'It's a crap shoot': Father of girl who wrote scathing letter to Ivy League colleges

QuantMech · May 27, 2013, 9:48pm

Clearly the college admissions process is not entirely random. It is not random among students who have essentially identical qualifications in terms of standardized test scores and GPAs.

However, I think that there is an element of randomness in the process, in that I do not believe that the outcome is perfectly replicable. That is, suppose that an experiment could be run in which precisely the same set of applications was sent to the same admissions personnel, but there were variations in terms of the order in which the applications were read. Does anyone believe that precisely the same set of applicants would be offered admission? I do not. In this sense, I believe that there is an element of randomness on the process, because it is not truly deterministic.

The National Science Foundation ran an experiment about 30 years ago, concerning the review of proposals in my field. After a round of reviews that were run normally, they ran a second round with the same set of proposals, but different reviewers. There were some proposals that were recommended for funding both times and some that were recommended for funding neither time. However, as I recall some appreciable number of proposals “flipped” between the two sets of reviews.

I think that the same thing happens in college admissions. Applications that are discussed by the entire committee may hit the table at favorable or unfavorable times. The representative for a geographical region may have a bad cold when he/she reads a particular set of applications. There may have been various family events, good or bad, that affect the mood of the admissions representative, and his/her reaction to particular elements of the application. Some reviewers may put drama/musical theatre/debate/dance in each of the 24 possible different orders of preference. Similarly with sculpture/water colors/oils/charcoal, or Chinese/French/Spanish/Latin, or violin/oboe/bassoon/timpani. To say nothing of different preferences for activities among the different categories! I am certain that the admissions personnel try to compensate as far as possible for any bias that might result from their preferences. But it is difficult for me to believe that they eliminate it completely. So just having some difference in the set of people who read an application in depth could affect the outcome. This all assumes that the application is quite strong to begin with.

calmom · May 27, 2013, 10:15pm

Why do you have such a disdainful attitude toward high school kids? They don’t need to set out their life’s experiences – they can write an essay about 5 minutes of their lives. It doesn’t have to be about something special – there’s a Dad on CC whose daughter wrote about leaving her dirty socks around the house. We CC old-timers all remember that, because the Dad was so angry he yanked out the cable from the computer, almost preventing the d. from submitting the app in time! The kid was accepted to Stanford (no thanks to Dad, who wanted her to study harder for the ACT instead of volunteering to help back stage for the high school play).

Unless your kid has been hiding under a rock, the kid must have done or seen something that is interesting enough to fill a 500 word essay. My son volunteered one day at Habitat for Humanity. He came home and had stories to tell me. My daughter signed up to work at a polling place on election day — it was an opportunity offered to every high school senior at her school (and probably every other school at town) – and she was even paid for it. Boy, did she have stories to tell at the end of the day! Neither kid used those for college essays, but they certainly could have.

Here are some things they did write about: My son wrote a “challenge you have overcome” essay about how he had let down a friend by not completing his end of a group project, and how he dealt with that. My daughter answered a question about daily routine by describing her 40 minute daily commute to schools. One school asked what she did the previous weekend – she wrote about going ice skating. When asked to write about a literary character she admired, she chose a character from a children’s picture book.

The “vast majority” of kids are not going to get into Ivy League or other highly selective colleges. But one of the qualities that the Ivy League definitely wants is students who have their eyes open, are participants in life, will recognize and take advantage of the opportunities before them, and have communication skills that are developed enough to be able to write a short essay sharing their point of view.

I don’t have a clue how many kids fit that description, but certainly there are enough to fill up the ranks of the Ivy League several times over. I do understand that many kids at age 16 or 17 aren’t there yet – but that’s a personal quality that the Ivy League does look for. That’s why they emphasize qualities like “leadership”. (Which doesn’t necessarily mean getting elected class president or team captain-- it can be showing initiative or maturity in a many different contexts.)

