44 thoughts on “Dumbed-down U.”

1. If you apply this same analysis to the past doesn’t the shortage of top tier students just get worse the further back in history you go?

   • Not necessarily. What Arnold’s analysis basically boils down to is that the top universities expanded enrolment faster than student population growth. Thus, by definition, they had to accept students from lower percentiles than they used to to fill up all the slots.

     • Then why are acceptance rates at top universities so much lower than they used to be? The top schools are more selective now.

       • Let’s define “acceptance rate” here as “percentage of applicants who are offered admission by the college.” Acceptance rates can decrease at a college even when SAT scores of students who enroll don’t change, if more students apply for the fixed number of seats. The Common Application protocol has facilitated the practice of applying to numerous colleges.

         Trends in acceptance rates and trends in mean SAT/ACT scores at a college can diverge because the mechanisms partly differ for each trend.

         As I noted in another comment, the mean SAT/ACT percentile score of students at selective colleges steadily increased 1980-2007. (I don’t have data for the trend after 2007.) The number of applicants with high SAT/ACT scores increased faster than capacity at highly selective colleges, thanks to deeper national and international integration of the market for higher education. Market integration was facilitated by development of college rankings, innovation in information and communication technologies, decreases in travel costs, greater prosperity, etc.

         • Besides any increase in foreign student applications, we must account for other important reasons SAT scores are higher, of which one very large and one smaller reason are especially clear:

           First, the SAT itself was severely dumbed-down in 1995 and simplified again in 2005.* So even if selective schools admit roughly the same proportion of high-IQ applicants today as they did decades ago, those applicants will have higher SAT scores without actually being any smarter.

           Second, immigration has greatly increased the supply of high-IQ, which means high-SAT-scoring, applicants of East Asian extraction, who have, moreover, high average predilection for applying to selective colleges.

           *As demanded by colleges and universities that wished to hide differences in the proportions of top scorers by race and sex. The campaign to make the SAT less predictive was lead by University of California President Richard Atkinson.

       • Because the amount of universities a student applies to is much higher now. If each student goes from applying to 3 universities to applying to 6, that doubles the number of applications and cuts the acceptance rate in half. Yet there are still the same number of students and seats available.

         • Rohan,

           I don’t think it’s that simple. One student can still only go to one college at a time no matter how many they apply to. If students more than double the number of colleges they apply to, but that they couldn’t possibly go to, then won’t colleges find the same number of acceptances yield them fewer student actually attending their college and more choosing one of their other five shots?

           I’m thinking of this as a simple supply and demand issue. The number of elite colleges has not increased nearly as fast as the student population in the last couple of generations. Sure they are taking somewhat larger classes but I am skeptical that makes up the difference. Average IQ’s are going up, not down, due to the Flynn Effect.

           I haven’t researched any of the real numbers but the numbers John supplied seem to support this.

         • Rohan,

           I see that I bungled my first paragraph there. The points I made in that paragraph are in fact entirely consistent with what you are saying.

           I think that (more applications per student) is part of the story here but not the biggest part. I am skeptical that there ever were any glory days when the social sciences were much more rigorous at the college level. As evidence for that I submit that bungled first paragraph and blame it on my long ago college level social science education.

           I agree that too many people likely go to college now but I think it’s likely that that excess and dumbing down is seen much more at lower tier colleges and not at elite schools.

         • Greg G,

           My understanding (and sadly I’ve forgotten where I read this) is that the universities have gotten *really* good at predicting who will accept an offer. Apparently, part of the US News ranking is “percentage of offers accepted”, so the universities are trying really hard to avoid making offers which will be turned down.

           So what happens is that if you apply to a First-Tier and Second-Tier university, and you have the scores for First-Tier, the Second-Tier university may very well not give you an offer, predicting that the First-Tier will. Or possibly it won’t give you an “early acceptance” offer, but may give you an offer at a later date if you haven’t accepted First-Tier.

           So there’s all sorts of shenanigans going into this. The universities are trying to keep both their “acceptance rate” and “offer rejection rate” as low as possible.

