How many colleges are your kids applying to?

Parents Forum
41

The probability models in question don’t require identical probabilities, just conditional ones.

In terms of the dating analogy, suppose there were 10 women looking for a date, and 100 men looking for a date from one of those 10 women. The men can only accept one date.

We know 90 of the men will get no dates no matter what they do, since there are only 10 dates total. Like even if all 100 men ask out all 10 women, necessarily 90 of the men must get no date offers.

OK, but then if the asking out process happens simultaneously, particularly desirable men may get more than one date offer. As an aside, this will mean the women have to figure out how to make more than one offer. Perhaps they will invent yield models, waitlists, binding early decision, and so on.

OK, so men, understanding the odds, then each ask 4 women. Some immediately get multiple offers, lots of women like them. Some only get one offer, but that works. But how many women will get the best date they can out of the 400 total offers?

Well, potentially all of them, particularly if they have good enough yield models and waitlists and such.

And this is where women having different preferences actually helps speed this along, if the men can figure out who they might match up better with. So maybe the women try to describe what they prefer, and the savvier men do a good job paying attention. Again, potentially all 20 dates available get filled this way after as few as 4 offers each by the 100 men.

OK, but if 4 doesn’t quite do it, what about 5? 6? Each increment represents 100 more offers total, many more opportunities for women to find their matches by the time yield and waitlists and such unfold.

Again, plausible models are going to make it extremely unlikely the women will need each man to make 10 offers in order to find the 10 they collectively prefer. Of course 1 a man won’t likely work either. But the matching process will likely be complete with something substantially more than 1 and substantially less than 10. Call that N.

OK, so now you are some individual man, and I gain the knowledge that you did your best to identify the best matches, asked out N, and got zero offers. I now know you are in the 90 and not the 10. That’s going to happen to most of you.

OK, last step, but then you ask out the remaining worse matches too, so ask out all 10. Does that change my assessment? No. We already know from the fact your best N matches failed that you are in the 90.

That analogy may be more trouble than it is worth, but that’s getting closer to something realistic. The really critical underlying facts to make such models plausible are that there are far more applicants hoping for a match than matches available, and that the entities deciding who will get offers can communicate individualized preferences, and then use yield models and waitlists and such to get the offers they actually want to take.

3 Likes
42

I’m sure the moderators will soon step in and say CC Parents Forum is not a debating society, so I will just leave it by saying we may be talking past each other.

It just seems super obvious to me that different colleges have different (and opaque) preferences around applicant’s intangibles, and so even if you’re a very strong high school applicant, you will likely have your heart broken if you fixate on any ONE of them (aka “I must get into Stanford”).

So if your goal is the undergrad experience provided by a highly selective college, isn’t it wiser to apply to many highly selective colleges – as many as you can do well without burning out – with the hope that what YOU can offer as a compelling candidate will be reciprocally viewed favorably by at least ONE of these selective colleges (“okay, maybe Stanford doesn’t like what I have to offer, but perhaps Yale will”). After all, it is unlikely that the intangible desires of Stanford and Yale and __ and __ are all exactly the same.

Remember, I started this part of the thread with the following premise, which means that we are already well beyond the candidate being academically qualified and deep into holistic review (and subjective) territory: “if your student is very strong (e.g., 1540+, top 3% of senior class, pointy EC’s, compelling recommendation letters), then . . . .”

1 Like
43

See the quote above.

As fascinating as the off-topic discourse may be, it deserves its own thread if it needs to continue. But really, it seems like back and forth restating opinions. So for this thread, let’s return to the OP’s question.

2 Likes
44

We’re just finishing out the selection process for D26 - and applied to 15. All of them except 1 were targets/safeties. That 1 was a marginal reach and was her only waitlist. D26 is set in what program wanted and what type of school she wanted to take it at (small to mid size, more rural than urban, etc etc). Beyond that it came down to total expected cost and while we’re unlikely going with the cheapest one we were able to trim the list by about 1/2 when we got the acceptance/merit. All of the schools were Common App and many of them didn’t have an application fee (total of $420 paid).

The brain trust on this site was great at giving potential colleges to go after if you give more details about the type of program, stats, etc.

4 Likes
45

At my school we have some kids who’ve applied to 10, 15, or 20. About 15 years ago we had one kid apply to 30 schools! But, these are the outliers. I’d say most kids about 4-10. My kids were in that range. Our college counselor says to go for quality over quantity. Only apply to schools that you actually want to attend and have at least a chance at getting into. Also, you want to think about finances. Don’t apply to MIT if you have no chance of getting into MIT. Or don’t apply to University of Washington if you’re on the fence about going there.

