This is the second in my weekly Monday series on Codechella Madrid (May 25-28, 2026). Last week I gave you the basics.
This week I want to tell you what’s new. New material and Claude Code stuff!
Last Monday I told you I’d be back every week with more about what we’re doing at Codechella this year (sign up information here), and I meant it. So here I am — a little more caffeinated, a little more organized, and ready to tell you about two things that I think make this year’s edition meaningfully different from the first two.
New Material
The difference-in-differences literature hasn’t slowed down. If anything, it’s accelerated. And this year’s Codechella reflects that. We’re adding several important topics to the curriculum that weren’t part of the workshop before:
Continuous difference-in-differences, including event studies — this draws on Callaway, Goodman-Bacon, and Sant’Anna, which has been conditionally accepted at the AER
Synthetic difference-in-differences (Arkhangelsky et al. 2021, AER)
Triple differences (Ortiz-Villavicencio and Sant’Anna)
Compositional changes in treatment and control cohorts within a DiD framework
Functional form and parallel trends (Roth and Sant’Anna 2023, Econometrica)
Bounding exercises when parallel trends is violated (Rambachan and Roth)
That’s a lot of new ground. To make it work, I’m reorganizing the talks so the material flows more naturally from foundational to modern to frontier. But the goal hasn’t changed: this workshop is practical, applied, and accessible. We assume very little background. If you know what a regression is and you’re willing to get your hands dirty with code, you’re ready.
Claude Code
Here’s the other big change, and honestly it might be the one that gets some of you most excited.
I’ll be doing my replications and demonstrations using Claude Code throughout the workshop. While we are still trying to figure out if we want to also run a separate workshop on it, at the least I will be constantly showing you how to work with these new estimators by illustrating it inside of Claude Code environments. The idea is therefore that you learn how to use Claude Code for applied quantitative research, and contemporary diff-in-diff and synthetic control methodologies more specifically, as I think maybe learning Claude Code in the context of using it for things you were already planning on doing anyway (e.g., making event study graphs, tests for violations of parallel trends, making beautiful tables and beautiful figures!) is probably the best way most non-computer scientists are going to both learn about AI agents, as well as maybe be brought to care about it too.
But if you’ve been reading this Substack, you know I’ve been writing about how Claude Code has changed the way I do empirical research. CodeChella is where you’ll get to see that in practice. I’ll be using it repeatedly across the diff-in-diff and synthetic control material, which means the workshop doubles as a hands-on introduction to AI-assisted applied research.
And I think the timing matters. AI coding agents like Codex are arriving fast, and tools like Claude Code are already here. One of the most important skills we can develop right now is verification — knowing how to check what these tools produce, how to build habits that catch mistakes, and how to structure your workflow so that speed doesn’t come at the cost of credibility. That’s part of what I want to teach. Not just “look how fast I can run this,” but “here’s how I make sure what I’m running is right.”
So CodeChella this year is two things at once: the best workshop we’ve ever built on modern difference-in-differences and synthetic control, and a real-world demonstration of how to use Claude Code in applied empirical research. I think the combination is going to be something special.
Come to Madrid
Codechella runs May 25-28 at CUNEF Universidad in Madrid. Four days, 9am to 5pm, with morning coffee and pastries included. Pricing:
Students: $220
Post-docs: $300
Faculty: $500
Tickets are available on Eventbrite here. If you’re a student or post-doc and need a discount code, email me at causalinf@mixtape.consulting. I want cost to be the last reason you don’t come.
Madrid in late May is perfect. The weather, the food, the city — it’s one of my favorite places in the world to teach. Come learn with us.
I’ll be back next Monday with more.
Codechella in Madrid sounds perfect Scott! Will you also be live streaming or recording as well?