- EphBlog - http://ephblog.com -

The Parable of the Privilege Pill

This comment from abl leads to the Parable of the Privilege Pill.

Imagine a family with twin sons, just entering 9th grade. The boys are average, both in their natural abilities and in their academic inclinations. Son 1 goes through high school with average grades and average test scores. According to Williams Admissions, he has an Academic Rating of 9. If he applies, he is rejected, as are all AR 9s. Note that Williams is not punishing him for bad performance in high school. The purpose of admissions is neither to punish nor reward. Williams rejects Son 1 because AR 9 high school students, on average, do very poorly at elite colleges.

Imagine that Son 2, on the other hand, takes a magic Privilege Pill on the first day of 9th grade, a pill which dramatically increases his academic performance for four years. He will receive excellent grades in high school and do very well on the SAT. Williams Admissions will rate him an AR 1 and, probably, admit him if he applies.

Williams would not (and should not) admit Son 2 if it knew about the Privilege Pill. By assumption, the pill only lasts for four years. After that, Son 2 becomes identical to Son 1, an AR 9, highly unlikely to perform well in an elite classroom. Admission to Williams is not a reward for strong performance in high school; it is a forecast of academic success in college.

The same reasoning applies to the Anti-Privilege Pill. Imagine a different family with twin daughters blessed with academic talent. Daughter 1 does very well in high school, is rated AR 1 by Williams and (probably) admitted. Daughter 2, unfortunately, takes an Anti-Privilege Pill at the start of high school and does much worse in terms of grades/scores than she would have done if she had not taken the pill.

Williams would (and should) admit Daughter 2 if it knew about the Anti-Privilege Pill. Recall that the pill, by definition, only lasts 4 years. Daughter 2 is, in truth, an AR 1 student whose underlying abilities have been masked in high school. We expect her to do as well at Williams as Daughter 1. Rejection from Williams is not a punishment for poor performance in high school; it is a forecast of academic struggles in college.

Things are different, however, in the case of a Privilege Pill (or Anti-Privilege Pill) which is permanent in its effects rather than temporary.

Consider a car accident in 9th grade which, tragically, leaves Daughter 2 with permanent neurological damage. Through no fault of her own, she will do only average in high school and will be scored as an AR 9 by Williams admissions. She will be rejected because, on average, high school students with AR 9, regardless of how they came to have an AR 9, do poorly at elite colleges. Even though she would have been an AR 1 (like her twin sister) were it not for the car accident, that sad fact does not influence Williams admissions.

The same reasoning applies to a Privilege Pill whose effect is permanent. If the Pill turns an average 9th grader into an AR 1, then Williams should admit her because she will, we expect, do as well as all the other AR 1s. The source of student ability — genetics, parenting, schooling, luck, wealth, special tutoring, magic pills — does not matter. Admissions to Williams is not a value judgment on the source, or justness, of student achievement in high school; it is a forecast of success in a Williams classroom.

With this framework, we can evaluate abl’s question:

If there are two students alike in every material respect (1450 SATs / 3.8 GPAs at the same school with comparable resumes), and you know that one student achieved her SAT scores after working with a private tutor with a long history of success stories while the other student did not have that opportunity — who would you accept?

The student without the tutor, obviously! In this scenario, the tutored-student has taken a Privilege Pill which, by assumption, is only temporary. She isn’t truly an AR 2. She would have scored 1300 without the tutor. She is really an AR 4 (or whatever). She is likely to do as well as other AR 4s at Williams. So, we should reject her (unless she is an AR 4 that we really want).

I honestly don’t see how any rational, clear-minded person can say that they aren’t going to accept the student who achieved her score on her own. That’s not because we are prejudiced against the student who got help: it’s that we don’t (or, at the very least, we shouldn’t) believe that her 1450 represents the same level of accomplishment and potential as the 1450 of the student who took the test cold.

