analysis · Claude · free
Audit this claim before I put it in a slide. The claim: [PASTE — e.g. "users who do X have 3x retention"] The data behind it: [DESCRIBE — obs / experimental, N, source] The proposed action: [WHAT WE'D DO IF WE BELIEVED IT] Deliver: Is this correlation, causation, or noise? Verdict + confidence. Selection bias present? (Who ends up in the "X" group and why?) Reverse causation possible? (Does Y drive X instead?) Confounders that could explain the relationship Survivorship / recall / measurement bias The one experiment that would actually establish causation Weakest version of the claim I could defend today The strong version I'd need more evidence for If someone would call this a "correlation not causation" tweet dunk, tell me.
#causation #bias #critical-thinking