
The Mixtape with Scott Episode 6 of the Odd Couple: How Will We Draw Our Diff-in-Diff?
Apr 14, 2026
Caitlin Myers, a Middlebury professor and applied economist who studies reproductive health policy, helps sketch the study design. They visualize clinic-closure driven travel-distance changes across Texas and debate how to present staggered spatial shifts. Maps, distance categories, population denominators, and placebo checks take center stage as they plan trend plots and next steps.
AI Snips
Chapters
Transcript
Episode notes
Dissertation Ruined By A Coding Topcode Error
- Scott discovered he had top-coded missing values in his dissertation NLSY variable and accidentally replaced missing responses with 10, which invalidated his results.
- He only found the error years later when reconstructing a messy do-file folder, then abandoned the paper despite initial JHR interest.
Enforce Clean Directories To Avoid Silent Bugs
- Organize code and directories intentionally to prevent hidden data-cleaning mistakes and make analyses reproducible.
- Scott shifted from relying on syntax knowledge to enforcing directory structure, clear file names, and re-running do-files consistently.
Figures Must Stand Alone With Clear Labels
- A clear figure should stand alone and convey the main story without requiring readers to first read the paper.
- Caitlin emphasizes precise titles and explicit captions, noting the map clearly states it shows Q1 2013 to Q1 2014 distance changes.