Exactly how do you propose that Williams admissions determines “the student who achieved her score on her own?” While I am happy to answer your hypothetical question, the sad truth is that Williams has no (reasonable) way of determining which students achieved on their own and which did not. High quality SAT tutoring is available for free at Khan Academy, for example. How could you possibly know if a given applicant “took the test cold?” Answer: You can’t.

There strikes me as being a reasonable debate to be had about how and whether admissions officers should take these sorts of advantages into account in the admissions process. There is no reasonable debate to be had about whether or not privilege plays a role in student achievement as measured by SAT scores and by GPAs.

Perhaps. But the key question becomes: Are the advantages of privilege temporary or permanent? Does the Privilege Pill last through 4 years at Williams? If it does, then we can ignore it. Admissions to Williams is not a value judgment on the source, or justness, of student achievement in high school; it is a forecast of success in a Williams classroom.

Fortunately, this is an empirical question! Define “privilege” however you like, while using data available to Williams Admissions. I would suggest: A privileged applicant is one who attends a high quality high school (top decile?), will not need financial aid at Williams, and comes from a family in which both parents attended an elite college. (Feel free to suggest a different definition.) We can then divide all AR 1 Williams students into two groups: privileged and non-privileged. If you are correct that privileged students benefit from things like high quality SAT tutoring which makes them look temporarily better than they actually are, we would expect the privileged AR 1 students to perform worse at Williams than the non-privileged AR 1s. The same would apply to privileged versus non-privileged AR 2s, AR 3s and so on. Director of Institutional Research Courtney Wade could answer this question in an hour.

But don’t expect that analysis to be made public anytime soon. Courtney, and the people who do institutional research at Williams and places like it, are smart. They have already looked at this question. And the reason that they don’t publish the results is because of the not-very-welcome findings. Privileged AR 1s do at least as well at Williams as non-privileged AR 1s, and so on down the AR scale. The effects of the Privilege Pill are permanent. If anything, the results probably come out the other way because the AR scheme underestimates the benefit of going to a fancy high school like Andover or Stuyvesant. But let’s ignore that subtlety for now.

The last defense of the opponents of privilege is to focus on junior/senior year. Yes, the poor/URM AR 3s and 4s that Williams currently accepts don’t do as well as the AR 1s and 2s in their overall GPA. But that is precisely because of their lack of privilege, or so the argument goes. After a couple of years, Williams has helped them to catch up, has made up for their childhood difficulties and obstacles.

Alas, that hopeful story isn’t true either. AR 3s/4s do worse than AR 1s/2s even after two years of wonderful Williams.

Summary: Admissions to Williams is not a value judgment on the source, or justness, of student achievement in high school; it is a forecast of success in a Williams classroom. It does not matter why you are an AR 1: intelligent parents who value education, luck in your assignment to a charismatic 8th grade teacher, wealth used to pay for special tutoring, genetics, whatever. All that matters is that your status as an AR 1 provides an unbiased forecast of how you will do at Williams. The Parable of the Privilege Pill highlights why the source of academic ability is irrelevant.

If Williams wants better students — students who write better essays, solve more difficult math problems, complete more complex science experiments — it should admit better applicants.

Facebooktwitter
Comments Disabled (Open | Close)

Comments Disabled To "The Parable of the Privilege Pill"

#1 Comment By sigh On March 15, 2018 @ 8:29 am

Let dead horse lie. That comment is from October, 2017. You’ve written this (flawed) parable multiple times. I hope this was a mistake repost.

Williams just hired a new president. That deserves multiple posts. Not this. Ffs.

And if you say one freaking thing about me joining as an author, so help me God…

#2 Comment By David Dudley Field ’25 On March 15, 2018 @ 8:46 am

sigh:

1) We definitely need multiple posts on our new president. And you are just the person to put those together! Please join us as an author.

2) Several of my most loyal readers have mentioned how much they like abl’s contributions. So, I will continue to highlight them, along with my responses.

