The thing is that “given that you are competitive” it is still hard to predict which one you will get into. Our results were 1 acceptance and 2 WL from your Ivy example, and 2 acceptances and 1 WL from the LAC example. We (nor the CC) could not have predicted these results. I know lots of students that got into 1 Ivy out of several.
2 Likes
I have no formal numbers on that, and I think it depends on whether you are talking Ivy and Ivy Plus unis, or similarly-selective LACs. Anecdotally, based on the cases we know about, I would say for NESCAC LACs it may be roughly half, possibly even more in some cases, most of those recruited athletes. I’d say it is probably lower than that for at least most of the unis and non-NESCAC LACs.
Like, if you told me one of our kids was attending Colby, not knowing anything more I would guess there was a good chance they were a recruited athlete. Claremont McKenna, probably much less so.
2 Likes
“Fit”. “Overlap”. People have different ways of describing the same thing. For example, if Brown is your “dream school”, it might make sense to bundle it with Wesleyan University, Tufts, or even Amherst for their open curricula, adjacent commercial strips and liberal student bodies. IMO, your college counselor is making a simple concept sound more complicated than it really is.
3 Likes
Yes, the risk is that in the college specific sections of the Common App or unique college applications, the tendency/temptation is to cut and paste across multiple applications as the app count gets higher. For highly selective schools, the decisions will likely turn on LoR’s and essays (and how the EC’s line up) as the “finalists” are probably academically indistinguishable except for extreme outliers. AO’s are internalizing 2 questions, how will my school benefit this applicant and has the applicant made a case for this and perhaps more importantly, how will this student benefit my school’s community. When they sense the essay is mostly a cut and paste job, it hurts the applicant’s pitch that this is an ideal fit for both me and the school.
1 Like
So that is consistent with what they are suggesting. Meaning in this case, adding a 4th school in either of those categories that you personally ranked after all of the first 3 would have made no practical difference, as there was at least 1 acceptance out of the 3 you personally ranked higher (by hypothesis, obviously I don’t know how you actually ranked schools by preference).
Your example does, though, suggest it should be more than 1. I also think your Ivy example potentially shows it could be a good idea for it to be 3 and not 2, although that is tricky because “Ivy” is a bit broad in some respects. Like, I think it is fair to say some Ivies are a bit more selective than others, and you might want at least 2 in each Ivy selectivity range. Or maybe not, if your top Ivies include at least 2 in the less selective range, or you prefer some “Ivy Plus” colleges in that range to Ivies, or so on.
For the record, that is me, not our college counselor. I am personally interested in things like probability theory and I find it interesting and helpful to translate the counsel she gives us into such terms. But her job involves counseling a wide variety of families, most of whom have no such interests.
So yes, this is a simple concept, and I think one that can be understood and implemented without all the conditional probability talk.
Still, my experience is some families who end up rejecting this counsel are basically doing some probability reasoning of their own. Meaning it just sounds wrong to them that rolling a die or flipping a coin 10 times won’t give you better chances of 1 good outcome than rolling/flipping it only 3 times.
Being me, I immediately recognize that as confusing fully independent probabilities with dependent probabilities, and I know the math works out very differently in cases where there are dependent probabilities.
Still, does me pointing that out actually help anyone who is thinking like that? No idea, but at least it is something they can consider.
2 Likes
There is enough randomness in outcomes that applying to multiple schools makes sense to a certain point. We all know of the kid who got into Harvard but was rejected by Stanford while another classmate had a flipped outcome. AO’s after all are human beings and who you draw as a first reader (and how your app resonates with him/her) can make a significant, maybe even determinative, impact. It would be interesting to see if we looked across Academic Index deciles (and somehow controlled for rigor), what percentage of applicants were shut out from all highly selective schools (pick 10 or 20) they applied to, how many got into only 1 and how many got into several. That might give us a better picture of the interplay between independent and independent probabilities.
I overplayed my hand and was for that reason unclear. Yes, of course, many people decide to not apply to excellent schools of all varieties because of their location. Often, the issue is the school is too isolated, or, less frequently, too urban, or too sleepy or whatever. But it doesn’t typically mean, IMHO, conveniently located near an R1 or other school. My point in the hypothetical move of Amherst to Middlebury is that not being near UMass or another school would not deter the typical Amherst student from applying to and attending Amherst. That’s all I’m trying to say there.
Sincere answer: of course not. I’ve been thinking we were talking about type of college and not where the colleges are located geographically.
