Most of nutritional "findings" are based on sketchy statistics run on inaccurate data obtained from really bad experiments. If anything, what little evidence there is suggests that there may not even be such a thing as "healthy" or "unhealthy" food, but food that may be more or less healthy for a specific individual. And what is more certain (from large scale statistics) is that the effect food has on our health (within a reasonable margin) is quite small in most healthy people without some genetic anomalies. What makes it seem large is simply magical thinking (or placebo, if the two are different).
Which raises the far more interesting question of why is it that otherwise scientifically minded people treat nutrition as a science rather than the "science" it really is? The psychologist and anthropologist Jonathan Haidt believes that this is a psychological manifestation of a drive for "purity" (one of several common moral drives he's classified in many societies) as it expresses itself among some people, especially liberals. He says conservatives channel purity to sex and liberals to food. He says people in all cultures believe that some behaviours contaminate us; conservatives tend to think those behaviors are sex related, and liberals think they're food related. Which is why accounts of people "feeling more energetic" after changing their diet read just like conversion stories -- they are simply two manifestations of the same phenomenon. And just like what religion people convert to doesn't matter, the accounts of being "more energetic" and "feeling better" are the same no matter what diet people switch to (of course, sometimes you read other accounts, but such is the nature of placebo -- it's complicated). As long as there is the slightest excuse to justify a particular choice of diet -- and in the case of science fans that can be a "science" paper -- psychology kicks in.