Tuesday, May 31, 2011

Making Failure Safe

I’ve been chewing on Adapt, Tim Harford’s latest. (He’s the Undercover Economist.) It’s a wide-ranging survey, but the basic idea is that adaptability requires making failure safe. If failure is not an option, then innovation will not occur, and eventually history will render your perfection irrelevant.

He gives the example of photography. Early digital photography couldn’t compete with the quality of film, but it didn’t need to; it justified its own existence with other virtues, like the ease of sending as email. That new virtue gave it enough market strength and time to improve its weaknesses; at this point, film is of mostly historical interest.

In some ways, Harford’s argument seems tailor-made for academics. Academic freedom should be the epitome of “making failure safe.” In a perfect world, colleges and universities would be hotbeds of experimentation, with success speaking for itself.

But it isn’t. In fact, it’s hard not to notice that academe is more tradition-bound than most other institutions. That’s especially true at the community college level.

Some of that has to do with the vulnerability of the student population. Given students whose previous track records suggest real challenges, there’s an argument to be made for playing it safe. For these students, failure is patently unsafe; the dropout rates for students who fail a class are dramatically higher than those for students who pass. These students generally don’t have elaborate safety nets; if anything, they’re often stretching just by enrolling. The scion of wealth and privilege at Snooty U can afford to take a flyer on a risky class; the single Mom who’s barely making it, can’t.

Some has to do with transfer. Annoyingly, many destination colleges use the “checklist” model for determining transfer credits. “Intro to Psychology” is on every checklist, and it transfers without issue. “Topics in Psychology” -- the kind of course in which we’re allowed to take risks -- generally doesn’t transfer at all. If it does, it gets “free elective” status, which amounts to not transferring. Since the classes have to fit predefined slots or they won’t carry over, no matter how good they are, there’s a powerful incentive to keep it vanilla. Colleges that run interdisciplinary freshman seminars for their own students won’t take interdisciplinary seminars in transfer.

In practice, experimentation becomes a class privilege. Those who have the resources to survive failure are allowed to experiment; those who don’t, aren’t.

The tragedy of that is that the traditional system works most poorly for the students with the fewest resources. In a very direct way, they have the least to lose by trying something different. They’re the ones who most need the breakthroughs, but the least likely to get them.

Harford’s examples unwittingly make this point. He repeatedly cites Google as exemplary in trying hundreds of things and sticking with the handful of riotous successes. But Google is an insanely profitable company; it can afford to eat a bunch of small losses. In the public sector, turning operating profits like that would be considered politically unacceptable, even if someone figured out how to do it. And without a cushion like that, repeated failures are fatal.

That said, I’d love to figure out ways to make pedagogical failure safe. The alternative is to keep doing what we’re doing, which implies that we’ll keep getting the results we’re getting.

Wise and worldly readers, have you seen or figured out a way to make failure safe in a setting without a huge resource cushion?