Maybe for things like ultimate frisbee, etc., sure. But you’re not going to find club soccer, lacrosse, basketball, baseball, track, etc. You can find those at the club level at large universities, but typically not at LACs. At any rate, you’ll be able to definitively confirm that one way or the other when your son starts making some decisions.
Good luck!
2 Likes
Not enough for data but this year I know of 2 valedictorians from competitive schools who got shut out from T20s. Great kids, but unlikely to meet any institutional priorities.
3 Likes
I’ve backed off a little bit on my anti-shotgunning stance. In a perfect world we could concentrate on fit all the time, and not aspire to attend a highly rejective school because of notions of prestige … but some kids and families abide by them, and with single-digit admit rates and holistic policies obfuscating the clarity of the process/decision-making, shotgunning increases the chances that one of those highly rejective schools will fill a need with the applicant.
Of course, the question arises: if you are shotgunning, can you really put your all into every single essay? Can you really show each school how you fit there, and how you would contribute? That you really want to be there? I suppose some kids can pull it off while others struggle. Regardless, it’s bound to take a great deal of time trying to personalize 20 essays.
I don’t think anyone should feel they must attend an Ivy Plus (U or LAC), without compelling needs beyond “prestige”.
I still think the best way to form a list is to choose your schools based on:
- Cost/affordability
- Academic fit (majors, curriculum, class size preference, etc.)
- Environment/Location
- Vibe (Greek, athletic, sociopolitical, etc.)
If, doing that, you end up with an all-Ivy list, so be it. Just make sure you have a couple safeties lined up as well, that you like and are affordable.
I am just a bit more understanding of the shotgunning approach. We can choose our schools based on fit… but we don’t know which highly rejective schools will return the favor.
3 Likes
I note again that is consistent with the dependence theory in question. Reach shut outs (No, No, No, No) will, unfortunately, be pretty common. Multiple mixed reach decisions (Yes, No, No, Yes) will be common. Maybe a few people will get into all their reaches (Yes, Yes, Yes, Yes).
But, apparently, it is relatively rare for someone to be something like (No, No, No, Yes), at least if that Yes is unhooked and at a school generally as or more selective as the first three.
1 Like
Fwiw, my kid considered 4 of the NESCAC LACs and 11 other LACs in the MidAtlantic, Midwest, and upstate NY area. I would say he was pretty typical at his school.
As for odds, what his CC said was this : Your stats are in the range for any of these schools. Most have about 500 in a class. That means 250 men. Of those, pull out every category you do not fit - recruited athletes (maybe 40-50) legacy/major donor applicants, (10?), another Questbridge/FGLI (20?). You are now vying for one of 175 spots. Those men in the preferred group will determine the needs admissions still has for the rest of the class, whether instrument played, academic interests, geography, URM, ability to pay, etc.
It’s possible that they’ll have your doppelganger 5 times in that preferred pool (especially true for a sporty white kid from the northeast who is not a recruit) or it’s possible they’ll still be looking for kids just like you. And then, it’ll depend on who is in that remaining pool. All of that is beyond your control and a matter of chance. It was the CC’s advice, based on a preference for LACs, that he apply to at least 10 schools, simply because of this.
At bigger schools, the numbers are bigger and the “class-crafting” is a bit less personal.
As for overlaps, there are some in the NESCAC, but plenty outside it. Tufts and Brown, Colby and Dartmouth, Wesleyan and Vassar, for example. I don’t think most people think about the athletic conference as much as you think as they apply. And within it, there are kids who simply like Maine and will apply to 2 or 3 BBC schools and there are others who apply to Williams and Bowdoin, or Tufts and Wes. It depends so much on what priorities are.
I think you really need to look at each one. And back to the consortium idea, if you are feeling that “LACs are great, BUT…”, check out places like Rice, Rochester, Tufts and Case Western. All a bit different, but a far cry from a huge state flagship that offers everything but may limit access by major.
12 Likes
I think that happens all the time. Honestly, your point is confusing me. You are saying someone is unlikely to get into one of four reaches?
2 Likes
I suppose it is possible, but these days this seems highly unlikely without at least one admissions hook.
1 Like
Anecdotally not our experience. When my son attended Bulldog Days, a relatively large number of attendees he met were cross admitted to at least one of HPS plus other normal suspects. This year, the 2 interviewees of mine that got into Yale also got into Harvard, Stanford, Brown, Penn, Duke between the 2 of them. Both were unhooked for all the schools. One of the statistics Yale admissions pays close attention to is the yield on cross admits. It must be a significant number for them to track so closely.