3) I have only written this parable twice. But, no worries! I will be publishing this piece, with improvements, each year in March for the next 40 years or so. Indeed, as you should have noticed, my own core contributions to EphBlog are evolving into a collection of 50 or so essays, each published once a year. Feedback is always welcome!

4) Speaking of which, I am honestly curious about what you think of this parable . . .

#3 Comment By sigh On March 15, 2018 @ 9:01 am

Why would I give you my opinion when you can’t even go one sentence without respecting my request not to ask me to be an author? When you claim people like abl’s comments but don’t include his comment itself nor his reply to the parable?

I call bs.

But fine: my feedback is the same it has been for over a decade: these hypotheticals are not accurate representations of how admissions to elite colleges actually works, nor are they accurate to the goal of admissions (the best class, not the best academic individual students). I’d only add that I echo zsd’s point: parables are supposed to be short.

Finally, there is a large literature on college admissions and these issues. Read more, repeat yourself less.

#4 Comment By David Dudley Field ’25 On March 15, 2018 @ 9:16 am

This version of the essay is different from last year’s precisely because of the some of the comments abl made. Thanks to abl for those comments!

Of course, I can’t address all of abl’s comments on that post without making the essay twice as long as it already is! In particular, he asks:

You take for granted that Williams’ admission choices should be based on who is most likely to be academically successful at Williams. Why?

Excellent question, but one that requires 1,000 words to answer, sadly.

You write:

there is a large literature on college admissions and these issues

Indeed! And, conveniently enough, I know that literature, both the academic and popular, better than all but a handful of Ephs. Stand by for some knowledge dropping this summer!

#5 Comment By equus On March 15, 2018 @ 12:14 pm

Here is a parable, apropos of nothing specifically. Alas it is also long.

For as long as anyone can remember, the most famous racehorses have all gone to the Surrey Horse Academy. SHA graduates dominate races everywhere. Few horses are faster. Horses from pastures all over the world dream of sending their foals to the SHA. Nothing is more satisfying than seeing your foal enjoy a decorated career as a champion racer.

Foals at the Surrey Horse Academy train very hard. They work out twice or even three times a day. It’s not an easy life but it’s a great experience for young, fast horses. They have the best coaches. They eat the best grass. They have the most learned veterinarians. Many older horses, long put out to pasture, remember their time at the SHA as the best years of their life.

Ungulates the world over accept that the SHA graduates the fastest horses because the SHA *makes* the fastest horses. All of that wise coaching, all of that hard training, all of that excellent grass — this is what goes into racing fast, as a horse. Nobody is really sure how it happens, though. At base the principles of horse training are the same everywhere. Many lesser horse academies have similar programs. Nor does the SHA have a monopoly on champions. Even lesser academies count some winners among their graduates, though not nearly as many as the SHA.

The Surrey Horse Academy will not train just any hopeful foal. Spots are extremely limited. Among other things, foals must post amazing race times to have even the slightest prayer of getting in. The official line is that the SHA wants to select those foals who are likely to do the most with the amazing training it provides.

The curmudgeons of the horse world, the mules and the donkeys and the odd zebra, have their own theory of what is going on. They think SHA graduates win races primarily because the SHA selects the fastest foals. Such is the social prestige of the SHA and other august horse academies that, collectively, they have a lock on each year’s most promising racers. This leaves only the slower foals for lesser academies. While it may seem that the SHA makes fast racehorses, in fact the SHA merely selects fast racehorses. In almost every case, SHA graduates would’ve been extremely fast whatever horse academy they went to. Conversely, had the SHA admitted slower foals, those foals would’ve remained slow despite all that excellent training and all that amazing grass.

Nobody likes this curmudgeonly hypothesis very much. As fervently as Aquinas believed in the Trinity, ungulates hold that the SHA makes fast racehorses. SHA coaches preach this orthodoxy and they certainly believe it. Why wouldn’t they? Who wouldn’t want to think that they are part of a magical academy that moulds raw unformed foals into winners? SHA coaches think so much of themselves that they come to believe that they could do more good for horses everywhere. They believe it is within their power to improve the lot of the entire species. They decide to admit foals from underrepresented horse populations. In doing so they believe they will reverse historical injustice by gradually turning these horses, too, into amazing racers.

