College admission unpredictable

[Data10]

@1NJParent

^I don’t think anyone claims (at least thus far) that college admissions are random. But that doesn’t mean they’re predictable. Lack of randomness and unpredictability aren’t mutually exclusive.

If college admissions are as predictive as some of you claim, perhaps you should be in the college admission counseling business. Lots of applicants and their families would love some certainty.

If you’ve been reading the forum for long, you’ve probably seen many posters describe admissions decisions as random, crap shoot, luck, nobody knows, reach for everyone, and similar. I’ve seen posters claim that nobody has higher chances than the listed average admit rate for their score. So the kid linked in the early example who was accepted to all 5 of HYSPM would have no better than the previously referenced 10% admit rate for high scores at MIT and similar chance of admission for others, meaning that the chance of him being admitted to all of HYPSM would be astronomically low.

I think one of the key reasons for this belief is not understanding the admission process well and underestimating the importance of less visible and less objective parts of the application. For example, assuming a small different in stats is more important than large differences in out of classroom activities, large differences in backgrounds/evidence of positive character qualities, etc.

I agree with CU123 that there is a varying degree of certainty for different applicants and colleges, rather than absolute 100% certainty of decisions for all applicants at all colleges. This is generally how predictions work in any field that has a human component to the evaluation criteria, such as social sciences, as well as many fields with limited human components. For example, a meteorologist might say there is an 80% chance of rain tomorrow, rather than say weather is unpredictable because it is impossible to know the outcome with 100% certainty.

[racereer]

@Data10

The more interesting cases would be the unexpected decisions. Why wasn’t he accepted everywhere? For example, why was he rejected by Cornell, which is often thought of as the least selective Ivy when he was accepted to HYPSM + Penn + Brown, which are all generally thought of as more selective.

So what about the other schools? Do we think most of them could be yield protection or is it possible that some are just as difficult for admissions for CS especially the state schools since he was OOS?

[privatebanker]

The problem is we tend to look at the outliers as the experience of the majority. Yes. A Tahitian superstar with unrivaled excellence will have a more predictable path. An athletic recruit with outstanding credentials as well. Also for an Obama Hogg or Bezos or acclaimed teenage actor sure.

The academically unqualified and no preference student is also fairly predictable.

Once we eliminate that statistically small cohort, you’re left with a large swath of academic and ec excellent group. Even more accomplished than what we would call the average excellent. There are thousands of these students and very limited space.

The confusion and feeling of randomness occurs within this group.

Multiply this by the variables of many different schools and combinations of results.

Multiply it over several years of admissions cycles and there are many who can come to CC feeling this way.

Why x seemingly similar to y gets into A. But y gets into b and Y doesn’t.

Multiply this by how one views A vs B in terms of selectivity and attractiveness. Personally.

There are so many variables it’s impossible to quantify at an individual student level.

Especially if y has a slight edge in knowable facts, like gpa and board scores. It feels like a crap shoot to many because they can’t explain it.

Also people love their candidate and it gets personal.

[MWolf]

One of the reasons that there is a large element of unpredictability in acceptances is that nobody can tell how many people will apply to any given college. These numbers have been increasing for the most popular school fairly steadily for a few decades, at least.

As @privatebanker writes, at colleges with very low acceptance rates, the unicorns are pretty easy to predict, as are kid with clear hooks. So are the unhooked kids with stats in the college’s bottom 25% or lower. There is, however, a wide swath of kids, of whom the popular colleges fill about half their class.

When 8,000 kids applied to Stanford, kids who were in the “best” 10% of the applicants were fairly likely to be accepted, since there were about 800 of them, and and Stanford was looking to accept 2,000 students. However, when there are over 47,000 students,the “best” 10% make up 4,700, they’re more likely to be rejected than not.

By “best”, I mean “best fit” based on whatever holistic methodology the colleges is using, not by simple stats.

In 2016, about 36,600 students applied to Stanford, versus about 47,500 in 2018. That meant that there were more students who would be at all “ranks” of the applicants, meaning that fewer would be accepted, unless the were the very very top, which would still be relatively few in number. Somebody calculating the chances of a Stanford applicant in 2018, using the acceptances of the class of 2020, could overestimate the likelihood of acceptance for all but the very very top applicants.

Since Stanford no longer publishes the number of applicants, this will make these predictions even more difficult in the future, especially if the number of applicants grows at the rate it is doing now.

