· Everything is significant when you have big data – Dave Giles on how we should be using different significance thresholds when the sample sizes get big.
· In Science last week, Stefano DellaVigna, Devin Pope, and Eva Vivalt highlight the potential usefulness of getting ex-ante predictions about research results. For example, they note this could mitigate the bias against publishing null results “if priors are collected before carrying out a study, the results can be compared to the average expert prediction, rather than to the null hypothesis of no effect. This would allow researchers to confirm that some results were unexpected, potentially making them more interesting and informative, because they indicate rejection of a prior held by the research community; this could contribute to alleviating publication bias against null results.” They have launched a prediction platform to help centralize the collection of these priors.
· Banerjee and Duflo on myths about migration, on DowntoEarth.
· Stata code by Hans Sievertsen on how to display a specification curve that shows how your parameter estimate varies under a range of different specifications. See the curve itself on twitter here. See also this paper by Simonsohn et al. on this approach to robustness.
· Dave Evans summarizes all of Esther Duflo’s papers.
· Give Directly’s 10 things they got right and wrong over their first decade: #5 is deciding to saturate entire villages instead of targeting specific households within them.