Yeah, that’s why I went and spent like, $30 or $40 or something like that for each kid on books by experts in this stuff for them to read.

If you can only buy one book, by the way, go with Harry Bauld’s.

Madaboutx · May 27, 2013, 11:48pm

And yet, Quantmech, has there ever been a solely deterministic method used for selecting kids for elite schools admissions?

I remember the good ole days when blacks couldn’t attend, women couldn’t attend, methods were create to limit Jewish presence. The poor were excluded for a very long time. Legacy held great weight and wealth and power reigned supreme.

These things still exist but the admissions process is far more inclusive today. Perhaps more inclusive than ever. Kids from low income families, black families, public school and trailer trash pedigrees all now have a shot of admittance. It’s more meritocratic than ever before.

As I’ve said, admissions is a human process. Put away the charts and graphs and lets look at how people make us feel. It’s as random as people are random but that is part of the beauty of it. Not relying on ugly, cold figures to tell us how to think and feel. Hearts and minds connecting is great.

QuantMech · May 28, 2013, 9:46am

My statement is merely that there is an element of randomness in the selection process. Actually, I think this is inevitable. There are some students who would be selected with any reasonable set of circumstances surrounding the consideration of their applications, and others who would not be selected in any case. However, I do think that for some students, the “luck of the draw” applies.

I think it is better than the “olden days” when I think the East Coast prep schools could pretty much place their students in the Ivies. I am pleased that it is more inclusive.

I think that “how people make us feel” is inherently pretty random, because it hinges so much on who the “us” are. I doubt that the admissions personnel at any of the colleges are so uniform in reactions that the element of randomness is removed.

LakeClouds · May 28, 2013, 10:12am

These rather prosaic topics prove the point others have been making about admissions being a crapshoot.

calmom is like many who believe they have mastered a random process rather than just having had sheer luck in their outcomes.

Here’s an excerpt from an interesting study on overconfidence:

Humans have a natural tendency to be overconfident. We tend to overestimate our knowledge of a particular subject, overestimate our skills, overestimate our ability to control outcomes and underestimate the risks involved in uncertain situations. In one classic study (Svenson, 1981), a group of American drivers was asked to rate their own driving skills as being either better or worse than the average person. If the study participants were accurately describing their own driving skills, we would expect that approximately 50 percent of the respondents would claim to have better than average skills and 50 percent would claim to be worse than average. However, an astonishing 93 percent of the participants claimed to be better than average drivers, which demonstrates a significant tendency towards overconfidence. Similar results have been found when asking people to rate their own intelligence, popularity, or the time that it takes them to use an ATM

calmom is like so many stock-pickers, gamblers or others who play games of chance and then insist they have a “system” that will overcome the randomness involved in the process.

xiggi · May 28, 2013, 1:37pm

Prove the point? How silly a conclusion!

Some asked Calmom to “document” what she meant, and she answered by providing details. Did she claim to have mastered a random process? Or did she shared how she applied common sense to a process that appeared arcane by researching the issue and developing a strategy that might help her children. Along the way, she made the efforts of attending the workshop of Van Buskirk, listened to experts such as Harry Bauld, and picked what appeared to be helpful to her kids’ scenario.

Does this work for everyone? In a world of single digits admissions, the answer has to be a resounding NO. But does the alternative of flying blind and relying on faulty assumptions (as that poor girl who is the subject of the thread was) better? Absolutely no, and from where do those claims of “randomness” mostly originate? A set of misguided assumptions and a poor understanding of what it takes to elevate the life or a typical high school student to a competitive applicant.

Having been here for more than a decade, and read plenty of applications “packages” I can, without much hesitation, that the overwhelming number of applicants is poorly advised and has failed to put his or her best foot forward.

