However, the acceptance rates for these schools occasionally may invert from that suggested above. For the classes that entered in 2020, for example, Middlebury’s acceptance rate was 22% and Carleton’s was 21%.
Then, there’s always the Groucho Marx take on the whole subject which is, “I would never join a club that would have me as a member.”
1 Like
It’s nice that the strategy worked well for your son, but it’s not a reliable strategy to expect that your admission chances will be similar for colleges with similar admit rate. Different colleges often have very different applicant pools. A 30% admit rate among an applicant pool largely composed of high achieving students is not the same as a 30% admit rate among an applicant largely composed of low achieving students. Colleges also often emphasize and deemphasize different criteria when deciding who to admit.
For example, the following colleges all have had approximately the same overall admit rate in latest IPEDS – Berkeley, Colgate, Navy, and Tulane. However, a particular high achieving student is likely to have very different chance of admission at the different schools, in spite of having similar admit rates.
Tulane offers ED 1 , ED 2, EA, and RD. Their latest IPEDS lists an ED acceptance rate of 68%. There have been posts about Tulane reps stating an RD acceptance rate of 1%. I can’t verify whether accurate. However, I do believe that Tulane places a greater emphasis on choice of ED vs EA vs RD than the other colleges above, which is consistent with listing level of applicant’s interest as “important” for admission decisions in CDS. The CDS also lists things like volunteer work as being important for admission.
Unlike other schools on this list, Colgate is need aware. Having high financial need decreases your chance of admission. In the old Chetty study, Colgate parents had the 2nd highest median income of all colleges in US. More Colgate students came from top 1% income families (>$600k income) than bottom 80%.
Berkeley takes the opposite approach to financial need as Colgate and may give a significant boost for criteria well correlated with being financially challenged. Unlike Tulane placing strong emphasis on applying early and level of applicant’s interest, Berkeley doesn’t give the option to apply early. They also emphasize and deemphasize a variety of different criteria. For example, Berkeley is test blind, so an applicant won’t get a boost from high SAT/ACT scores like they would elsewhere. Berkeley selectivity can also be vary dramatically depending on which major you apply to and whether you are in state vs out of state.
Unlike Berkeley, Navy places a strong emphasis on test scores (test scores are relatively low due to differences in applicant pool, rather than lack of emphasis) , but they also place a strong emphasis on things like physical fitness, including CFA score based on 1 mile run time, how many pull-ups you can do, etc. Congressional nomination is also important.
3 Likes
Of course that was a very unusual cycle due to COVID.
Overall, here are the last 10 CDS for Middlebury and then Carleton, with total number of applications and then acceptance rate in parens:
14-15 8195 (17%) / 6297 (23%)
15-16 8891 (17%) / 6722 (21%)
16-17 8819 (16%) / 6485 (23%)
17-18 8909 (17%) / 6499 (21%)
18-19 9227 (17%) / 7092 (20%)
19-20 9754 (15%) / 7382 (19%)
20-21 9174 (22%) / 6892 (21%)
21-22 11906 (13%) / 7915 (18%)
22-23 12952 (13%) / 8583 (17%)
23-24 13297 (10%) / 6464 (22%)
Pretty interesting. Things were pretty stable before COVID with Middlebury getting more applications and having a somewhat lower admit rate. They appear to have had a different strategy for dealing with that COVID cycle, and then Middlebury’s applications really exploded post-COVID. Carleton’s were up too, but not that much, until 23-24. I am not actually sure why Carleton’s applications dropped that much in that cycle
But reportedly they were back up to 7869 in 24-25 (18% acceptance rate), and Middlebury’s were down to 12541 (12% acceptance rate). But that still means Middlebury’s applications appear to have gone up a bit more in the post-COVID world than Carleton’s.
If my memory serves Carlton said they didn’t count partially finished apps for the 23-24 year so that is why the numbers were lower. I remember that popping up when my S24 was looking heavily at Carleton. I also remember thinking I doubted there were that many apps where the student never submitted his transcript.
1 Like
Not sure this would really be accurate. For the class that started in '20, they would have applied in late '19 and early '20. Acceptance decisions would have prior to any understanding of what COVID was and any impact it might have on the Fall 2020 cohort.
Where numbers could certainly have skewed some would have been admitted student % - by 5/1 deadline families would have had only a slightly better definition of what COVID would mean.
I think we see the COVID impact in the table for the '21-'22 cohort.
