Mixtape University series: Diff-in-diff with a checklist. Where does Parallel Trends come from?
Today’s video is the 3rd video entry into my new Mixtape University series, “Diff-in-diff with a checklist”. I had planned to introduce a fourth but due to a filming error, I will have to do that next time.
I introduced the concept of both the Mixtape University and the series on here last week:
Introducing Mixtape University: The Diff-in-Diff Checklist (and Why I’m Doing It)
Today, I'm excited to announce something I've been wanting to do for a long time: Mixtape University. Many of you may have noticed this tab quietly floating around my Substack for a while, but now I'm ready to bring it to life. Let me share a little about what this is going to be about.
Today’s video is kind of fun for me for a few reasons. First, I get to go from basic description of averages to potential outcomes to causal effects to forms of bias, all using handwriting on an iPad with an Apple Pencil. I don’t know what it is, but I love solving stuff using pencil and paper, and I’m going to include the Apple Pencil in that circle of love too.
But basically, the video is a short 15 minute video in which you see how I personally formalize diff in diff from a simple calculation to identification of average treatment effects and the situations where the biases are zero. The video focuses on deriving the parallel trends formula, and also shows how to recognize it (it’s a diff in diff on Y0 equaling zero), as well as where the “no anticipation” assumption shows up.
I highly encourage students starting out to just work through these substitutions themselves over and over until it feels natural. Potential outcomes is funny because it’s like its own language. But I’m horrible at languages! My friends are like Shakespeare with potential outcomes, but I’m not. But I actually enjoy exploring all the diff in diff stuff with potential outcomes because it’s so easy to do. It sort of feels like you’re solving a jigsaw puzzle.
These videos are part of Mixtape University which is my effort to create more value for my paying subscribers. Consider becoming a paying subscriber if you aren’t already and you can then get Mixtape University, the Mixtape Mailbag, and all explainers throughout the week (as opposed to only half in expectation due to randomized paywalling policy I use). And if you are already a paying subscriber — thank you!
Keep reading with a 7-day free trial
Subscribe to Scott's Mixtape Substack to keep reading this post and get 7 days of free access to the full post archives.