For at least a generation, the SHA strives to make space for underrepresented horses among its graduates. In that time, a mysterious thing happens. Discussing the nuts and bolts of this initiative — alluding to the outcome of this natural experiment in social and athletic engineering — becomes oddly taboo. SHA coaches and SHA administrators give media interviews in which they obfuscate about the true nature of their selection process and its meaning. Admissions officers hold their cards extremely close to their nostrils.

Why should this be so?

Should the SHA and its peer institutions not be trumpeting the success of their program from the barn tops?

The donkeys and the mules and the zebras suppose that the SHA isn’t succeeding in making racehorses out of those slower foals. Instead, the initiative threatens at every moment to illuminate the true nature of the SHA — that its success is a function of selecting those foals who are already the fastest, and then associating itself with their inevitable speed. Nothing shows this so clearly as the long-running experiment in admitting foals from underrepresented horse populations and hoping that the magic of the SHA will make them faster. Statistics show that these slower foals aren’t any faster, even upon graduation, than comparable foals at lesser academies. The SHA may have the best grass, but that has nothing to do with running fast. At the same time, those faster foals who might’ve been admitted according to strict race-time criteria, but under the fuzzier egalitarian regime have been denied a spot, haven’t really been harmed. They’re doing just as well, winning just as many races, as we would’ve expected anyway. It turns out foal racing times predict adult racing times regardless of what horse academy the foals went to.

Meanwhile, in the stables of the Surrey Horse Academy, cognitive dissonance reigns supreme. Most deny that racing ability has any innate component at all. The new orthodoxy is that any horse can be fast. Everybody says remarkably untrue things all the time. They say that the race times of young foals don’t predict anything. They say that the majority horse culture somehow suppresses the speed of underrepresented horses. Coaches, who have spent their lives among fast horses and have trained hard to be fast themselves, wonder what it means to be fast in the first place. Coaches who pride themselves on their quarter mile ask why running slower should be so bad. They point out that races are a social construct. They insist that the SHA admits only the very fastest horses and that no standards have ever been compromised. They take pride in their choice to admit some foals for attributes other than speed. They say that this choice has had no effect on speed at the academy. They say that you don’t want a class full of fast horses anyway (even though they also assert that our classes are full of fast horses regardless). They say that nobody wants a speed grind.

(Meanwhile, in much the rest of the world, horse academies do admit only the fastest horses and they have somehow managed to avoid social apocalypse.)

There are horses who dream of excluding race times as an admissions criterion altogether. The donkeys and the mules and zebras hope that the ideologues will actually go that far. Too many slow foals among the graduates, for too long a period, would give up the game and prove their position–that selection, not training, is the most powerful force and the secret ingredient at the Surrey Horse Academy. For now horse academies everywhere protect themselves by acting in concert, like a cartel. They all deemphasize race times in roughly the same way, in favor of underrepresented horse populations, and this makes it easier to elide the truth about the power of selection.* The mythology, though, is unstable, because it is not true. Sooner or later some academy somewhere might get back to selecting foals more rigorously based on race times, and unless the SHA follows suit, its graduates will no longer be among the fastest.

*For the very same reason, the SHA and other elite academies also form varsity math leagues. These enforce strict league limits on how many favors can be shown to slower foals with better math ability force everyone to operate at the same selection (dis)advantage. The alternative is an arms race among horse academies where varsity math is a valued tradition, which would divert the force of SHA’s selection, trading prestige among peer institutions for prestige in the wider world horse racing.

#6 Comment By sigh On March 15, 2018 @ 3:25 pm

FFS, ephs are terrible at parables.

