Thread by @TDeryugina

A new blog post with some basic guidelines for how to conduct robustness analysis in quasi-experimental empirical work (+ some general tips).

deryugina.com/some-tips-for-…

#EconTwitter

Overarching principle: think first. Do not try every possible permutation of samples/fixed effects, etc. Think about what concerns readers might have & design robustness checks to address those.

Pick your preferred regression specification/sample before you know what your results are. You are not necessarily wedded to this, but it will help you be on firm ground conceptually.

When doing robustness checks, change one thing at a time based on the most likely threats to your empirical approach. There may be good reasons for changing >1 things at a time in some cases, but doing so is more likely to lead to too many tables!

Consider adding some falsification/placebo exercises, again based on possible threats to validity.

And, finally, remember that not all robustness check "failures" are actual failures.

See post for more details & please add your own suggestions!