· Scott Cunningham interviews Amy Finkelstein. Part of the discussion is about how she found out about the Oregon health experiment lottery, and how some of the issues of designing data collection for that experiment differed from what development economists normally have to worry about – including much higher attrition in surveys.
· John List offers his tips for doing better field experiments and getting your work published. These include considering power and how much a study will shift priors, thinking of stealth within-subject designs to recover other parameters than the ATE, making sure to deal with multiple hypothesis testing “The chief reason for false positives in experimental economics is not p-hacking, file drawer issues, or lack of replication, it is not adjusting for multiple testing”, think about blocking/stratifying on key situational factors that will matter for scaling, he doesn’t like the term “lab-in-the-field”, and more…
· Dave Evans offers a round-up of research on education presented at the recent RISE conference.
· Nick Huntington-Klein on interpreting coefficients from interrupted time series regressions and the importance of being clear whether you are interested in a discrete or instantaneous change.
· Connectedpapers.com allows you to put in a paper and it analyzes co-citations and overlapping references to identify which other papers are most similar – a potentially useful way to check your literature reviews. (via Sarah Rosenberg).
· In Nature, Abel Brodeur and co-authors have a comment describing the Institute for Replication’s replication games approach, and its efforts and goals to more systematically re-analyze papers. The journal has illustrated this with a lovely image that makes literal the games aspect, showing papers lifting weights, hypotheses being thrown like a shot put, and more…
· In a VoxDev talk, Robin Burgess and Kelsey Jack outline some broad takeaways on research on environmental economics in low- and middle-income countries- as well as some of the key policy questions emerging, such as whether social protection schemes need to be designed differently when shocks are now climate shocks that are aggregate (and hard to risk-share) and can be catastrophic. Also a good reminder for the ongoing BREAD-IGC virtual PhD course on environmental economics, which has the slides and recorded lectures up at this site.
· Call for papers: The ThReD Conference on theory work in development will take place at MIT Sloan on March 1-2, with submissions due November 1.
Join the Conversation