Here’s an attempt:

A restaurant cooks a great steak–it’s got 4 stars. It’s famous for it’s steaks. It has a recipe for steak that is very simple–steak, butter, salt, pepper, a sprig of rosemary. The recipe is published yet others cannot make the same delicious steak. The key to its steak? They buy from only one pasture they know raises a particularly high quality cattle and then house-ages the steak for 45 days. Anyone could do it, but they got the in with that pasture.

One day, the restaurant decides it should be more environmentally friendly to get new customers, so it introduces grass-fed steaks to the restaurant from a different pasture that has high quality cattle, but not the same breed. The grass-fed steaks make the restaurant friendly to more customers who care about sustainable meats, but some of the regulars grumble about the grass-fed beef being chewier compared to corn-finished. The restaurant continues to get 4-stars, continues to cook the same way it always has, but it now has another option for people who like grass-fed beef and the steaks are delicious but more varied. The first pasture now sells some of its steaks to other restaurants to cook, but it’s pissed off that it no longer only has one customer, because now it experiences a small bit of struggle when it previously had it relatively easy in the low-margin business of cattle. Some of the regulars are pissed off they now have a harder time getting reservations. The regulars spend some of their free time writing parables to explain why they deserve those reservations more than others.

#7 Comment By ZSD On March 15, 2018 @ 3:37 pm

https://literarydevices.net/parable/

The form is characterized as short or concise, probably because a parable was meant to be told.

Through the bottom of comment six, my count is 1500 words.

I am sure that whatever points are trying to be made, they can be helped by good editing including the more frequent use of para breaks.

#8 Comment By equus On March 15, 2018 @ 3:53 pm

I don’t know about paragraph breaks, but your link calls Andersen’s ‘Emperor’s New Clothes’ a parable, and that also weighs in at around 1500 words.

#9 Comment By ZSD On March 15, 2018 @ 4:03 pm

… but Anderson could write

#10 Comment By David Dudley Field ’25 On March 15, 2018 @ 4:11 pm

ZSD: If your claim is that Anderson writes better than I do, you win! He does.

More useful would be specific feedback. Which word choices are awkward, which sentences redundant, which thought experiments unclear?

You claim it is too long. Which paragraphs would you cut?

#11 Comment By dcat On March 15, 2018 @ 5:30 pm

Not sure if this is the place for it or not, but “The American Scholar” this month has a fantastic essay on privilege that really does take a strongly critical view of the concept, at least inasmuch as while it does not reject that there is privilege (obvs) but that it’s current manifestation, especially in the academy, is more of an accusation that grants unearned sanctimony rather than poses an actual argument that is willing to engage others.

#12 Comment By dcat On March 15, 2018 @ 5:53 pm

So, I subscribe and read it in hard copy, but here is a link to Robert Boyers’ article:

“The Privilege Predicament”

#13 Comment By abl On March 15, 2018 @ 7:39 pm

DDF –

The framework of your post is at points strong. Many of the assumption that you make in it, however, are not. For example:

The sad truth is that Williams has no (reasonable) way of determining which students achieved on their own and which did not. High quality SAT tutoring is available for free at Khan Academy, for example. How could you possibly know if a given applicant “took the test cold?

The Khan Academy videos do not constitute relatively high quality SAT tutoring. Are they better, for most students, than nothing? Sure. I’m not even sure they’re all that much worse than the standard SAT tutoring options used widely by middle class consumers — Kaplan and such. But they are materially worse than the sort of boutique extra-tailored tutoring service preferred by the wealthy — especially at achieving the sort of outcomes necessary for schools like Williams. Khan Academy classes are probably similarly likely to Kaplan at raising scores from 1000 – 1200. They are far less likely, however, than the best of the boutique services to raise a student’s score from 1200-1400. (And for purposes of Williams admission, it’s the latter improvement that’s relevant.) They also require far more self-motivation. The average prospective Williams applicant is far more likely to complete a course of 1:1 fancy SAT tutoring than she is to complete the equivalent amount of Khan Academy prep.

