Randomized controlled trials are kind of a big deal in development economics right now. A recent article in The Economist shows a sizeable rise in the use of RCTs in economics overall over the last 15 years, and recent analysis by David McKenzie shows that RCTs make up a large minority of development papers in top journals (see the figures below).
Source: The Economist on the left; McKenzie (2016) on the right.
In his new book Experimental Conversations: Perspectives on Randomized Trials in Development Economics, Tim Ogden has assembled interviews with a distinguished group that interacts with RCTs in every imaginable way: you have those who pioneered the use of the method in development economics, the next generation of researchers, the chief critics of the method, and consumers of development RCTs at organizations like GiveWell, the Ford and Grameen Foundations, and the Center for Global Development. You also hear from one broader observer of economics as a field (Tyler Cowen) and one of the scholars who pioneered the use of RCTs in U.S. policy (Judy Gueron), to give added perspective.
- Interview with Mark Rosenzweig: “One of the advantages of studying developing countries is that it’s cheap to collect data and the response rates are much higher than in the United States. I’ve helped lead a survey in the U.S. of 8,500 households — it cost $23 million. In India, where the questionnaire is probably eight times longer, the total cost is about $750,000... Five hours is a substantial commitment of time, what’s the response rate? Our response would be somewhere around 90%. People enjoy telling you about their stuff. I’ve surveyed a lot of farmers in India and they want to show you everything. They enjoy it. People there value their time differently than we do. In most villages there are no cinemas or shopping centers there. There’s no television. They enjoy talking to people. That’s different than here. We all have better things to do than sitting down and answering silly questions over the phone, let alone allowing somebody into your house. Sitting down and talking to people is an interesting activity for these folks.”
One of the things I get asked when people are designing experiments – when they are either interested in or worried about spillover effects – is how to divvy up the clusters into treatment and control and what share of individuals within treatment clusters to assign within-cluster controls. The answer seems straightforward – it may look intuitive to assign a third to each group and I have seen a few designs that have done this, but it turns out that it’s a bit more complicated than that. There was no software that I am aware of that helped you with such power calculations, until now...
- A webcast of the AEA panel on “publishing in economics journals: the curse of the top 5” (h/t @DurRobert) – Heckman, Akelof, Deaton, Fudenberg and Hansen discuss. Some interesting discussion and comments – Deaton notes he didn’t have any papers rejected until he was famous; Heckman had a lot of data, including this one which shows (first column) which journals account for most dissemination of the ideas of the top development economists – with WBER number 1:
In my very first experiment, Suresh de Mel, Chris Woodruff, and I gave small grants of capital to microenterprises in Sri Lanka. We found that these one-time grants had lasting impacts on firm profitability for male owners. However, despite these increases in firm profits, few owners made the leap from self-employed to hiring others.
In 2008 we therefore started a new experiment with a different group of Sri Lankan microenterprises, trying to see if we could help them make this transition to becoming employers. Eight years later, I’m delighted to finally have a working paper out with the results.
- Fiona Burlig blogs on her new paper about how to do more accurate power calculations for experiments that use panel data (more T). There is apparently also Stata code, but I haven’t been able to download it yet and play around with this.
- In time for those on the job market, the CSWEP newsletter has advice from a number of economists on how to discuss the dual career search process: lots of different perspectives and advice. One piece discusses how ambiguity aversion means it can be helpful to reveal your status, whatever this is. The majority seem to be suggesting you should disclose this information around the time you are invited for a flyout.
- How much does interviewing with bystanders around affect survey responses? The inaugural blog post from Kantar public Africa and Middle East reports that i) bystanders are present in about half the cases. Bystanders are mostly non-family and extended family members, such as neighbours, domestic staff, but also children - not the spouse. Ii) bystander presence has little effect on non-sensitive question responses, but do for some sensitive questions; iii) the presence of the husband or wife can sometimes improve accuracy.
- From the IDB Development that works blog – a program in Bolivia manages to reduce malnutrition, but made the kids overweight instead.