So now Columbia U won’t have to figure out, how best to “produce” these figures without running afoul of their own Math department.
Basically, they were already ahead of the game 
Tbh, I don’t think general college rankings matter very much. Things like pay and academic influence vary by field and are usually more important than a general ranking. Pricing will also differ, some may qualify for merit, need, in state, etc. Overall, most, if not all, factors that are included in general rankings are irrelevant to students.
1 Like
What if they simply dropped any still-remaining factors and pretense, and JUST published their ranking.
1 Like
Possible but obviously not likely looking at the schools ahead of it. Its social mobility rank isnt higher than its peers and from what I can tell their graduation rate hasnt increased so the only thing left is the reputation score. Which could increase but thats not likely.
The proportion of small class sizes at my latter school, University of Alaska Fairbanks, was much higher than at my original school, Ohio State. So UAF would have scored better with US News in that respect. Most of my class sizes at UAF were in the 20-25 student range, and some only had a dozen or so students. Only a few of the largest introductory classes at UAF had 50-60 students. At Ohio State, most of the introductory classes had hundreds of students in a single lecture hall.
Yes 
I also have a theory that if you create a new ranking list that assigns 50% weight to the reputation survey score and 50% to endowment per student, you’ll end up with rankings very similar to USNWR’s current list.
2 Likes
That would suggest “honest” class sizes, unlike a school with lots of 19- and 29-student classes that looks like it is gaming the rankings.
I don’t either. Most students go to a particular school because it’s close to home, far from home, their friends go there, they like the sports teams, they like a particular program, they like the campus, or it’s a cheaper option. The proportion of prospective students who factor rankings into their decision is going to be very small.
1 Like
Class of 2026: 60,377
Class of 2027: 57,129 (-5.4%)
However, the cause is not certain - it might not be related at all to US News rankings:
2023 is the first reporting year, where colleges had been explicitly instructed to ONLY count COMPLETED applications. Some colleges had quietly been counting all (started) applications (resulting in a smaller acceptance rate), even if a certain percentage never bothered to submit everything that was needed to actually be considered for admission.
So comparisons to prior years can be misleading, if one doesn’t know what number a college had used up to 2022. It’s quite possible for a college to actually have an increase in completed applications, even if on paper it looks less than last year’s “more liberally counted” figure.
3 Likes
USNWR changes their methodology most years. When making these updates, USNWR intentionally chooses a weightings formula that allows the colleges USNWR readers expect to appear at the top to actually appear on the top. I doubt this will change in the most recent revision. However there may be notable changes further down in the rankings.
2 Likes
But enough to be noticeable that some colleges actively game the rankings, and colleges near the top of the rankings attract enough applications that they are difficult to get admitted to.
2 Likes
Yes, some game the rankings, but I doubt that class size is one of those things they try to game. More likely they’re trying to do things like get more alumni to contribute money, and increase applications so they can reduce the acceptance rate. Even if schools did reduce class sizes in an attempt to improve their rankings, smaller class sizes also do improve learning.
A college that may split 48 students taking a class into two sections each gains a USNWR ranking benefit from having one section with 19 and one section with 29 instead of two sections of 24 students. Or it can gain an even greater USNWR ranking benefit by having two sections of 19 and leaving 10 students unable to take the class.
How Northeastern University Gamed the College Rankings quotes former Northeastern University president Richard Freeland saying that class size of 19 was a way to game the rankings. If you look in Northeastern’s class schedule, you may find that 19 is an unusually common class size.
That page also mentions other things that can be done. Admitting students to start in the spring (sometimes with a fall study abroad programs or some such) can be a way to load-balance the semesters (because fall would otherwise be more heavily enrolled than spring), but can also be a way to game the USNWR rankings by admitting lower stat students in the spring where their stats will not be counted with the stats of fall start students that are counted for USNWR rankings.
2 Likes
One school does not a pattern make. And like I said, smaller classes are better for learning, so splitting the larger class into smaller classes actually improves teaching and learning, even if that’s not the main motivation for doing so.
Is there actually an example of a school that did this, or are you just stating a hypothetical? I would think a school in a situation where they have 48 students would have three classes, each with a cap of 19. You may end up with 3 classes of 16 students; a class with 19, a class with 17, and a class with 12; or all kinds of other combinations, but again that’s a better teaching and learning environment than being in a class of 48. Or if they did only have 2 classes and 10 students were left out, then those leftover students could take the class in the future.
A large school could teach all of its basic introductory classes in big 500-1000 student lectures, and then offer a whole bunch of little freshman / sophomore seminars with 19 students each. This would result in a lot more bonus points, compared to spreading out faculty resources to make introductory classes smaller (50-60).
Also, smooshing as many students as possible into a single giant lecture would seem to be more favorable on the “percentage of classes with small class size” metric, because that’s just one class…? A ratio of one giant class to 9 tiny ones would mean that 90% of classes offered are tiny, even though the majority of students would be enrolled in the giant classes.
6 Likes
Or have a giant lecture with a whole bunch of discussion and (if applicable) lab sections limited to exactly 19 (not 20) students. After all 500 or 1000 is the same as 50 for the USNWR class size categorization.
4 Likes
Yes, if discussion / lab sections count as “classes,” that would definitely be the most rankings-optimized way to teach any given subject to 1000 students. You would just need 1 professor and 53 TAs.
2 Likes
I think most schools, and by most I mean Cal and other underfunded schools, won’t optimize for that. They would rather optimize for other directly beneficial things - <8 hours per week per UGSI so they don’t have to provide tuition waivers. Mainly private schools play the rankings game (NEU, Columbia etc.) the publics have other priorities.
I agree, I don’t think that schools with huge classes (like Cal) are organizing their schedules in order to optimize for rankings. They are just trying to get that material taught somehow, within budget.
My point was just that I don’t think a “percentage of small classes” metric is very meaningful, because cramming a lot of students into a single large lecture can make that number look more favorable.
I think that students looking for smaller classes are actually more interested in something like “average class size in the average student’s schedule.” Since class size often gets smaller as you get deeper into a major, you could break out the numbers by lower division / upper division, or by year (freshman, sophomore etc).
1 Like
I agree with that. Just pointing out that the rankings and the associated gamesmanship is mainly a luxury only the privates can afford. All the more reason I really respect public schools who punch above their financial weight.
5 Likes