Also, Williams could easily determine (or begin to determine) what students have studied for the SAT and how: Williams could ask that question on the application. Something as simple as “did you study for the SAT; if so, what service did you use (and at what cost)” could potentially do the trick. For a more imprecise measure, Williams could just look to family income. Common sense indicates (and my own experience in this industry supports) the notion that the average wealthy applicant pays substantially more for SAT prep services than the average impoverished applicant. This obviously is far from a perfect measure of anything. But admission to top schools is all about applying rough general assumptions via weak and imperfect signals — and this should be one such signal taken into consideration.

Rejection from Williams is not a punishment for poor performance in high school; it is a forecast of academic struggles in college.

At best, this is an oversimplification (because intra- and extra-curricular contribution — in addition to pure academic performance — are and should be part of the admission consideration). At worst, this is simply incorrect. As a normative matter, you haven’t resolved when measurable academic aptitude becomes relevant. Is it on arrival at Williams? E.g., to the limited extent that pure academic aptitude matters, should Williams accept the class most likely to perform highest in the first week of its first year? Or is it on graduation? E.g., should Williams accept the class that, in the aggregate, appears most likely to be the highest performing at the point of graduation? These are each easily defensible and reasonable approaches. Neither could necessarily be tested via GPA (the former looks for the best raw talent–which stops being relevant once Williams starts–and the latter looks to performance at graduation–for which even final-semester GPAs assess too early). And again, I want to emphasize: pure academic aptitude is only one of several factors that does and should motivate admissions decisions.

They have already looked at this question. And the reason that they don’t publish the results is because of the not-very-welcome findings.

This is probably the weakest part of your piece. You make confident assumptions on a topic about which you know very little. There are many reasons why Williams would study (or wouldn’t study) these sorts of outcomes. And there are many reasons why Williams would release (or wouldn’t release) the results of said studies. The fact that you don’t have access to the results of any possible studies along these lines doesn’t mean that Williams (a) has conducted these studies; and (b) the studies support your point. If you’re planning on reposting this for years to come, please throw out this poor assumption (and rewrite accordingly).

— Finally, I want to repeat, again, that this privilege/aptitude argument is not only grounds on which affirmative action practices are justified. Even if the effects of the “privilege pill” (or some versions of the privilege pill) last for 4 or 10 or 40 years, that would only obviate one of many of the reasons to admit underprivileged applicants. —

#14 Comment By David Dudley Field ’25 On March 15, 2018 @ 8:52 pm

Khan Academy classes are probably similarly likely to Kaplan at raising scores from 1000 – 1200. They are far less likely, however, than the best of the boutique services to raise a student’s score from 1200-1400.

This is delusional. There is no evidence that any sort of tutoring has anywhere near that magnitude of an effect. See my post from 8 years ago. See this from last week.

Myth: Test Prep and Coaching Produce Large Score Gains

If tests were easily coached and coaching was only available to the wealthy, there would be an equity problem, even if tests are generally useful. Commercial test prep is clearly expensive, so this is a critical issue.

Researchers have conducted a mix of experimental studies and controlled field studies to test this question. They have generally concluded that the gains due to test prep are more on the order of 5 to 20 points and not the 100 to 200 points claimed by some test prep companies.

One review found a typical gain of 15 to 20 points on the math portion of the SAT and 8 to 10 points on the verbal portion. One of us conducted a more in-depth analysis of 4,248 high-school students and, after controlling for prior scores and the differing propensity of students to seek coaching, we estimated a gain of 14 points on the math test and 4 points on the verbal.

These are just averages, and among students who prep, a small percentage do realize 100 point gains. Why? The research suggests that they fall into two overlapping groups. The first consists of students who are fundamentally well prepared but are rusty on some basic concepts. The second group has not put even basic effort into understanding the questions and the flow of the tests. Gaining simple familiarity is one of the surest ways to achieve quick increases in scores.

There no evidence that it is common, much less prevalent, that a rich family can purchase “boutique services to raise a student’s score from 1200-1400.”