On a last note, the prosaic topics presented by Calmom do NOT confirm the randomness of the subject, but indicate a deep understanding of what a compelling answer to an EASY question really is. Over the years, I have used the “slice of life” example many times. The trick, if there is one, is to use a very simple topic to convey a vivid and memorable insight into the fabric of the applicant that is not reflected in the sea of numbers and scores. But it would be a mistake to think that the simple essays are simplistic and basic. They only are on the surface!

And, when the director of admissions at “XYZ hyper-selective school” handwrites a comment on the admission letter that refers to one of those “prosaic” topics, one can reasonably concludes that … it worked. With simplicity and elegance!

QuantMech · May 28, 2013, 2:10pm

I’ve served on a lot of selection committees over time, and I have seen that a single person with a strong opinion can often sway the entire committee. I have also seen situations where one committee member is unlikely to be able to push through a string of candidates if the candidates are discussed sequentially, while each of the candidates might get through, if they were discussed at separated times. There seems also to be some “horse swapping” going on, where committee members support each others’ choices (or not). While I believe that committees devote earnest efforts to reducing the biasing effects of these practices, nevertheless, the selection is a human process.

I think that there is effectively an element of randomness as a result, even though there is an “explanation” for each of the decisions (which the applicant will in general never know). As above, that is not to say that all of the decisions are random!

rhandco · May 28, 2013, 2:20pm

The randomness of the process comes down to who they review first. A school that sets limits on, for example, students from State A, B, and C because they are neighboring states, might review applications from those states last purposefully. A school that has a lot of student athletes would not look kindly on other students who have sports as their main extracurricular, but will not play in college.

I do think that having a lot of different activities, but doing well in them, does help. Being “in the band” and “in the French club” does not help as much as “Select Band” and “President of the French Club”. And sports help a lot in general. Showing commitment to anything in the summer helps, having a summer job helps.

Yes, there is randomness, but there is no way to control that. And really, I know that at my college, people who did not have very diverse interests would be rather miserable. Even our engineers were good at liberal arts and our history majors good at science. There are many other schools where a hyperfocused person would do well but not at my alma mater.

I teach college and I also interview for my alma mater. I basically tell people that even the “sure things” get rejected. And some of the “no ways” get accepted. When a “bad” GPA is 4.0 and a bad SAT score is 2000, some of those can make their way into an Ivy. And some with a 4.6 GPA and 2300 SAT will not make it. I know that even happened back when I applied to college; my stats were like the first group and I got into both Ivies I applied to. And some friends with stats like the second group did not get in, even then when admission rates were higher at Ivies because far fewer people applied.

Hanna · May 28, 2013, 2:43pm

“One big lesson: you would be amazed how many bright and accomplished people are sending out terrible resumes and cover letters. It’s not that my d. has an amazing talent, it’s that a lot of people really, truly need serious coaching to get that simple first step right.”

Do they ever! You would not believe some of the terrible resumes and cover letters that cross my desk – even though all of my career counseling clients are practicing attorneys and graduates of top law schools. Job-seeking requires a skill set very different from the skill set required for the job. You have to get it right, or you’re dead in the water.

Hunt · May 28, 2013, 2:58pm

I don’t think it’s a crapshoot, and I don’t think it’s poker either. I think it’s golf. Think about it–in golf, you’re trying to get the ball into a tiny little hole that’s far away. You’re competing with others who are trying to do the same thing. How do you win? Well, talent and skill are huge elements, but there’s a chance element, too. A gust of wind, an invisible irregularity in the turf, lots of things. But you can generally look at a golfer’s stats and determine whether he’ll get on the green–or at least if he’ll makethe cut.

But I think there are a lot of misconceptions–some kids seem to think that if they hit the ball hard and with sincerity, they’ll win. Others seem to think you have to hit a hole in one every time to win.

