There is something double strange about the p values.For the main effect, the broad based decline, p < 0.0001. But for the result for men at "normal" weight, 18

There is something double strange about the p values.For the main effect, the broad based decline, p < 0.0001. But for the result for men at "normal" weight, 18.5 < BMI < 24.9, the p value is given as p< 0.05. The first strangeness is that this hints at a much weaker effect. The second strangeness is reporting the p value rather than the effect size. I clicked the link to a pdf at the bottom of the article, but it is not the paper. shrug