Tuesday, May 18, 2010

Esther: Randomista

Esther Duflo is kind of a big deal right now. At least among the Economics and Development crowd. Duflo is a French economics professor at MIT who is at the vanguard of a movement for better data in economic development research using randomized controlled trials (RCTs). Randomized controlled trials are an attempt to make real life into a laboratory with studies that include control and treatment (intervention), much as scientific medical research does. The idea is to get stronger evidence about what actually works in development. For donors, its about knowing whats going to get the most bang for their buck.

RCTs have been around for a number of years but the debate surrounding them has gotten really lively and actually quite interesting as of late. So much so that Ms. Duflo was invited to give a prestigious TEDtalk about randomization in economic research. While RCTs have pros and cons like pretty much everything, the idea of using better information when donating to social causes is important to everyone. At the very least people should be thinking about how aid groups are going to spend their money and on what type of intervention. The data is a lot stronger for some interventions than others. I suggest everyone take a look at Esther Duflo's Ted Talk here. And even if there's are points that are debatable, the average person should be enlightened by her message. She even admits that its not a magic bullet and that poverty is going to be around for a long time, but why not employ it if it provides us with useful data? If the content isn't quite up your alley, I still encourage you watch the video if for no other reason than the fact that she has a slightly humorous French accent.



If the idea of RCTs really resonates with you then I would recommend you pickup a book I just finished reading called WHAT WORKS IN DEVELOPMENT. It's a collection of essays edited by the popular economist WIlliam Easterly about the merits and drawbacks of using randomized controlled trials in development research. RCT detractors claim, among other things that there's an "external validity" problem, in other words, what appears to work in one place may or may not work elsewhere. Confounding the problem is the fact that once a certain type of intervention study is published once, there's less incentive for researchers to replicate it because its easier to simply trumpet the original findings than spending the time fundraising for and coordinating replications.


Still the debate has fascinated me as I've begun to delve deeper into the realm of more complex economic research. Who knows if these types of experiments are just a trend or not, but at least the conversation surrounding them is making people more aware about the importance of measuring results when it comes to poverty reduction efforts. Another blog hits the nail on the head with its title "Good intentions are not enough". Not every idea is a good one in development and not everything works. A flop of a concept called 1 million t-shirts is an excellent example.

The whole aid effectiveness discussion is about results and accountability. Because seeking better information about what works is something we can and should do. Better data can spur innovation in business and social enterprise. It can lead to initiatives that are more efficient at helping people to lead healthier lives and opening doors for future generations by convincing them to get an education. A great example of the freeing of data is aiddata.org.

I'm a newcomer to all these issues, but I'm enthralled and I'd like to see where this conversation goes. Now has anyone done a RCT in the Gambia?

No comments: