I can't find the actual article they refer to, but on the surface this sounds like a badly set-up experiment. For starters, sample sizes of 8 and 9 are extremely low. There is no repetition. The two variables ("water" and "high fat-high carb") are confounding.


Any first-year graduate science student learns how to set up a priori a basic Analysis of Variance test that would separate the variables and detect the fat-carb interaction in the results. If this is really all they did I wonder how on earth they got funding from the William G. McGowan Fund for it.