Have a look at this article from April 2020: Admissions office markets Midd in a year roiled by economic downturn, public health crisis - The Middlebury Campus
Even with heightened virtual efforts, however, colleges around the world are facing Covid-19-related challenges that could lead students to either enroll in the fall as planned, or defer their admissions — a process experts refer to as “melt.” As a buffer, Middlebury admitted more students than usual to ensure it would reach its desired yield of 725–740 enrollees. The result is a pool of admittees that is nearly 700 students larger than last year’s and an acceptance rate that surpasses 20% for the first time in more than a decade.
Here’s how I think about it-
Missouri S&T has an admissions rate of over 70%. There are hundreds of students at U’s with far lower admissions rates studying engineering whose ^&* would be handed to them if they were at Missouri.
You see a kid with a 3.6 in nuclear engineering from Missouri? That’s impressive. Far more impressive than the kid with the 3.9 from a bunch of private U’s with far lower overall admissions rates-- and less rigor, easier to game the GPA, etc. Ditto aerospace, petroleum, and a bunch of other programs.
So while these periodic discussions on CC have a certain appeal, my “lived experience” hiring people from all over the country (and parts of the world) tells me that there are far more useful metrics than acceptance rates. And that the USA really needs to beef up its 3rd and 4th grade math curriculum so that people understand that a “rate” is just a function of a numerator and a denominator.
4 Likes
There’s a little bit of circular reasoning going on here. You began with the premise that there is something called, “locational arbitrage” which defines a college’s peer group (in this case, Conn College) as private colleges in the northeast. Then, you question that location on the basis of the college’s competitive edge (or lack thereof) with regard to the same peer group. So, which quality are we solving for, location or acceptance rate?
FWIW, Conn has nearly two dozen highly selective private colleges within its general geographic vicinity. Stated another another way, Carleton has maybe six?
1 Like
Yes, my understanding is there was in fact ultimately a drop in applications most places, including as people started withdrawing applications they had previously submitted. Colleges were further uncertain what their yields would look like, so adopted different strategies.
Middlebury as described apparently adopted a strategy of admitting more applicants. going from 1498 the year before to 2022 that year. Their yield was 30%, down from 40% the year before, so they ended up with 602 students actually enrolling, right around the same as the year before with 605.
Carleton also admitted more students than the year before, 1460 to 1400, but obviously not as a big a difference. Their yield was 31%, down from 37.5%, so they enrolled 453 students rather than 525 the year before.
Anyway, this all helps explain why that was an anomalous year. Middlebury was able to keep its enrollment more or less level despite fewer (unwithdrawn) applications and lower yields, but that required admitting a lot more people. Carleton faced the same situation in terms of fewer applications and lower yields, but did less to correct with more admits, and so experienced a much more significant enrollment drop.
The effect on fall 2020 data was a number of enrollment deferrals (due to uncertainty and many closed campuses) and the resulting somewhat larger than typical number accepted off waitlists over summer 2020. Waitlist acceptances are probably not very noticeable in the overall acceptance rate.
The effect on fall 2021 included slightly reduced available seats due to the 2020 deferrals then coming in, as well as test optional policies increasing both apps and later yield (TO admittees were more likely to yield).
So that is not quite right. The semantics does not really matter, but I am identifying two different dimensions along with you might identify college clusters.
One is something you might roughly call “academic peers”. Even that is too simplistic, but for that purpose I was looking at things like the last list of unlocked US News peer reputation survey results I saw.
The other dimension is location.
My point is very frequently, if you identify academic peers along the first dimension, you can detect patterns of different admissions statistics that seem to reflect differences along the second dimension.
Of course this is not a scientific study so there is a lot more work you would have to do to really confirm all this to a high standard. But to me the patterns are so evident that I felt comfortable using it as what I would call a working hypothesis.
Exactly. I was referring to this situation when I wrote:
Of course my comparison there is to St Olaf–which again I identified using the US News peer reputation survey results (if you are curious, in that version St Olaf got a 3.6, Connecticut College a 3.5, which is not in my view a significant difference, whereas Middlebury and Carleton both got 4.3, which I do think is a significant difference). But St Olaf of course has essentially the same location as Carleton.
OK, so then you can observe St Olaf had a higher yield in its last CDS than Connecticut College. Does this show St Olaf is an inherently better college than Connecticut College?
And my answer was no. For the same reason you cited–Connecticut College undoubtedly is facing a much more intense competitive set for regional cross-admits. So a lower yield in that situation would be expected from what were otherwise academic peers.
OK, so what does all this mean for a given applicant? Well, obviously it depends on how strongly they care about location. But, I do believe that a given applicant who had good numbers and other solid qualifications for both St Olaf and Connecticut College would often have at least less uncertainty about actually being admitted to St Olaf. Maybe not no uncertainty, but often not as much as they would have at Connecticut College.