#15 Comment By David Dudley Field ’25 On March 15, 2018 @ 8:57 pm

I want to repeat, again, that this privilege/aptitude argument is not only grounds on which affirmative action practices are justified.

I agree! I never meant to suggest otherwise.

This is probably the weakest part of your piece. You make confident assumptions on a topic about which you know very little.

There are no more than a handful of Ephs who know more about the academic literature about student performance in college.

sigh: Please correct me if I am wrong. Isn’t it true that students at places like Williams who score 1500 on the SAT do much better than students who score 1300, not just freshmen year, but also senior year?

#16 Comment By sigh On March 16, 2018 @ 8:39 am

If you know the literature so we’ll, why do you need to ask me? State how much SAT predicts “do better” (defined as?) and give your citation.

#17 Comment By David Dudley Field ’25 On March 16, 2018 @ 9:32 am

Jen Doleac ’03! See here:

Each 100 point increase on the SAT correlates with an additional .18-.22 points on the freshman GPA, or .13-.19 points on the cumulative GPA for all 4 years (this is consistent with other studies, which show SATs best predict freshman academic performance, but become less relevant over time). An improvement of 1 point on the AR, similarly, correlates with an improvement of .16 points on the freshman GPA, and .14 points on the cumulative GPA.

But EphBlog readers are always eager to learn more. What citations would you recommend?

#18 Comment By sigh On March 16, 2018 @ 9:56 am

I like her point on page 27 that SAT is overemphasized in the calculation of AR.

Those models are somewhat barebones on page 23 (not her fault, you model the data you have), fwiw. NLSF data through sophomore year show a smaller effect on gpa. But even with her larger estimates, a 200 point Gap is the difference between a B/B+ student and an A- student. And the sat score “explains” bout 25 percent, give or take, of gpa variation. Woohoo.

#19 Comment By abl On March 16, 2018 @ 1:23 pm

There is no evidence that any sort of tutoring has anywhere near that magnitude of an effect.

Yes there is. And you’re mixing up two points. Nobody argues that most people will improve 200 points or that the average improvement will be 200 points (or that 200 points of improvement is anything other than an exceptional outcome). The question is whether these programs achieve a meaningful improvement (and as little as 50 points is often meaningful) regularly enough to color the applicant pools of top schools. The answer to that is definitely yes. Nothing you posted rebuts that.

Put another way, elite prep programs can fail to be a sufficient condition for elite college admissions (because most people who enroll in them will fail to realize the sorts of gains necessary to obtain an offer of admission) and yet be a necessary condition for elite college admissions (because obtaining an offer without being in the minority of students who do receive a boost is prohibitively difficult). I’m not arguing for anything so significant — the reality is more complicated and the impact of these programs is more of a sliding scale than anything. But the necessary/sufficient example explains why your rebuttal misses the point.

I would agree that the average outcome of programs like Kaplan or Khan Academy is middling and that even the elite boutique programs (which I do not believe the WSJ article covers) fail to miss their advertised marks more often than not. But these programs–especially the elite boutique services–are enough successful with enough regularity to significantly impact top college admission. Here’s another example: it may be true that 19/20 of Kaplan enrollees do not significantly improve and that 4/5 elite boutique enrollees do not significantly improve — and yet for the pool of students admitted to Williams to be comprised disproportionately, or even entirely, of the 1/5 elite boutique test preppers and the 1/20 Kaplan test preppers who did see success.

This is probably the weakest part of your piece. You make confident assumptions on a topic about which you know very little.

There are no more than a handful of Ephs who know more about the academic literature about student performance in college.

sigh: Please correct me if I am wrong. Isn’t it true that students at places like Williams who score 1500 on the SAT do much better than students who score 1300, not just freshmen year, but also senior year?

What argument are you trying to make? I was pretty clear about what specific confident assumption tainted your piece. This is an entirely separate point. I have never argued that there is not some weak correlation between SAT scores and uGPA.