MitchKreyben · May 28, 2013, 4:14pm

I just spit up a bit in my mouth after reading the last few pages of our chat. I think we can all agree that we’ve beaten this topic to death. Some want to put huge time/effort into gaming the system. Some thing it’s random. Some think it’s a process with some quirks, and sloppiness. Some think it’s opaque. We’re pretty much all stuck on our way of seeing it. And while I believe you should control what you can control, I am recommending that my offspring give it their absolute best shot after hearing best practice concepts…but I’m also harping to them that they’ll be happiest where they are wanted…not where they bamboozled some poor adcom schlub with a bunch of b.s. prose that had little to do with who they really are. I know, I know…some of you will say “i never said to to THAT!”. Fine, respond with that if you want, but I won’t be here to read it.

calmom · May 28, 2013, 5:42pm

QuantMech – I think you are confused the concept of “random” with a subjective process.

If you were running a randomized, clinical trial and you assigned volunteers to either a test group or control group after a process in which they were all interviewed and submitted essays, and then a team of doctors reviewed each file and decided which individual would go to each group – that would be considered the very opposite of random.

It may appear “random” to an outsider but that is because of lack of information.

Yes you are right that on any given day a result could be different, due to a whole array of factors – but that is no different than most things in life. Results vary because of multiple factors.

A truly random process would function as a true lottery; each eligible applicant could be assigned a number, and a computer algorithm could be created to generate a random list of accepted applicants. But the influences you cite are specifically why it is not at all random.

Compare it to a close election. You don’t know before the election which candidate will win, and no one can predict the outcome with certainty – at best, the pollsters can give a set of probabilities. Unforeseen events can influence outcome: perhaps voter turnout is lower than anticipated because of bad weather. Had the election been held on a sunny day, maybe the outcome would be different. But again – far from a “random” process.

calmom · May 28, 2013, 5:45pm

I like the golf analogy, Hunt. :)

Madaboutx · May 28, 2013, 5:58pm

I think having an “element of randomness” is a lot different than it “being a crapshoot.”

Maybe we are saying the same thing in the big picture but…

IMO, saying its a crapshoot is meant to help the rejected feel better about themselves and saying its a well organized, thoughtful process boosts the egos of the accepted.

I wish there was a simple, more elegant way of describing the process that doesn’t make the rejected feel too badly about who they are and at the same time doesn’t detract from the hard work and talent of the accepted.

QuantMech · May 28, 2013, 7:49pm

Right, I think you understand my point, Madaboutx. Clearly the process is not random, as an entire process.

I would say that a process had no elements of randomness if it were perfectly reproducible–that is, if the same set of applications went to the same admissions staff, but the aspects that appear to be irrelevant could vary (e.g., order of reading applications, whether a particular staffer had a bad cold or not on a given day, order of discussing applicants in committee, even whether an applicant was the last discussed before lunch or the first after lunch).

This experiment cannot really be run unless the admissions committee members had very weak memories. However, among the posters who argue that the process does not have any elements of randomness, do you think that the selected class would be exactly the same if the selection were re-run as I have described?

No question, there are people who do not have the picture of where their application stands in the overall pool. No question, from some of the brief descriptions of accepted applicants, it may be hard to see why they were accepted, yet if one had access to the entire file, it would become obvious.

calmom · May 28, 2013, 7:55pm

But why does a person need to deal with rejection in a competitive process by refusing to acknowledge the simple possibility that others who were not rejected had qualities that were more appealing to the people making the decisions? I mean, why the fragile egos?

No one’s ego should be hurt because Harvard turned them down. Yes, there is the sting of disappointment – but I don’t understand the need for the rant. The nature of the process is that a few get in, and many are turned away. If a school accepts 7% of applicants and the applicant isn’t the most amazing student ever to cross their threshold, then of course the odds are that the student won’t get in. But that doesn’t change the fact that there is a process in place, and that some students with equivalent stats do a better job than others at framing their applications.

LakeClouds · May 28, 2013, 9:28pm

I really don’t think you understanding what QuantMech is saying. If you had the adcom reread the entire set of applications from scratch (after using the MIB Neuralyzer Pen on them) they would come up with a different set of admitted students based on moods, ordering of reading, etc, etc. Some students would be accepted in both “experiments” (legacies, athletes, et.) and some rejected in both (unqualified applicants) but the bulk in the middle would not be the same. This is what makes it a random process.