         • @RohanV

           Really good?

           Here’s UCLA’s freshman class profile:

           https://www.admission.ucla.edu/prospect/Adm_fr/Frosh_Prof18.htm

           They admit just 14% of applicants.

           But of those, only 39% end up attending, which is about the same as it was 15 years ago.

         • In fairness, a 39% acceptance rate is pretty good, given that most kids are probably applying to, what, 7-10 different schools?

           I don’t think colleges are being much more selective, I think that, at least in part because of the common application, most kids are applying to more schools than they used to.

       • The argument I expect is that self-selection was higher before.

         I think self-selection is a pretty big factor. The acceptance rate at U of Chicago tends to be higher than at the ivy leagues (I think) but in my opinion is far more rigorous. High school students historically have been far more self selecting when it came to Chicago than Harvard or Yale.

2. Agree completely!

3. The mean SAT/ACT percentile score of students at selective colleges steadily increased 1980-2007. See Figure 1 in Caroline M. Hoxby, “The Changing Selectivity of American Colleges,” Journal of Economic Perspectives 23:4 (Fall 2009) 95-118, at p. 98.

   The article is available online, ungated, at the link below:

   https://www.aeaweb.org/articles?id=10.1257/jep.23.4.95

   Dumbing down, as you define it, did occur during the 1970s at selective colleges, except at the Ivies (99th percentile). (In the 1970s, flagship State universities continued to expand after the baby-boomers entered the workforce, and before the massive growth in international demand for admission to selective colleges in the U.S.)

   I haven’t checked data about the past dozen years. I won’t be surprised if the trend has persisted—despite reduced reliance on standardized test scores in admissions—because global integration of the market for higher education has outpaced growth in enrollments at selective colleges in the U.S.

4. The University of Massachusetts, Amherst, the flagship state university, has about 22,500 students. 3,373 (15%) are in the Commonwealth Honors College. Among other things, CHC students have a separate residential area, have some small courses, and have to do a senior thesis. There is a separate freshman admissions process (and a small percentage, ~12%, are admitted from students already on campus). I never know how much to believe a school’s website but it says CHC students work closely with professors, “delve deeply into topics of interest and contribute new and original knowledge to their fields of study.”

   The website provides figures: “entering Honors College students … Combined SAT: 1411”. I’m assuming that’s a mean, so not many would be below 1300. I don’t know the mix of courses CHS students take, nor whether some of them have special CHS sections. I assume most of are ordinary, potentially dumbed down, regular university courses.

   I don’t know how common this sort of thing is but the website informs me that there is a “The National Collegiate Honors Council, which supports and promotes undergraduate honors education”.

5. The progressive view is that college is for everyone. (Yes this very elitist and libertarian economist usually sound like they want to return to the glories of the US economy of the 1910s.) Also some of your percentages need to include foreign students here as well.

   That is a little strong here for most Progressives and I bet most Democrats are closer to Obama college and/or vocational training.

   1) Do we know the failure rate was correct in 1990? And how do you know it is inflated today?

   2) There are more office working class jobs that do require more writing and math skills. Yes corporate offices are too ‘Degree’ heavy here but there is a value to getting degree.

   3) Again, what do you propose for the 20 – 90% of the population? The top 10% will do fine in general but you need programs for this population. (More military, AA degrees and vocational?)

   4) In terms of young people failure, most of it comes in the post-HS years since 2000. From 1950 – 2000, the failures of young people centered in High School but now that has appeared to have moved to 18 – 25 years old.

   Why is this long term important to capitalist elitist?
   1) The best way to turn young people in conservatives (leaning conservatives) is have them get married, have children, a career, go to church and mortgage. And a lot of these occur together in mid-20s but now is moving to 30ish. When did the Boomers become Reagan supporters? When they started achieving the above in early 1980s. One way to get people away from identity politics and make put their identity with their career and church.

   • Anyway when you analyze the biggest name colleges how many of their students are legacies that don’t have the SAT to be accepted? 10 – 15% here?

6. You are so absolutely correct – you cannot provide a rigorous social science major to students who are below 1300 on the SATs.