BTW, one chose Yale over Harvard and the other chose Stanford over Yale.
1 Like
Right, I understand kids receiving multiple acceptances – I specifically meant acceptances from all highly rejective apps. (OP’s “Yes, Yes, Yes, Yes” example)
Gotcha. I was just trying to illustrate that it seems to me that dependent probabilities play a big role in ultimate success when looking at highly rejective schools. It goes with my normal advice to apply to at least 1 reach early (not necessarily ED) kind of as a canary in the cage to see how strong the application is both in terms of how high the applicant can reach and if there are potentially problems with the essays are even LoRs.
3 Likes
Getting back to your experience with those two interviewees – man, those must have been tough decisions. I think if I were back at that stage of life, I’d rather just get into one of the highly rejectives – let them make the decision for me. hehe
I can’t imagine choosing between H/Y/P/S, or Duke/Northwestern/Brown. No thanks. I’d always wonder about the road(s) not taken.
4 Likes
I don’t want to belabor this if it is not clicking for you.
But briefly, conditional probabilities are of the form if A happens, then what is the probability that B happens?
If A and B are independent, that means the probability of B happening given A is the same as the unconditional probability of B. Meaning what is the probability I will get no heads in two flips of a coin? Well, if it is 50% each time, and they are independent, then there is a 50% chance I will get heads the first time, still 50% the second time, so only 25% chance of flipping no heads.
But if B is dependent on A, then the math is different. Like, suppose there is an unconditional 50% chance of A, and an unconditional 50% chance of B, but if A doesn’t happen, now there is only a 10% chance of B, and vice-versa. What is the chance of neither A nor B? Well, 50% again for A (or B, in this case it doesn’t matter where you start), but then 90% chance of not-B if not-A, so there is a 45% chance of neither A nor B–higher than 25%. And actually, 45% is not much different from 50%. Because of the assumed dependence.
Let’s do a third. If independent, you have only a 12.5% of no heads. Now let’s say the conditional probability of C given not-A AND not-B is only 1%. OK, then 45% chance of not-A and not-B, then 44.55% chance of not-A, not-B, and not-C. C barely matters now, because of this dependence relationship.
OK, so suppose you have an unconditional probability of admissions at a given “reach” school of 10%. If they are independent, then for 3 “flips”, you have a 72.9% chance of getting into none, so a 27.1% chance of getting into at least one. Sounds good! But wait, if I apply to 10, then I can get that up to 65.1%! And so on.
Except they are not independent. We actually kinda knew that, because then everyone would apply to enough reaches to get into at least one, and we know that isn’t how it works.
OK, so suppose instead it goes like this. 10% chance of reach A. If not-A, then only 3.33% chance of B. If not-A and not-B, then only 1.11% chance of C. In this model, each no is just reducing the conditional probability by 2/3rds.
OK, 3 applications, you get a 13.9% chance of admissions to at least one. Better than 10%. Not nearly as good as 27.1%.
OK, what about 4? Well, continuing the model, if you got not-A, not-B, and not-C, then the conditional probability of D is down to 0.37%. Total probability of at least one went up from 13.9% to 14.2%. This is becoming a meaningless distinction. Do all 10, and you will see it quickly converges.
OK, so that’s how conditional probability math works. And that was just a simplified model, and in real world cases it should definitely be applied flexibility.
So, got an unusual special circumstance that explains low grades one year that may or may not help? Maybe apply to more than three reaches–although again, if it doesn’t work the first four or five times, it may just not be happening. Applying to schools which deemphasize academic screening and emphasize personal fit? OK, maybe apply to more of those reaches too–but again, four or five well-chosen are probably about the limit of helpfulness.
And so on. Our college counselors are not dogmatic about this, and if you have any sort of good reason to apply to four or five reaches (at least realistic reaches, as in you have that 10% unconditional probability or similar), they will likely support that.
But if you are a fairly standard applicant looking at fairly standard reaches, they will also be supportive if you cut that down to two or three that fit you particularly well. Because they know if it doesn’t happen at two or three well-chosen colleges where you submit a really well-done application, there isn’t much chance adding more will meaningfully change your odds.
Which is because something like that dependence model above fits this situation better than the coin flip model. Which we kinda already knew, right?
1 Like
Consider the point belabored, and let’s move the conversation forward
5 Likes