And if you accept the premise these are academic peers, this is what I mean by locational arbitrage.
2 Likes
Safety - we use the same some/many would assess who may use it as a factor for consideration. Campus safety resources, area based statistics / info for that immediate area as well.
I grew up in a town with what many consider a highly regarded university. Knowing that area full well, having worked at that university and knowing what has taken place there, not a chance she’d be going there. Campus is too spread out and much of it requires students to leave the campus and walk to get from building to buildings.
Obviously what is and isn’t safe is opinion based. Urban/Suburban is not a factor, at least for the schools that D26 has picked as finalists. In fact, her first pick is urban.
Our school probably has the internal stats for acceptances at her picks, but those numbers are not shared. Yes, we use Maia Learning and other resources for acceptance, but CDS is still the driver for us. Should it be? Maybe / maybe not. That’s our go-to for now.
I agree geography is a significant factor in the difference in application counts between Middlebury and Carleton.
I think another factor is Middlebury doesn’t require a supplemental essay while Carleton does. For some schools dropping required supplements was a Covid-era change, but I don’t know when Middlebury dropped theirs.
On the other hand, Middlebury has an application fee while Carleton does not. My guess would be no essay would increase applications more than the app fee would decrease, especially since they make it easy to request a waiver for hardship, but that’s a guess. Perhaps it’s less of a guess to say that different types of students would be dissuaded by an extra essay versus an application fee. Personally I think supplements make sense at colleges that prioritize fit.
Thought experiment: “Carleton College is my dream school, but in case I don’t get in there, I’ll also apply to…”
I think different people will answer that differently. For some it might be Grinnell and Macalester. For others it might actually be Swarthmore and Haverford. Or Amherst and Wesleyan. Or indeed various universities. My S24 actually ended up deciding between offers at Carleton, Vassar, and WashU (those were the three we visited post-offer, plus St Andrews but that was before he knew about the other offers).
…but not Connecticut College.
That certainly seems to be how it worked at our feederish HS. It seems like a lot of people “added” RD applications to Middlebury once they did not get into their ED1/REA/SCEA school. And part of the reasoning is it was so easy to do that.
And to my knowledge, every one of those kids got waitlisted or rejected. Waitlistings seemed particularly popular.
So I am not sure how Middlebury knows, but it does seem to know when kids from high schools like ours are doing something like that. And yet it seems happy to encourage them.
And I suppose some of those waitlistings eventually work out. Not a lot–Middlebury in the 2023-24 CDS offered 2778 waitlist spots, 2734 accepted a spot on the list, but that only turned into 36 admitted.
By the way, Carleton only offered 782 a waitlist spot, only 356 accepted, but then 35 were admitted. Very comparable outcome with a very different path.
Anyway, either way I assume these colleges all value those waitlist admits. So if their respective strategies are working for them, I am not going to question it.
But I also think in practice, you should understand your last-minute throw-in easy RD app to Middlebury may be basically applying for a spot on their waitlist.
I mean I am sure it has happened, but probably there are not a lot of lists like that.
I note it can be fun to play around with Parchment for these purposes, just to see where they have data. Carleton has enough cross-admits in their database with, say, Macalester to say it is statistically significant. With Wesleyan they report a number but it is not statistically significant, implying only a few cross-admits. And then with Connecticut they report no matchups (which doesn’t mean there were none in the real world, just none in their data). Whereas Wesleyan does have some cross-admits with Connecticut, but the report is the same as with Carleton and Wesleyan (meaning only a few). And so on.
Since I was suggesting in this case St Olaf would be an arbitrage opportunity with Connecticut College, I checked that too–no matchups.
Of course in this context, this is a feature not a bug. If you are applying to multiple schools as a hedge against each other, I would think the less overlap they have in applications, the better. Because that would likely help reduce correlated outcomes, aka statistical dependence, which is the mathematical expression of why you can’t guarantee admissions to one of a set of similar colleges just by applying to enough of them.
And that’s sort of the point of all this–to try to figure out when it might be helpful to zig when others like you are zagging.
So, say if you like Wesleyan as a Reach, maybe you also apply to Connecticut College as a Target (with merit potential). And I am not discouraging that.
But if you feel like it, maybe you also add St Olaf as potentially a Likely (depending on your qualifications). Because my bet is the sorts of kids who are reaching for Wesleyan will be less likely to have St Olaf on their list too, and I think that is potentially going to help you get a good option to consider.
1 Like
That is wrong on so many levels.