   Recently I saw yet another ad for education, for the poor children of Africa. An immediate conclusion is this: rigorous education is wasted on those whose scores are below 1300.

   There has long been a huge civilization problem – “what do we do about the poor?”
   For many, the answer has long been education.

   But this has failed, because we can NOT ask the real important question – “what do we about those who are below avg in intelligence?”

   My own answer is now a Job Guarantee, which includes a National Service that accepts everybody and gives them work to do, and pays 80% of the lowest Army enlistment starting wage.

   Because there is a big racial issue involving IQ, this problem will not be honestly discussed. But it’s part of the truth.

   • I think that immediately jumping to the idea of a Job Guarantee or UBI builds bad habits. The implicit assumption is that people who score below average are always a net-cost and need to be isolated from broader society.

     I’d rather we have more efficient sorting so that people find jobs that maintain their interest and where they perform well. I think would also be beneficial if social status was less tied to job title – working a service job shouldn’t be shameful, especially if your life is otherwise fulfilling.

     You don’t need a degree to be useful to society, and you certainly don’t need a 1300 SAT score either.

   • I support a form of the Job Guarantee, but not that specific program.

     Rather than offer a job to all comers, I would have a program in which the federal government pays to create enough jobs such that, if all filled, unemployment would be very low, say 2.8%. That would be about 1.5 million openings in the current labor market. These jobs would be designed to be useful, even if MB < MC. As a quick example, one minor citizen headache is that people are forced to wait in long lines at the BMV. One set of these jobs could be to open supplementary BMV centers to keep waits down and provide more flexibility on hours. Another example could be auxiliary police officers to provide extra manpower for cities that really need it.

     Rather than a fixed wage, it would have multiple job and management tiers, with a weighted average salary of 75% of GDP per capita. These would be treated as actual jobs, so that a job guarantee stigma doesn't arise and it also would allow people to earn a comfortable middle class life.

     These jobs could be targeted preferentially at areas with high unemployment or poor job prospects, in particular the inner cities and small towns which have lost factory employment, helping to revive both sets of communities. Hopefully this would kick start business activity in these depressed communities.

     It would also be useful in the event of automation-related job losses, much more so than something like a basic income, as it is far more targeted to actual need and it doesn't create dependency.

     Ideally, we'd replace most income-based transfer payments with a three tier approach:

     1) A program to encourage employers to keep employees during downturns, modeled on Germany's Kurzarbeit
     2) The federal jobs program described above
     3) Disability insurance for those who cannot work

     • I love wonkiness too, but my new-ish support for a Job Guarantee is precisely because so many Libs, like Kling and C. Murray, seem to be in favor of UBI.
       In my view, it’s a unicorn UBI theory that will NOT be politically supported in practice.
       To get a program started, it needs to be simple and clear, to start.
       I support ALL Job Guarantee schemes that offer more jobs. But claim that the simpler proposals are more likely to get passed. However, state gov’ts could more easily get wonky clever, and that would be fine, too.

       JG is much better than UBI.

       • I agree that UBI doesn’t really have a way forward.

         If it’s small enough to be affordable, it doesn’t do much practically to help people and the left won’t be on board, esp. if part of the funding is through a VAT or cutting back other social benefits.

         If the UBI is large enough such that all of the other benefits could be safely repealed, it will probably be too costly and too large to work well without significant economic cost.

         I think a more limited JG proposal would be both more politically successful, more effective at helping people and more cost effective.

         For example, say you have a $15/hr JG wage with good benefits. Ceteris parabus, that JG probably pulls 20-30 million people out of the private sector plus maybe another 5 million into the work force. Of course the private sector won’t allow that many workers to be lost, and will raise wages, but it can only raise as much as business conditions will allow. Some $9/hr jobs will become $16/hr jobs, but others will be automated away or eliminated with business failure. Let’s say we only have 10 million people pulled out of the private sector and 5 million new workers, 15 million total. On a wages paid basis + 10% for benefits, that’s about half a trillion and probably double that to account for non-wage related costs of employing people or $1 trillion total.

