known_biases: Shared Columbia/Ginsberg controlled-feeding lab, protocol and small young-healthy-volunteer pool across the 1994 men’s and 1995 women’s dose-response trials; a systematic lab measurement bias or the narrow selected population would move both estimates together.