I’ve been importing IPEDS data, which is a treasure trove of stats about college. One notable consumer of this data is U.S News, which uses IPEDS to rank schools. Looking over their methodology, I was surprised to learn acceptance rate is no longer included in their rankings:
Acceptance rates, yield and alumni giving are among several factors no longer part of the ranking formulas. Data for them continues to be updated for informational purposes on school profiles on usnews.com and in Academic Insights.
Acceptance rate is defined as the number of students accepted to a college divided by the number who applied. Yield is enrolled students divided by accepted students.
Acceptance rate thought experiment
Imagine a universe with just two schools. Both enroll 1,000 students each fall and there’s no particular distinguishing feature except:
- School A gets ~1,000 applications a year and
- School B gets ~2,000 applicants.
Which is likely to be a better school? My grandfather had a joke he told whenever he saw a cemetery. “That must be a great graveyard. People are dying to get in!” Along the same lines, there must be a reason one school is getting more applications than the other and a reasonable assumption is that applicants see something in the school that isn’t otherwise obvious.
But suppose I told you the two schools had the same budget and School A spends as much as possible on instruction while School B pays for applications. I don’t necessarily mean giving students cash directly, but spending on various forms of advertising, including non-academic programs such as sports and clubs. Assuming your goal is to get the best education, School A turns out to be a better choice even though it’s not selective at all.
Acceptance rate data
Since I have the data, let’s look at the top 10 schools by acceptance rate:
| school_name | applied | admitted | enrolled | acceptance_rate | yield |
|---|---|---|---|---|---|
| Minerva University | 20816 | 208 | 150 | 1.0 | 72.1 |
| Fairfax University of America | 382 | 8 | 4 | 2.1 | 50.0 |
| California Institute of Technology | 16626 | 448 | 224 | 2.7 | 50.0 |
| Harvard University | 61221 | 1984 | 1646 | 3.2 | 83.0 |
| Stanford University | 56378 | 2075 | 1736 | 3.7 | 83.7 |
| Columbia University in the City of New York | 60879 | 2404 | 1540 | 3.9 | 64.1 |
| Massachusetts Institute of Technology | 33767 | 1337 | 1136 | 4.0 | 85.0 |
| Yale University | 50060 | 2289 | 1554 | 4.6 | 67.9 |
| Brown University | 50649 | 2562 | 1717 | 5.1 | 67.0 |
| University of Chicago | 37522 | 2039 | 1729 | 5.4 | 84.8 |
Until I ran this report, I hadn’t heard of either of the top two schools. Fairfax University is in the town where I graduated high school and just down the road from George Mason University where my parents taught. But it’s enrollment is much too small to be included in the US News ranking. It doesn’t take many applications to be highly selective when you only admit 8 students!
Minerva is another oddball. According to Cryptid56 on a Reddit post:
But there are a couple sketchy things I wanted to run by you all. [Minerva is] known for having a super low acceptance rate but I have heard they skew it by counting any submission of info on their website as an [application][1] (email, name, anything). Also, they are super vague with their outcomes stats among other things.
So you can see that my thought experiment isn’t exactly hypothetical. Still, the 2–10 school ordered by acceptance rate do pass the smell test. The top schools in the country are highly selective.
Acceptance rate is a backward-looking statistic
Of course this is not surprising. There’s no real limit to the number of schools a student can apply to and name-brand schools benefit from people taking a chance on their dream school. According to a 2023 article by Jessica Grose, changes in the application process have magnified the phenomenon:
Since applying is easier (or at least involves fewer postage stamps than it did back in the day), does this mean more students who have relatively little chance of getting into these selective institutions, based on their grades and résumés, are applying? Connie Livingston — who was an admissions officer at Brown for 14 years and is now the head of college counselors at Empowerly, a private counseling company — told me that before the pandemic, qualified applicants made up around 75 to 85 percent of the applicant pool.
Now she thinks there are some students who are “throwing their hat in the game just to see what happens” and that “the number’s probably down to about 60 percent, 65 percent of applicants,” who meet the recommended standardized test scores and grades of the schools they apply to. Livingston cited the common app as a reason for the influx of applications, but she also noted that the loosening of standardized testing requirements played a potential role as well.
A good ranking system should surprise us from time to time.[2] We want to see an up-and-coming school that’s poised to break into the rankings and acceptance rate isn’t helpful for that.
The OP said “acceptance”, but I believe that to be a typo. ↩︎
I spent a good deal for time looking for a BIll James quote to that effect, but I can’t seem to find it. Perhaps I hallucinated it. There’s probably a useful quote in “A Mathematical Theory of Communication” but I decided not to get distracted by looking for it! ↩︎