DATA IS FTAWED

Data is flawed

Data collection for nutrition studies is severely flawed because people simply don't report accurately. They tend to underestimate what they eat, but even more fundamentally, they don't remember what they ate to report it back. And if they eat any meals out, they don't know what all the ingredients in the food were (for example what oil was used). Finally, they (or the nutritionists running the study) don't have accurate data about the actual nutrient content of their food, which can vary depending on freshness and where it was grow or raised.

As Marion Nestle, a prominent nutritionist, says in her book Food Politics, "The most intellectually demanding challenge in the field of nutrition is to determine food intake."
Food in a restaurant

Furthermore, in studies comparing one diet to another, it is not always possible to control what people eat. For example, in a study comparing a group eating a low-fat diet to those eating a higher-fat diet, those in the low-fat group cheated and ended up eating a higher fat diet than the researchers recorded, while those in control group (normal diet) lowered their fat intake voluntary (perhaps because the study made them aware of it). In the end, unbeknownst to the researchers, the two groups were not far apart in the amount of fat they were eating. Thus it was not surprising that the "low-fat diet" group didn't show any health benefits!