[ucbalumnus]

@jzducol

Actually, GoBears2023’s outcome serves as a perfect example of predictability. Cornell ED acceptance would have required Cornell to offer him a full ride. He had a hook—being Tahitian, who can be counted as native/indigenous american

Tahiti is not a US territory (it is part of French Polynesia), so he would not be considered “Native American” (but would be considered “Pacific Islander”).

[ucbalumnus]

@Data10

I think one of the key reasons for this belief is not understanding the admission process well and underestimating the importance of less visible and less objective parts of the application. For example, assuming a small different in stats is more important than large differences in out of classroom activities, large differences in backgrounds/evidence of positive character qualities, etc.

That, plus most outsiders have no way of determining how a given student’s subjectively graded application characteristics compare with the other students in the college’s applicant pool. Indeed, for most applicants, there is no one person outside the college’s admissions office who sees the entire application (even the student, who commonly does not see the recommendations, nor necessarily knows which potential recommenders will write the best recommendations for him/her).

So even though a college may score applicants’ on numeric scales for each aspect of their applications, and then those with better scores are more likely to be admitted, getting from subjectively-graded application characteristics to those scores is opaque to outsiders, including the applicant and those who would be advising him/her (except possibly at a well-connected prep-school, although the presumably-well-connected college counselors would not be as “outside” as most outsiders).

[Data10]

In 2016, about 36,600 students applied to Stanford, versus about 47,500 in 2018.

According to CollegeNavigator/IPEDS, the CDS, and other federal reporting sources. The number of applicants to Stanford was as follows. There was indeed a significant increase between 2016 and 2018, but no where near 36k to 47k. Stanford will continue reporting to federal sources like this, regardless of what information they publish on their website.

2016 – 43,997
2017 – 44,073
2018 – 47.452

A lower admit rate without change in admission criteria and preferences doesn’t inherently mean ability to estimate chances is better or worse. It just means the admission threshold has changed. A smaller portion of applicants are highly likely to be admitted, and a larger portion of applicants are highly likely to be rejected. The portion of borderline applicants for which admissions decisions are less obvious may or may not change. Rather than this degree of difference in admit rate, I’d be more concerned about year to year changes in the applicant pool and preferences.

I did not mean to suggest I can predict Stanford admissions with perfect accuracy. However, it has been my experience that admission decisions meet expectations far more often than not, rather than just a crap shoot of unpredictable after crossing a stat threshold for basic academic qualification.

I interview students for Stanford. As part of interviews, I rate students in multiple categories. I realize that these ratings and interviews in general are given little weight in the admission process, yet acceptance decisions so far have matched up almost perfectly with the median of these ratings, even among which students are deferred/waitlisted, rather than outright admitted or rejected. Stanford is one of the few highly selective colleges that defers truly borderline applicants and rejects the vast majority in the early round.

The correlation with decisions is especially surprising given how limited information I have about the applicant. I believe a large portion of this correlation relates to the interview rating categories being similar to what Stanford says they are looking for on their website and similar to some of the key more influential rating criteria/categories used by Stanford admission officers. Students who do very well in this criteria tend to be accepted, and students who do not do very well tend to be rejected. Obviously the latter group is larger than the former. It also helps to interview in a limited area, generally among the same few upper SES high schools.

[1NJParent]

To calculate the probability of acceptance at a specific college, one needs to know the following, even assuming s/he has seen the application package in its entirety:

1. How the college quantify, if it does, each element of the application. Only some, certainly not all colleges, quantify every element of an application. If the college doesn’t, the exercise to predict is already moot. If it does, how precise and consistent its method is is also highly uncertain in most cases. After all, quite a few elements in an application are subjective and don’t lend themselves to formulaic quantification.

2. How the college weighs each element of an application, even after each has been scored. We only have some information, only from statistical studies, on what Harvard may have done because of the lawsuit. We don’t know anything about the other schools. Even for Harvard, statistics are limited in scope (time, sample size, etc.) and are no substitute for actual practice used by Harvard.

3. How individual AO strictly follow the college guidelines and to what degree s/he is affected by emotions, personal biases, etc. AOs are humans after all.

4. How the dean/director of admissions exercise his/her power to override decisions on admissions at each college.

Can we quantify all these elements, beyond some guesstimate of saying an applicant has “low”, “medium” or “high” probability of admission? I doubt it.

[foobar1]

And if you add in the highly subjective “personal” rating (from the Harvard lawsuit) that is assigned to all applicants that makes individual admissions decisions extremely difficult to predict at a high-ranking college.