view that certain personality traits, which can’t be measured by a study such as this, have a fairly significant impact on outcomes in life.
Quick hypothetical.
You have a college of 3 per class, and there are 15 applicants you judge as all academically well-qualified. You choose one varsity captain, one passionate visual artist (who wants to major in Bio), and one person who did meaningful charity, on the theory they will help inspire each other.
After college, a study shows that each of them has outperformed what their incoming academics alone would explain, including relative to the 12 you didn’t admit to the extent they did not go to a similar college. But none of them outperformed the others.
This study would find no predictive significance to being a varsity captain in HS. Also none to being a passionate visual artist in HS. Also none to being involved in meaningful charities in HS.
It would not have shown your community-building project failed to add value during college.
The reason why I didn’t fully expect that particular result is because these Ivy+ colleges send disproportional numbers of their graduates into investment banking and consulting, two careers with very high earning potentials. Since hiring of investment bankers isn’t necessarily based on academic factors, I expected graduates of Ivy+ colleges would have higher earnings as a result.
I agree. I’d also add that these people would be successful regardless of which colleges they attended, if they attended any college at all.
I can’t agree enough with this. I know a number of folks who have been "exceptionally " successful, and they are all really smart. But they are also what I would call “the real deal”, in that they have great social skills. They are smart in how they handle their relationships, likable, and trustworthy. I know tons of people who are really smart and likable but not in the same way. They have done well in life but not like the first cohort. This is a form of character, for sure. And to be clear, not to say better than the second cohort, but one which leads to success.
No, in non religious institutions, people should be able to swear as much as they like.
Questionable hyperbole though? Absolutely needs banning!
Of course, if the college were not so small relative to the applicant population, it could have just admitted all 15.