         Now compare that to a limited form, which will start at 1.5 million people at 75% of GDP per capita (~$48,000). Throwing in $the same 33,000/employee of non-wage costs, the total program expense is $121.5 billion, about 1/8th of the take-all-comers JG program. If a moderate recession hits and unemployment rises by 3 million, the limited program would be expanded to 4.5 million and the total cost would rise to $364.5 billion, still barely more than a third of the full JG program.

         Automation is also well covered by this program. With the regular JG, the wage is flat $15/hr – not awful, but pretty meager. The limited form can have a higher wage, in this case roughly $24/hr on average, and with job tiers so that more experience workers who likely have more needs get even more than that, while new entrants to the work force get less because in part they need less.

7. I think Kling’s approach is a pretty decent Fermi Estimate. One of the key assumptions is that the number of students taking the SAT (i.e. 2 million) represents the total pool of potential post-secondary freshman. A quick Wikipedia search shows that the ACT test is an alternative to the SAT and that it has achieved approximately the same number of students as the SAT. Should the pool of freshman not then be 4 million instead of 2 million?

   • Good point, but they are not exclusive. Just one data point here, which says that about 13% of students submitted ACT only (in 2012).

     https://www.nytimes.com/2013/08/04/education/edlife/more-students-are-taking-both-the-act-and-sat.html

     So the pool should be something closer to 2.3 – 2.5 million.

   • That’s a good point; the top schools didn’t historically accept the ACT but they do now. Many students take both, of course.

     There are slightly more than 4 million 18 year olds in the US. Are the percentiles of test takers or vs the whole population?

     Alas, the whole thing is very counter-factual. Maybe UCSD could have a class with 1400+ SATs, but in reality the 75/25 numbers are 1410/1210. For TAMU:
     1360/1140. Brown’s numbers are not on Niche, but they were tied with Vandy which is 1570/1440. Princeton is also missing, Harvard is 1590/1460.

     • The only school I can find with a class like you’d expect is Caltech at 1590/1530. Of course, they don’t do admission based on legacy, donations, skin color, or skill with the fiddle or pigskin.

       I suspect Minerva would too, but numbers don’t seem to be available.

8. It’s also worth pointing out that the effects of dumbing down courses and degrees have all kind of pernicious ripple effects. For example, many jobs require applicants to gave a college degree as proxy screening mechanism, but the value of this signal has degraded substantially over time.

   That may not be so bad in the private sector, which can adjust to new educational realities and labor market conditions quickly, but it’s particularly hard on the government, where many of these requirements are deeply entrenched by being embedded in legislation, regulations, and case-law requirements, and there is little room for common sense judgment and distinction in terms of the degree of validity or worthiness of any particular degree at any particular institution (not to mention granular difficulties with alternative course tracks to the ‘same’ degree, and the open secret practice of grading on racial and gender curves). A degree is a degree is a degree, even though everyone knows that’s absurd.

   My own experience has been that plenty of 22-26 year-old applicants with 4-year degrees come in with less capability than the median 18 year-old graduate from my public high school. I’m talking about lack of completely rudimentary “3Rs” capabilities.

   The efforts required to deal with this problem and the corruptions required to make the process appear to follow unworkable official rules while necessarily involving personal preferences, unrecorded distortions in the selection process (which itself is prone to abuse), should not be underestimated and increasingly resembles dark humor jokes about life in Soviet bureaucracies.

   And of course, when having a “4-year degree” includes people who cannot compose a cogent paragraph, the pressure for genuinely talented young people to waste several more years of their life on obtaining a master’s or graduate degree of some kind is intense, as the law does permit distinction and extra points for those.

   The way around all this is to rely more heavily on standardized test scores, but of course those have highly disparate impact, and so the progressives naturally argue for their elimination (e.g., the GRE, the SHSAT, etc.)

   Finally, it’s worth pointing out that some standardized tests appear to be experiencing score inflation, and that the prudence of comparing percentile scores across generations depends on the validity of the assumption that the distribution of capability has not changed between generations, which is hardly beyond question.

