I haven't updated my blogroll in a while. Blogroll, now there's an obsolete term. One that I've been reading the last year or so is From Poverty to Power, by Duncan Green at Oxfam. Recently he posted a list of the blog's ten most-read posts from 2019. A couple are really interesting. One, by the outgoing CEO of Oxfam, is about strategic planning. It includes this chart, which I just love:
Another, by the feminist development scholar Naila Kabeer,* is about the blindness of randomized controlled trials in development to human agency; in other words to why interventions that appear to work, work. An extremely relevant critique especially now that Duflo, Bannerjee, and Kremer have won the Nobel.
On that subject, we read a post from another blog for Social Diversity, Inequality, and Poverty class this week about the limitations of RCTs in development. I've always been skeptical of RCTs, even when I first became aware of J-PAL and Esther Duflo 7-8 years ago (thanks Caryn). They're expensive and have little independent ability to learn anything outside the very narrow context that they set out to study. The author of that blog points out (paraphrasing) that learning something through an RCT about incentivizing teachers to show up to class in rural Kenya is unlikely to tell you anything that you couldn't already intuit through theory about incentivizing teachers to show up in rural India. And in fact, it's crazy to expect an RCT to do that! They're only good at measuring internal validity (a good new term for me).
We're going to spend a lot of time this term learning about measurement, which is exciting. I spent 10 years doing super conventional development project M&E: log frames, PMPs, baseline surveys, etc. It'll be fun to learn about those things anew and then to learn what else is out there in terms of how to figure out whether policies and practices are having the intended effects.
*Who rules, by the way.
Another, by the feminist development scholar Naila Kabeer,* is about the blindness of randomized controlled trials in development to human agency; in other words to why interventions that appear to work, work. An extremely relevant critique especially now that Duflo, Bannerjee, and Kremer have won the Nobel.
On that subject, we read a post from another blog for Social Diversity, Inequality, and Poverty class this week about the limitations of RCTs in development. I've always been skeptical of RCTs, even when I first became aware of J-PAL and Esther Duflo 7-8 years ago (thanks Caryn). They're expensive and have little independent ability to learn anything outside the very narrow context that they set out to study. The author of that blog points out (paraphrasing) that learning something through an RCT about incentivizing teachers to show up to class in rural Kenya is unlikely to tell you anything that you couldn't already intuit through theory about incentivizing teachers to show up in rural India. And in fact, it's crazy to expect an RCT to do that! They're only good at measuring internal validity (a good new term for me).
We're going to spend a lot of time this term learning about measurement, which is exciting. I spent 10 years doing super conventional development project M&E: log frames, PMPs, baseline surveys, etc. It'll be fun to learn about those things anew and then to learn what else is out there in terms of how to figure out whether policies and practices are having the intended effects.
*Who rules, by the way.
No comments:
Post a Comment