04. Causal Graphs: Organizing Your Thinking With Arrows

Analyze anything that wasn't a randomized experiment and report it, and someone in the room will ask: "did you control for confounders?" or "what if you add a few more variables and rerun it?" Deciding what actually needs controlling for — and what actively makes things worse — turns out to be hard to get right by gut feel alone.

What Happens in the Meeting Room

More variables feels safer. It usually isn't...

💎

Premium membership required

Upgrade to premium to access the full chapter.