Nic Duquette Mar 21, 2023
Is there a single JEP- or handbook-style paper on "how to run an economics (social science) experiment"? Thinking about a step-by-step explanation of writing a preanalysis plan, making ex ante power calculations, getting IRB approval, stuff like that.
Show threadOwen OzierMar 21, 2023

@nic the frontier keeps moving but:

the Glennerster and Takavarasha book is good

and the Duflo Glennerster Kremer “toolkit” paper ( https://scholar.harvard.edu/files/kremer/files/randomization_toolkit_dev_economics.pdf ) is excellent, though pre analysis plans basically arrive in the social sciences after these.

Miguel and Christensen (in both book and paper form) reflect on advances in transparency.

https://www.aeaweb.org/articles?id=10.1257/jel.20171350

https://www.nber.org/papers/w22989

Also tagging @ryancbriggs

Show threadRyan Briggs
@owenozier @nic those are my standard picks. Depending on your goals, I’d love if you included things on workflow or at least pointed people to that stuff: here is how you organize your folders and code, etc. I think people mess that up around as often as they mess up statistical issues.
Mar 21, 2023 at 2:01pmIceCubesApp
Show threadOwen OzierMar 21, 2023

@ryancbriggs @nic totally agree, great point Ryan. here is Julian Reif’s coding guide for projects along these lines!

https://julianreif.com/guide/

Stata Coding Guide

Julian Reif
Show threadNic Duquette Mar 21, 2023
@owenozier @ryancbriggs great resources, thanks @owenozier @ryancbriggs !
Show threadOwen OzierMar 21, 2023

@nic @ryancbriggs

Pam Jakiela and I teach this material. Our syllabi may be useful, plus some slides and other materials.

Our Barcelona School of Economics RCT course about this:

http://economics.ozier.com/bgse-2019/

https://bse.eu/study/summer-school/development-economics

And Pam’s Program Evaluation course at Williams:

https://pjakiela.github.io/ECON523/

BGSE Development: Randomized Controlled Trials, Summer 2019

Show threadNic Duquette Mar 21, 2023

@owenozier @ryancbriggs These slides are great, and I'm going to steal from them ruthlessly.

FWIW the use case I have in mind is a doctoral student who wants to run a survey experiment, not my own work or development RCTs specifically. It's not a topic I currently cover in my PhD level methods class, but I've decided to add it because of student interest in writing these papers. They need to know how and why to do a power calculation before they run something with N=60 and ten treatment arms.

Show threadRyan BriggsMar 21, 2023
@nic @owenozier a bit out of date, but maybe useful for that use case https://alexandercoppock.com/Green-Lab-SOP/Green_Lab_SOP.pdf