Blog links April 4: when not to measure impact, the impact of girls playing sport, bogus complaints, and more…
A couple of relevant pieces in the Stanford Social Innovation Review – Mary Kay Gugerty and Dean Karlan have a piece “measuring impact isn’t for everyone ” discussing under what circumstances you shouldn’t try to measure impact (basically if you already know it works well (e.g. vaccines), or if you can’t credibly measure the impact (so aren’t closing a knowledge gap by trying to do so). Holden Karnofsky of Givewell has a piece responding to questions over the impacts of “give them cash” programs like Give Directly, challenging those who claim big impacts of alternative programs to show us the evidence .
A special issue of the Review of Income and Wealth is on poverty, development and behavioral economics – includes, among others, Datta and Mullainathan on behavioral design, and Karlan, Ratan and Zinman on a review of research on savings by and for the poor.
At the NYU Development Research Institute blog, Eva Vivalt draws 5 lessons for better reporting of RCTs based on 400 studies in her AidGrade database.
On the Freakonomics blog, Ian Ayres claims a causal impact of girls playing sport on them being less religious and more likely to be single mothers later in life . Uses instrumental variables, based on the introduction of Title IX in the U.S.
Jason Kerwin on bogus complaints about the use of discrete variables (also summarizing new work in the JPE on using Progresa to test the collective vs unitary household).