Researchers rely on epidemiology to control for confounding factors, but it is often impractical to conduct long-term randomized control trials. This makes it easy to manipulate nutrition science to prove any point, and it is built on a "house of cards."