Blog Links February 7: Better Graphs, Changes at EDCC, Boring Development, and more…
Marcel Fafchamps has taken over as editor of Economic Development and Cultural Change: his opening editorial : “Scientific methods change over time. What was once seen as revolutionary - for example, long panels best exemplified by the Indian ICRISATdata set—was only a few years later seen as inadequate, if not irrelevant, by a new generation of economists seduced by the intellectual elegance of randomized controlled trials, best exemplified by the PROGRESA study inMexico. The only thing we can predict with certainty is that scientific methods will continue to change. As research practitioners, we must welcome new tools for the new opportunities they bring, while at the same time remembering that there never is any silver bullet. No experiment is flawless, no regression is perfect…. Our job is to extract the most accurate and useful information possible from flawed data. Personally, I do not believe I have ever estimated a consistent coefficient estimate: at the very least, my data always had some measurement error. … I do, however, intend to ensure that the journal continues to take advantage of the most current methods of serving our authors and of reaching out to our readers. I have some ideas in this area, one of which is to enable authors to post photographs, short videos, and other testimonial material about their fieldwork on the journal’s website…Development economists often differ from economists in other fields in the strong emphasis we put on direct observation. Let us document that.“
In the latest Journal of Economic Perspectives: An economist’s guide to visualizing data : Before and After plots from real papers and reports, showing how plots can be made to look nicer and get across the point better.
A new blog with a great name: The Campaign for Boring Development – “The real work of development is not glamorous. It’s not exciting. It doesn’t photograph very well”, ironically undermining its case somewhat with quite nice photos to make its points that we need to do more of things like building roads, planting hybrid seeds, and other ways to raise incomes.
Behavioral Economics key concepts presented through Dilbert cartoons on the Psychology Today blog (h/t @RFisman )
Should you use frequentist standard errors with causal estimates on population data? Yes says Abadie, Athey, Imbens, and Wooldridge as summarized by Cyrus Samii.