Syndicate content

We shall see how many funerals...

Let us take out of the Hospitals, out of the Camps, or from elsewhere, 200, or 500 poor People, that have Fevers, Pleurisies, etc... Let us divide them in Halfes, let us cast lots, that one half of them may fall to my share and the other to yours; I will cure them without bloodletting and sensible evacuation; but do you do as ye know (for neither do I tye you up to the boasting, or of Phlebotomy, or the abstinence from a solutive Medicine) we shall see how many Funerals both of us shall have: But let the reward of the contention or wager, be 300 Florens, deposited on both sides: Here your business is decided.

That was Jean Baptiste van Helmont in the 17th century. It took three hundred years for randomized trials to become widespread in the medical profession. Now the MIT Poverty Action Lab, among others, is advocating their use in evaluating the effectiveness of development projects (and other policy interventions). Since many projects are rolled out gradually, rolling them out with some randomization generates very good data without much extra effort required.


Submitted by Peter on
While the idea of having hard evidence for development policies and interventions is attractive to many people, I think there are complications with applying the idea of randomized trials to development policies. The example given on the MIT pages is about using the fact that a school-building program would require some schools to be built before others to randomly assign children or communities to having or not having a school. First, the ethical issues here, as with medical clinical trials, are immense, and magnified by having a community element rather than being individual trials. Secondly, a key element of success of development policies is cultural, and this is likely to differ from one community to another. Allowing for this in a statistically-controllable manner will not be straightforward. I speak as someone who has undertaken market research in developing countries, and encountered cultural challenges in applying standard western statistical techniques. Thirdly, in any case, assessment of effectiveness of actions is problematic. Over what time period and according to whose criteria will effectiveness of policies be assessed? These are not questions with simple, domain-independent answers.

Add new comment