Observational comparisons (e.g., Japanese vs Americans, Korean vs American soldiers' autopsies) can only generate hypotheses, not establish causation, because you don't know what variables you're not looking at — sugar consumption tracks just as well as fat, but is ignored because it isn't the hypothesis.

causalpending

Speaker

Gary Taubes

Evidence Quote

“you don't know what you're not looking at so you focus on fat because that's what your hypothesis about this is a endemic problem in public health”

Source

Gary Taubes on Fat, Sugar and Scientific Discovery 11/21/2011— EconTalk
Created: 6/13/2026, 7:03:37 PM

Your notes

Loading notes...