9. The bottom 100 schools do not have to be useless.

10. Does the US economy have a need for anything close to 260,000 social science majors to enter the job markets each year?

    And doesn’t this social science fire hose produce the very sort of extreme theories that the previous few posts have presented?

11. The highly rated, elite universities face the same challenge as other luxury goods makers; they want to sell a lot while simultaneously maintaining the cachet that exclusivity brings. Somehow there is enough demand for $5,000 wristwatches to keep several makers around. If exclusive schools based admission purely on SAT scores with publically available criteria, the pool of applicants might shrivel. The value of a degree from an elite school is directly proportional to the number of applicants that they reject. The long term prospects of the people who are admitted matters as much as the timekeeping ability of the $5,000 wristwatch.

    • I believe Cambridge and Oxford have very objective admissions criteria— and it hasn’t diminished their brands.

      It looks Oxford wants US student to have a (1470 SAT or 32 ACT) and (3 AP 5s or(?) 3 SAT Subject 700s).

      • I think those are minimum requirements/guidelines. Applicants might still need to take an Oxford-specific test and would certainly need to be interviewed:

        https://www.ox.ac.uk/sites/files/oxford/Guide%20for%20US%20students%202017.pdf

        As someone who had a disastrous Oxford interview, the college fellows have a considerable, and entirely justifiable, amount of discretion when it comes to whom they admit. Even in England, and even 30 years ago before grade inflation*, acing your A-levels was not sufficient for Oxbridge.

        My guess is that US students who get in have (average) SAT scores a long way above 1470. I was, much later, on the other side of those interviews for a couple of years. We admitted one student from the US; she had a perfect score.

        [*”However, by 2009, grade inflation at A-level was so widely recognised that a new grade – the A* – had to be introduced for the following academic year.” https://www.newstatesman.com/politics/education/2019/08/great-university-con-how-british-degree-lost-its-value%5D

12. There isn’t just one kind of elite student. There are excellent Engineering students who will never be anything better than mediocre writers at best and excellent Humanities students who would struggle mightily if they had to pass a 100-level Calculus class. And then there are the Art, Music, Dance and Theater students — a whole different crowd there.

    • This is correct, but also missing the point. Engineering is a fairly highly-paid field, and not one you can learn on your own. Billion-dollar bridges and skyscrapers do not make good learning experiences.

      If I knew a teenager who wanted to become a writer, I would advise them to blog daily (preferably under a pseudonym) and read heavily rather than go to college. You can develop that skill through trial and error, then demonstrate it with a body of work. The same goes for art.

      Since almost anything pays better than (clothed) dancing, I really would not recommend that major to anyone who could not afford to simply waste the money (i.e., almost everyone). Many of the humanities fall into the same boat. I used to work with a facilities guy with a degree in art history; I have no idea whether he was an elite student of art history, nor was it ever likely to matter.

13. Random thoughts:

    – Even doubling the college worthy student population with foreigners and ACT takers would only get you from #66 to say about #150.

    – There are about 5,300 colleges in the USA. The majority are of execrable quality in international comparisons.

    – Undergraduates are the least of it. USA institutions print over a million graduate diplomas annually. Although some graduate degree programs are cash cows, in general, they are much more heavily subsidized. A good chunk of undergraduate tuition goes to subsidize graduate programs.

    – Higher education primarily exists in its current incarnation as a result of tax expenditures, direct subsidies, subsidies disguised as research grants, and federal student loans. Only tax reform will divert the resources flowing to the education system morass and produce the critical reforms necessary to prevent another generation from sinking into the mire.

    – The excessive scale of the US education industry is perhaps best understood in anthropological terms as something of a potlatch ceremony in which parents destroy their savings as a sign of status and prestige.

    – Delaying college entrance until age 21 would at least get students work experience and independence and pull the industry’s hands out of parents’ pockets.

    – I was recently invited to deliver a talk to a class of doctoral students at a mid-tier school. The experience was profoundly disheartening. If the VA and DoD cared about service members and veterans they would stop paying graduate school tuition for them. The course syllabus a painful read. So much waste.

