Talking to People to Understand Selection is first best and if that still isn't enough, then I have some other suggestions: Step 4(f) of Pedro's Diff-in-diff Checklist (my last entry!)

When it comes to Difference-in-Differences (DiD) analysis, one of the most crucial steps is choosing the right control groups. For the past two months, we’ve been diving deep into Pedro Sant’Anna’s checklist for DiD, meticulously working through each step to ensure robust causal inference. We’ve spent the last few weeks specifically focused on Step 4, w…