46

I mostly agree, but your opening paragraph and #7 drive towards the current problem. As overall numbers of apps skyrocket, you’re almost forced to play the game because targets might become lotteries, safeties might not be as guaranteed. My child is high school class of ‘28 so I’m watching this unfold and wondering at what point the madness will stop.

We are very early days, but her initial list is about 8-10 schools which includes two reaches but is mostly targets and two safeties. However, in another year of thousands more kids are applying to those same schools her app could get buried in that avalanche. Will she end up having to contribute to that by applying to 15 or more to cover her bases even when we aren’t talking Top 20 schools?

1 Like
47

More is better. Everyone claiming that anything more than 3-4 apps is pointless assumes that the correlation amongst the choices, pairwise, is one, and therefore the overall probability of a desired outcome does not improve. In reality, nobody knows the true correlation between any two pairs of choices. It is very strongly likely to be less than 1. Therefore increasing the pool of colleges is strictly better, until the quality of thE application to each college starts falling due to lack of time. People can put in up to 20 or 25 before this starts happening. Of course you need to start working earlier on the apps — maybe even starting in the spring of Junior year. This obviously assumes there are 20-25 schools that you would pick over your safety. For some people there are 20-25 schools over their safety. Also it is expensive to go and visit all these schools ahead of time to figure out if you want to attend them. It is cheaper to just apply to any schools that is reasonably likely to be better than your safety, and then go visit the small number of acceptances that you have to eventually pick one school from. You only need to pre-visit your likely EA or ED choices.

3 Likes
48

We went for 6 including one auto admit on stats as a true safety, though being in California also had the ultimate CC-UC (or SLO) pathway as backup. Despite a potentially problematic profile my kid ended up getting into all of them, including a reach and low reach/high target. So I think 8-10, if well chosen as it sounds like your list is, should be fine. It’s also what the school college counselors suggested for kids (school has a proper college counseling office).

1 Like
49

So if the correlation was perfect, you would need only one.

With imperfect correlations, it becomes a math exercise. On fairly plausible, Bayesian-style models of conditional probability, the odds of getting a different result on the next attempt conditional on all the prior results being the same are going to become extremely low after something like 3-6 applications of a roughly similar sort (“roughly similar” being key to the model being plausible). It is true that technically, you may never quite exhaust the incremental increase in probability. But when the modeled incremental increases are extreme small (like say on the order of 1 in a 100000, or worse), that is negligible for practical purposes.

That’s one concern, but I don’t think the only one. What I have personally observed is that kids who are “shotgunning” Reaches may not always be doing a careful job choosing Targets and Likelies as well. Like, they might not even have any Targets or Likelies, just a long list of Reaches and then a couple “Safeties” they actually don’t even like. And in fact, they may not even be doing a careful job choosing all their Reaches.

So one common outcome of all that I have seen is a kid who gets into none of their Reaches, and is really depressed about having to go to a Safety they don’t like. Shotgunning didn’t work, and they didn’t have a solid Plan B.

Another is they DO get into a Reach, but they realize it is unaffordable, or doesn’t really have the academic programs they want, or so on. So they STILL end up at a “Safety” they don’t even like, and are extra depressed they got an offer they can’t take.

Of course in theory, a kid should be able to FIRST choose a great list of Targets and Likelies, THEN choose a great short list of Reaches, and then add more carefully chosen Reaches indefinitely.

In practice, I think a lot of shotgunning kids are basically convincing themselves, even if not consciously, that: (a) as long as they apply to enough Reaches, surely one of them will come through; (b) getting into ANY highly-ranked/highly-selective/”prestigious” Reach will make them happy by proving their worth to peers, family, and so on, and also guarantee a lucrative and high status career.

So for me, encouraging a different way of thinking about lists isn’t just about maintaining application quality, although I think that obviously becomes a real problem as well. It is also about trying to get kids to be a lot more reflective about the process, in a way that may end up producing more offers they actually can consider, and are excited to consider.

4 Likes
50

My youngest daughter applied to 20 targets and safeties, most oos. She was accepted into all of them, was offered merit at all but one (Rutgers, flagship, not surprised about lack of merit). Because she had a budget, and none of the schools offered automatic merit for gpa and scores, she had to apply to several.

2 Likes
51

The Bayesian analysis doesn’t apply because the applications are not happening sequentially for the most part. There are only two stages — early and regular. So there is a certain amount of Bayesian analysis that is applicable, but very limited. My kid had 17 apps for a full list, but eventually stopped at 6 due to getting some choices in the first round. People are rational about these things

2 Likes
52

Chasing merit is absolutely a very good reason to have a much longer list of Target and Likely colleges. All the logic about admissions per se isn’t really relevant to that, and indeed it is facially different because it isn’t a bunch of yes/no decisions, but much more complicated “how much” decisions.

5 Likes
53

It does apply in the sense we can take any N similar applications as the “first”, and then ask about any other applications as the “next” as a hypothetical.