14. What really dumbs down a university is accepting students who aren’t prepared (for example, don’t have the study skills) to get through college and graduate.

    The biggest single cause of this has to be affirmative action. Granted, there are a lot of disadvantaged and deserving minority people of that age, but admitting them to a university they’re sure to flunk out of accomplishes nothing and is a total waste of everybody’s time, money, and effort, especially theirs. This problem can only be fixed by fixing our K-12 schools (either within that system or by fostering competition against it).

15. I recently retired from a pretty good university (in the top 100 you link to, but near the bottom of that list). I taught English and Honors humanities courses for 41 years. Here are some reflections.

    Over those 41 years, the typical undergraduate became less and less well prepared for college. The students’ average reading ability and writing skills declined considerably. There were still plenty of students who could read and write at the college level, but in recent years I would put that number at around the top third of the student body, 40 percent at most. (That’s one reason I got involved teaching Honors humanities–almost all those students could do college-level work.)

    It seemed to me that the average undergraduate got next to nothing out of college. He/she could not read at the level necessary really to understand what was going on in the works I assigned. Students often could parrot back what their teachers told them, but they did not understand the works themselves; there’s a difference between the ability to grasp the teacher’s summary of a work and the ability to read it on one’s own after a bit of guidance and instruction. And once again, I was teaching at one of these top 100 national universities. One wonders what went on at lower-level schools.

    • Did you also teach STEM students that had to take some mandatory courses from the humanities? If so, how well did they do and how high/low was their interest level?

      • Yes, I taught quite a few STEM students. Their interest level varied, but I felt that part of my job was showing them how interesting these philosophical, artistic, and moral issues were. I got through to some, and they were among my favorite students.

    • The parroting back is the safe way to get through a required gen ed class. My niece was having trouble with a professor in a course named “Liberal Arts”. She wasn’t doing well because she is conservative. I advised her to decide whether the argument was worth losing the money paid in tuition to a bad or failing grade. She adapted, got through it and never looked back.

      In a recent interview on the Free Thoughts Podcast with Jason Brennan and Philip Magness on their book ‘Cracks in the Ivory Tower’. Brennan related they had found that 50% of students get nothing from college (excluding the magic parchment), another 40% get a little and 10%, maybe, get a lot from college. These latter are the ones who continue on to become professors and think tankers.

16. It’s going to get worse before it gets better.

    The colleges are unable to improve themselves, under current incentives.

    Revamping gov’t Civil Service tests to be more comprehensive tests of multiple subjects, allowing standardized comparison between different courses from different colleges might help — especially if one can take such tests based on on-line learning.

    On-line testing of mastery, with more AI evaluations of writing samples, is likely part of the near future of mixed steps for progress.

    • I’m more optimistic about post-secondary education. I think it goes hand in hand with my opinion that the main functions of education are ranking and matching. Transforming, that is, the teaching the types of skills people achieve mastery in, is important but further down the list. I also make a distinction between specialized knowledge, the mastery type stuff, and shared knowledge, the kind of stuff that helps us communicate and coordinate with other people. Professors lament how little students retain (i.e. they lack mastery) but this discounts the amount of shared knowledge that the student may have acquired. If mastery truly is important, education should be structured around spaced repetition learning. I suspect that greater post-secondary mastery would not translate into more/brighter entrepreneurs or more productive workers.

      What I worry about is K-12 education and the number of highly intelligent individuals that ignore/abandon the science and math aspects of their liberal arts education.

17. Why UC Berkeley is no longer in the list…

    https://www.berkeleyside.com/2019/07/28/uc-berkeley-kicked-off-the-u-s-news-world-report-rankings-of-best-u-s-colleges

18. IMO the anathema of being dumbed-down, is the ability to genuinely do critical thinking — but I see little that in today’s current academic system.

    As a physicist and citizen that greatly concern me.

    On a slightly off-topic matter, I really liked your report “The Three Languages of Politics” (2013). Have you updated that, based on recent developments?

    Feel free to email me directly.

Comments are closed.