United Nations ReliefWeb

[Nat Wilcox]

...and some of you who are involved in trying to get cooperation from NGOs may occasionally, and increasingly, find yourselves entangled with the competing demands of experimental social scientists.

We have seen this growing in the last five years or so, under the rubric "field experiments." Tried to hire a couple of these young bucks in the last couple of years as it is an increasingly hot area in Econ. Particularly in Africa, but also elsewhere. Typically, an established academic big shot at Harvard and some grad students plan some experimental intervention on the back of things NGOs are already doing. For instance (real example), some NGO might have an information program designed to inform teen girls in rural Kenya about HIV risk. So the experimentalists get them to randomly piggyback some kind of extra information on exactly half of the high school curricula they have. Then down the road, you look for the treatment effect of the added information on outcomes you do and do not want.

The thing is, uncontrolled and unplanned variation during a study period is anathema to experimental design. And this kind of work is booming right now--I mean growing exponentially. It is very hot. So I wouldn't be surprised if, in future, some of you find an NGO contact saying: "No, sorry, can't do that. We're currently in the middle of a controlled study for Herr Doktor Professor SuperBigshot..."