Assuming the conditional probabilities work in a plausible way, if we took any random six and hypothetically NONE of those were acceptances, the conditional probability of any of another six, or indeed another ten, or twenty, being acceptances can be very low.

But realistically, we can do even better. For example, good college counselors can help applicants determine which specific Reaches are relatively better or worse bets for them as an individual. This requires going beyond raw numbers and understanding their institutional policies and tendencies in a more sophisticated way. Kids and parents can also do this for themselves, but it can take quite a bit of work to get up to speed. That said, communities like this can help.

OK, so now the “first” six are no longer random, they are actually your six best bets after doing this more sophisticated analysis. This is making it even less likely that if none of your six best bets work, one of your worse bets will.

Part of the point of this observation is it DOES take real work to sort colleges like this. Again in my experience, shotgun applicants may not be doing that work. And it may well be that even with, say, twenty Reach applications, they haven’t included their six best bets. They may in fact have included as few as zero of their best bets, because there are typically a lot more than twenty Reaches they could consider.

By the way, having a potential list of 17 but being able to stop after 6 is actually consistent with this approach. That’s an example of one of the six best bets working, which means the other eleven didn’t really add anything in the end. Not that there was anything wrong with identifying those 11, it just is an example of why the math does in fact tend to work out this way.

1 Like
54

My older applied to 7 and it was too many. My requirements (these did not need to be unique schools, but I wanted all of them to be met) - one school close to home that was affordable (we know somebody who developed a health problem during their senior year and needed to stay home to get it sorted out while attending college, and kid did not have a year’s worth of CC to take), one big school, one small school, and one inexpensive school lest our finances change and our budget suddenly become very small. The big and small were because I’ve seen kids suddenly want a particular environment. Since kid was applying to some OOS ACC/SEC schools that had become unpredictably more competitive recently, they applied to 3 in this category. So, they wound up with 2 very affordables, one nearby, one small, and 3 big. They would have been better narrowing it down more, but until National Merit was announced they didn’t know that the affordables would both basically be free. After being accepted at all of them and finding that merit made all but the most affordable basically the same price, it was hard to eliminate options.

Younger has the same rules, but will need fewer schools to get there. With her major, the nearby can be the local CC because there is still a year’s worth of classes that she can take and still transfer. There is no need to apply to the CC unless it’s needed at the last minute, and she already does DE so she’d just be changing status. The ‘affordable’ is likely to be a small school where she is considering playing her sport. She has many more options because there are lots of places to study to be a teacher, compared to an engineer. At many small NAIA schools, there are great scholarships for kids with good scores and grades. She will apply to one big school just so that she has the option if she decides that she’d rather have that experience than play her sport. She may pick another school to apply to - maybe where friends are, or a second possibly sport school? But, at the most she’ll be looking at 4, and possibly only 2, applications.

1 Like
55

The first six are the first six because the state publics need you to be early. The next 11 or so in our case are all the privates that cannot be early after the EA. I don’t understand your example. There is not a lot to learn from some rejections to help with the next 11. There are at most 3 batches if you count ED 2. There is only one stage of learning, and the learning is imperfect because you can’t learn much from the publics for lessons about the privates etc. For the most part the 6 and the 11 are not interchangeable

56

Also, there is much less information about selectivity of non-automatic-for-stats merit than there is for admission. So the applicant chasing merit may not have any idea whether sufficient merit is reach, match, or likely.

3 Likes
57

Absolutely. Outside of the automerit colleges, the discretionary merit colleges are often largely a black box. People will look at prior results for others being posted online, but there are going to be hidden variables, and colleges can and will modify their discretionary merit policies anyway as they are reacting to what is happening in their competitive market. I gather some colleges even do this on the fly in a given cycle, which is why sometimes some admitted applicants get an unusually late merit offer, or increase in their merit offer.

So yes, I think in most cases the best you can do is find out if the sort of merit you need is realistically possible or not. Categorizing the odds beyond that may be largely impossible, and I agree that helps explain why a longer list typically makes sense.

1 Like
58

Times have changed! Most kids seem to apply to 8-12 schools now just to stay competitive and secure better financial packages.

1 Like
59

@Mjkacmom — I’d like to hear what your method is for deciding a college is a “match”, even if it’s approximate. Thanks

60

My kids made their application lists, and had majors which were widely available at the mostly public universities where they applied (business, science). They had a budget, and sometimes they’d submit applications to schools they were not going to attend regardless. My youngest daughter received financial offers that I never even saw, just said over budget (Lehigh, Bentley). They knew we were only paying so much, the rest was up to them ( we would co-sign and pay interest but wanted to be taken off after 12 months of repayment). Only one stuck to the budget, but he is the least likely to be earning as much as the others.

1 Like