A 42-year-old childhood obesity theory may be wrong