Central Perk: Seth's Blog : The truth about admissions

joi, 29 ianuarie 2015

Seth's Blog : The truth about admissions

The truth about admissions

One in five applicants to Harvard and Stanford are completely qualified to attend—perhaps 20% of those that send in their applications have the smarts, guts and work ethic to thrive at these schools and to become respected alumni.

These schools further filter this 20% by admitting only 5% of their applicants, or about one in four of those qualified. And they spend a huge amount of time sorting and ranking and evaluating to get to the final list.

They do this even though there is zero correlation between the students they like the most and any measurable outcomes. The person they let in off the waiting list is just as likely to be a superstar in life as they one they chose first.

Worth saying again: In admissions, just as in casting or most other forced selection processes, once you get past the selection of people who are good enough, there are few selectors who have a track record of super-sorting successfully. False metrics combined with plenty of posturing leading to lots of drama.

It's all a hoax. A fable we're eager to believe, both as the pickers and the picked (and the rejected).

What would happen if we spent more time on carefully assembling the pool of 'good enough' and then randomly picking the 5%? And of course, putting in the time to make sure that the assortment of people works well together...

[For football fans: Tom Brady and Russell Wilson (late picks who win big games) are as likely outcomes as Peyton Manning (super-selected). Super Bowl quarterbacks, as high-revenue a selection choice as one can make, come as often in late rounds as they do in the first one.]

[For baseball fans: As we saw in Moneyball, the traditional scouting process was essentially random, and replacing it by actually correlated signals changed everything.]

What would happen if rejection letters said, "you were good enough, totally good enough to be part of this class, but we randomly chose 25% of the good enough, and alas, you didn't get lucky"? Because, in fact, that's what's actually happening.

What would happen if casting directors and football scouts didn't agonize about their final choice, but instead spent all that time and effort widening the pool to get the right group to randomly choose from instead? (And in fact, the most talented casting directors are in the business of casting wide nets and signing up the good ones, not in agonizing over false differences appearing real--perhaps that's where the word 'casting' comes from).

It's difficult for the picked, for the pickers and for the institutions to admit, but if you don't have proof that picking actually works, then let's announce the randomness and spend our time (and self-esteem) on something worthwhile instead.