Thursday, 30 August 2012

Coop-eration: chickens, baseball and the Research Excellence Framework

Back in the day, baseball teams would pay huge amounts for superstars with excellent performance statistics, such as batting averages. (Baseball involves batting, right? Bear with me, this will become a metaphor.) One now famous baseball coach started out as an econometrician. He analysed team performances, looking for players who were on successful teams even though their individual statistics were not that great. The logic was that teams, not individuals, win matches. Buying these relatively cheap players, he took them to the top of the Superbowl (or whatever - I think this book has the details).

Chicken farmers have discovered the same thing. If you only breed from the individuals who lay the most eggs within a coop, you end up selecting for highly aggressive hens who destroy each others’ eggs. It’s better to breed from individuals within highly productive coops. (Source.)

The UK's Research Excellence Framework is approaching. It ranks academic departments by aggregating each individual's productivity. This is fair enough. The problem starts when departments try to increase their ranking by trying to hire academic "superstars". If this were what made a great department, we could all just work from home. Departments should be places where ideas cross-fertilize. The person who contributes to a vibrant working environment, or frames someone else’s idea in an important new perspective, may be as important as the person with a Stakhanovite publication output.

So, in terms of incentive structures, academics are some way behind sports coaches and chicken farmers. I hope they will soon catch up with these more forward-thinking elements of society.