Over the past several decades, acceptance rates at the most selective United States colleges and universities have dropped dramatically. In the mid-1990s, for example, Yale University had an acceptance rate of around 18% for freshman applicants, whereas its freshman acceptance rate in 2017 was only one-third as high. Assuming that acceptances rates for the high school class of 2018 are similar to those for the class of 2017, all freshman applicants to Yale during the 2017-2018 admissions cycle will compete in a pool from which approximately 6% of freshman applicants are accepted.

Which of the following would most weaken the conclusion of this passage?

(A) Applicants who apply to Yale through Single Choice Early Action are accepted at far lower rates than they were in the mid-1990s.
(B) There is a significant difference in the acceptance rates of Single Choice Early Action and Regular Decision Yale applicants.
(C) The most competitive applicants to Yale often gain admission to multiple Ivy League schools.
(D) A smaller percentage of students apply to Yale through Single Choice Early Action than apply Regular Decision.
(E) The demographic makeup of Yale’s freshman class has changed significantly over the past several decades.

Before I go into the explanation, I’m going to take a little detour into a recent Washington Post Education article by Jeffrey Selingo. Most of the article is devoted to rehashing how grades are scores are not longer sufficient to get into a top schools, how the most selective schools have become even more selective, yadda yadda yadda. About halfway through the article, though, Selingo starts to dig into the actual significance of the jaw-droppingly low admissions statistics that the most selective colleges seem to post each year, and things start to get interesting.

As Selingo points out, the admissions figures that schools release are based on both early and regular decision which typically have very different acceptance rates, with rates being much higher in the earlier round (or, in some cases, multiple rounds):

[C]alculating the acceptance rate is not as simple as dividing the number of students who applied by those who were accepted. It’s a number that schools have easily manipulated in two ways.

First, colleges have boosted the denominator in that calculation by purchasing names of test takers and employing business-like marketing techniques to encourage applications from students they have no intention of ever accepting.

Second, some elite schools have created multiple application cycles with binding commitments known as “early decision,” which in some cases fills half their classes. But when colleges release their acceptance rates, the number they announce blends all the various cycles together, including regular admissions…

Take Vanderbilt University as an example. It fills 54 percent of its incoming class through two rounds of early decision. To most students and parents that might sound like they still have plenty of spots to parcel out in the regular round. But remember, the vast majority of applications that colleges receive — and then later boast about — come through regular decision. In Vanderbilt’s case, the regular round yields only around 500 students from more than 25,000 applications.

(Note that this last sentence is potentially misleading as well: “yields” and “accepts” are two different things. in 2017 Vanderbilt accepted 2,382 students out of 27,841 applicants, for an admit rate of 8.6% — not the 2% implied in the article. Still, that’s a very low rate.)

So that question I posed earlier? The answer is (B): the assertion that all students are competing in a pool from which approximately 6% of all applicants will be accepted ignores the fact that Single Choice Early Action (SCEA) and Regular Decision applicants have two very different acceptance rates.

Although the overall acceptance rate is around 6% (6.3% to be exact), the “trick” is that there are actually two pools: one with an extremely competitive but not insane acceptance rate (15.5%), and one in which very few applicants have a realistic chance of getting in (4.6%).

At Penn, which is notorious for accepting a large percentage of its class early, the difference between the Early Action acceptance rate (25.2%) and the Regular Decision acceptance rate (7.3%) is nearly 18 points. And at Harvard, the RD acceptance rate is only 3.5%, as compared to a perfectly humane 21.1% for SCEA applicants.

When “hooked” status is factored in, the playing field becomes even more competitive. Although some of the strongest applicants do of course apply ED, many hooked students wait to apply RD precisely because they can afford to do so. Penn is clear that legacy applicants lose their advantage if they don’t apply ED, but a student whose parents attended Harvard is given the same advantage RD as ED.

So basically, what’s going on is that ED is the round in which more or less normal but still very high achieving applicants have a not implausible shot at getting accepted, whereas RD is the round in which most of the acceptances go to kids who are either hooked or off-the-charts exceptional in some way. For nice, typical 1500-scoring class president/yearbook editor/captain-of-the-baseball team applicants from average suburban public high schools, the RD acceptance rate at the top Ivies is probably more like 1-2%.

Of course there are schools like MIT, which makes a deliberate effort to keep Early and RD acceptance rates more or less equivalent, but among the uber-elite, they are the exception.

Then there’s the actual percentage of the class that’s filled early. Penn, which has binding ED, accepts 53% of its class early. One major reason for this is to protect yield –the lower the RD yield (that is, the more students a school is likely to lose to a more prestigious competitor), the larger the percent of its class a school will be tempted to lock in by December. The reality is that most Penn/Harvard cross-admits are going to pick Harvard, and Penn knows it will lose a good number of its RD admits in the spring. Accepting more students early is a way to keep yield rate artificially high, which in turn helps it preserve a high USNWR ranking.

Conversely, that is why Harvard and Yale can afford to have non-binding SCEA. The majority of admits to those school will ultimately enroll, regardless of where they are ultimately admitted, so protecting yield rates is less of a concern. If a small percentage of students admitted to Harvard in December ultimately decide they’d be happier at Stanford, Harvard can be generous and give them that option.

For students who have been groomed for the Ivy League more or less since birth and aren’t applying for financial aid, all this doesn’t pose too much of a problem. And incidentally, it poses less of a problem for the very poorest (first-generation) applicants than one might think: top colleges actively seek such students, and most can afford to meet their financial full need if they manage to get in.

The problem is for applicants in the middle: the ones who might get enough money to attend a top school if accepted, but who really need to compare financial aid packages and maybe see whether they can pick up a full-ride merit scholarship somewhere. Unfortunately, as things currently stand, there’s no easy solution. There are too many players, too many variables, and too much riding on the outcomes for things to change anytime soon.

The bottom line, though, is that the next time you pick up a college guidebook and check the acceptance rate, you shouldn’t automatically think, “Oh, school x accepts 27% of its applicants. That’s not too bad — I’ll probably be fine.” Go online and find out how many applicants are accepted early, how many spots are left over for RD applicants, and how many apply in each round. Unless it’s your absolute #1 choice and you’re planning to apply early, those are the numbers you really need to go by.