Weekly links September 6: Lots of goodies that caught my eye while on August break

|

This page in:

Here is a selection of some of the things that caught my attention while we were on August break:

·         Simon Heβ put up an FAQ on his popular ritest Stata command, covering issues like how to implement with diff-in-diff, and with multiple treatments. See our previous blog post for more on this command.

·         On the Stata blog, doing power calculations by simulation for regressions

·         Five examples where researchers in applied economics often appear to be estimating the wrong thing – John Gibson discusses issues like whether or not to use a tobit, dealing with sample weights, and quality-quantity adjustments to price movements that often come up in applied work  - noting that “Some have effects that are plausibly much larger than the effects of using randomized versus observational research designs”

·         Dave Evans on “how big is a big impact?” (in education, in standard deviations, in developed countries).

·         Q&A with the JEEA’s former managing editor on questions about what editors are looking for when you send your paper to a top general journal.

·         On the IDB Invest blog, Irani Arraiz reports on their replication of the Drexler et al. rules of thumb business training approach. They allocated 2,400 microenterprises into control, traditional training, and rules of thumb, and find no effect of traditional training, but that the rules of thumb increased sales by 7% and profits by 8%, with gains a bit larger in percentage terms for women.

·         A PBS Newshour story covers the ultra-poor experiment in Ethiopia. What I particularly liked was they talk multiple times with Tedros Kesete, one of the field team, and show the tough work of trying to track down households, and discuss the challenges of getting the control group to continue participating.

·         When field research goes awry – Stanford PhD students Odyssia Ng and Solene Delecourt offer a window into the messy process behind implementing a field experiment, discussing their work setting up a firm to test various explanations for the gender gap in profits. See my post from earlier this year for more on this type of experiment.

·         The AEA randomized trial registry introduces digital object identifiers (DOIs). Keesler Welch blogs about this change on the J-PAL blog – the main thing is that DOIs don’t change, whereas URLs can. It also allows one to make clear exactly what version of the registry entry a paper is referring to.

·         Dealing with the dreaded blank page? – Raul Pacheco-Vega offers advice for how to get started writing research.

·         Applications of Machine Learning in Development Economics – nice set of lecture notes by Michael Koelle.

·         On Let’s Talk Development, Humans of Policy Research interviews government counterparts and implementation partners in an ongoing impact evaluation in an Ethiopian industrial park. “What can researchers do to make their research more useful? Hayat (the implementation partner): The most important thing would be to produce data analysis and policy and program-specific papers throughout the course of the project, rather than waiting until the endline.”

·         Banerjee and Olken are offering an online class on political economy and economic development as part of the MITx MicroMasters program in Data, Economics, and Development Policy. Classes started this week and you can register up to October 4th.

Authors

David McKenzie

Lead Economist, Development Research Group, World Bank

Join the Conversation

September 09, 2019

As an addendum to the post above on power calculation simulations in Stata, here is a talk given at the 2019 Stata conference, and some examples, of some different approaches to coding these simulations:
https://github.com/tpmorris/TheRightWay