Arnold Kling  

Harvard's Decision on Early Decision

Consumption Mobility and Inequ... Elephants Clashing...

Greg Mankiw comments on Harvard's decision to end early decision in the admissions process.

The early admission process has been becoming increasingly strategic on the part of both schools and students, and this game playing does not seem to serve much social purpose... If a systemic change is going to occur, this is the place.

I think that the campaign against early decision is a classic case of a superficial reform that is likely to have unintended consequences.

Early decision is most useful at schools that are classic "safety schools" for students who would prefer to go to the Harvards of the world. Places like Rochester and Brandeis are flooded with applicants who really would prefer to go elsewhere. In April, they have no idea who is going to respond favorably to their acceptance offers. Early decision helps them to manage their admissions process without a huge waiting list.

If the anti-early-decision crusade succeeds in persuading most colleges to drop early decision, I expect that soon afterward we will be reading stories about the agony for high schools students of the waiting list process.

Comments and Sharing

COMMENTS (11 to date)
dearieme writes:

The way to avoid endless difficulty with admissions is to throw out all the freshmen who don't reach some acceptably high standard of performance.

richard writes:

I think you have early decision backwards. One applies early decision to their reach school not the safety school. When one applies early decision, one promises to rescind applications to other schools in the event of acceptance. Thus early decision is binding. That said, I think it is a net positive for both school and student.

GMUSL 3L writes:


I think you're missing his point. One man's reach is another man's safety .... so if a school that is one's reach is another's safety, by applying early decision to that school, that school gets somebody with a 99% chance of attending.

GMUSL 3L writes:

Additionally, think of the harm that this will do to students in the form of reduced certainty, if this catches on elsewhere.

Instead of finding out in december, the earliest students will find out is in march/april. This will lead to a time-delayed trickle-down effect as spots clear up at each successive tier of colleges as their best admitted students get in to the spots vacated at the next higher tier. Given admissions, decisionmaking, and the communications and processing time, a non-trivial number of students -- probably more than an order of magnitude more than at present -- will only find out where they are going in July, if not August.

I fail to see how ANYBODY is advantaged by finding something like that out at the very last minute.

Piso Mojado writes:

I agree that there will be unintended consequences that will result from Harvard's decision to eliminate early decision. One advantage, though, is that it will be easier for parents of high school seniors to convince them that their senior year does, in fact, matter and that they need to continue to study even as they await the results of their college search.

nwhitehe writes:

College admissions is a classic example of the "marriage problem". In the original problem, there are n male suitors and n available females. The males all flock to their first choice of female, then the female can either reject any male or keep his hopes up. Any rejected male goes to the next choice on his list. Women behave strategically, only rejecting a man when they know they have a certain better alternative ready. When a stable equilibrium of one man for every women is reached, the women marry their man.

The reason the problem is interesting is that it originally seems like the females have more "choosing" power, but actually the males have it better off. The marriage algorithm yields the best stable match to the male preferences, and the worst stable match for the female preferences.

For the college admission process, the colleges are effectively the males and the applicants are the females. The reason is somewhat convoluted, but it is mainly that applicants don't get rejected after each application. Applicants have to apply to all colleges at one time, then the colleges can accept and reject students on their own time.

Early decision applications I think give more bargaining power to applicants. It would be even better if every school had several early decision dates. That way students could apply early decision to their top school, then if they get rejected apply for 2nd early decision at their 2nd school and so on. I think this would make applicants the males in the marriage problem and thus get more advantageous matches.

Alcibiades writes:

Agreed mwhitehe,
The more early decision(s) the better. As it is now, everyone applies everywhere. That means more apps to send out for students, and more to read for colleges. Early decision (or multiple early decisions) save a lot of wasted effort on both ends.

John Thacker writes:

"Thus early decision is binding."

Actually, Harvard has (or had) what is called "early action" instead of early decision. Early action is non-binding on the student, unlike early decision, which is binding on the student.

Most universities have early decision, but several of the most elite schools have early action instead. (Presumably the lower ranking schools are more concerned about getting commitments from students who might otherwise choose Harvard in the spring.)

nwhitehe writes:

John Thacker-
You're right, I didn't notice that distinction the first time. In the context of the marriage problem, early action strictly helps students because it involves the school committing sooner without requiring the student to commit. So for Harvard the motivation for eliminating it is obvious: they end up with the students they want more often with the program eliminated.

Robert Tsai writes:

The National Residency Match Program already provides us with a system to study what happens in the absence of early decision and early action. It matches up medical school graduates with residency programs. Basically, all candidates interview wherever they can interview. Then all candidates submit a ranked list of their desired residency programs to the NMRP, and all residency programs submit a ranked list of their desired candidates to the NMRP. The NMRP runs a computer algorithm to ostensibly come up with a mutually best-fit allocation of residents to residency programs.

One problem with a direct comparison is that residents are older than college students and usually have many more factors to influence their choice of where to go (spouses, family, etc.).

Robert Tsai writes:

(And yes, each year on "Match Day", the blogosphere and the press are full of articles on the agony of residents waiting to hear about where they will spend the next 3 to 7 years of their life.)

Comments for this entry have been closed
Return to top