I actually like your idea of running it more like a lottery process.

calmom · May 28, 2013, 11:50pm

Here is why it is not “random” --the “element of randomness” notwithstanding.

Let’s imagine an elite school that accepts 10% of its applicants. It will receive 10,000 applications and offer admission to 1,000 student.

There is some very small percentage of students who are essentially shoe-ins for admissions. For our hypothetical, let’s say that is 5%.

There is a much larger segment of applicants who have virtually no chance of admission. They simply are not close to meeting the minimum that the ad com is looking for. Let’s say that is 30%.

Let’s say you can program a computer to run simulations based on assigning a scored set of variables to each applicant. No matter how many simulations you run, that top 5% will always get accepted and the bottom 30% will always get rejected.

So now you are left with the 60% who are essentially discretionary admits – they may or may not be accepted. But they don’t all stand as having equal chances – because each applicant has a multivariate score, the relatively likelihood of each acceptance is influenced both by the weight of each of the variables, and by factor such as the order in which the application is considered.

What you are going to get is a probability as to each applicant being admitted. It’s going to look like Nate Silver’s electoral map – there are going to be some applicants who get accepted 90% of the time, some who get accepted 80% of the time, some 70% of the time – and others who get rejected 90% of the time, etc. Some of those applications are going to fall within the 40-60% probability range, and those are the ones most likely to seem “random”.

What I am saying is that if the applicant knows something about what variables are being considered and how they may be valued, then the applicant can do things that will likely change the probabilities attached to their likelihood of admission.

So if application A has a 70% chance of admission and application B has a 40% of admission, it is possible that B will get in and A won’t – but it is far more likely that A gets in over B. If you are an applicant who has a 10% chance of admission, and there is way that you can boost your chances to 20% – then you have effectively doubled your chances.

So why wouldn’t you want to spend some time focusing on increasing those chances?

My guess is that the margin of variability at Ivies is probably much less than all of the “crap shoot” adherents think, in part because of the way that SCEA and ED skews the odds. You are right that if the same process were run all over again, you would end up with a different set of admitted students, but it probably wouldn’t be that different – I’m guessing there would be 80-90% overlap between any two sets of outcomes in that case.

xiggi · May 29, 2013, 1:02am

There is really ZERO evidence to support this conclusion. The chances that the accepted class would essentially be identical is much higher than being mostly different. You are simply clinging to the erroneous belief that adcoms are clueless and that the decisions are entirely whimsical. The fact that some who could have been accepted were rejected does not make the process random. It actually demonstrates to be either objective or subjective, but as far as random as it could be.

Time has come to understand what random means.

QuantMech · May 29, 2013, 8:05am

I agree quite closely with calmom #538. I’d set the percentages are fairly closely as well–with about half of the admitted class being admitted under any scenario. They would be 5% of the applicants, at a school that admits 10% of the entire group of applicants. I don’t know how many students would not rate serious consideration, but 30% of the applicant pool does not sound too far off to me. At this point, we have 50% overlap between the admitted groups every time the “experiment” is run.

My guess is that, if the experiment could actually be run, the overlap between any two admitted classes would be more like 70-75%, rather than 80-90%. I am basing this on the frequent statements by admissions personnel at top schools that they could fill the class 2 or 3 times over with no diminution in quality.

It is possible that the characteristics of the 5% who are admitted under any scenario leave gaps of specific types to be filled in the class. If the process causes them to be admitted first, then their admission might constrain to some extent the qualities that are sought in the rest of the applicants, pushing the overlap up a bit. I would be very startled if the overlap were as high as 90%, though.

I agree with the other points that calmom makes.

'It's a crap shoot': Father of girl who wrote scathing letter to Ivy League colleges

CONNECT WITH US