Here's what I think (I have a Bachelors in Biological Sciences and worked as a research assistant in a physiology lab as an undergrad):
The human body is so complex that science has a hard time with it, and due to ethics, humans make terrible experimental subjects. Plus, science tries to boil down the observed results of an experiment to a single cause, or at most a handful of causes. But, at the level of the human body, there is essentially never just a single cause for anything - its all an interplay of many many factors. For this reason, it is almost impossible to do a well controlled experimental study of the effect that a substance/behavior (drug, food, diet, type of exercise, etc) has on humans and maintain relevance. Basically, for most of the "controlled" studies that are done on diet/exercise/health their results are either:
a) So ambiguous that its almost all up to the interpretation of the researchers to draw any conclusions, and thus the personal/structural/funding biases of the researchers come too much into play.
b) Conclusive, but only conclusive because the setup of the study predetermined the outcome. For example, a study that shows that a "low fat, high fiber" diet reduces heart disease/cancer, etc. Well, if you look at a lot of the studies that have shown that, then you'll see that the researchers are comparing the low-fat, high fiber diet to the Standard American Diet (SAD), which is basically McDonalds and Velveeta... While the results may be applicable to the so called "Standard American," I'm not the standard american myself (for many reasons), so the conclusions of that study have little to no relevance to me. Furthermore, I think that the SAD is probably about the worst diet you could possibly eat, and therefore ANY change is going to show improvements... Ironically, there are some studies that have tried to look at multiple diets (ie, atkins, vs. ornish, vs. US Govt. recommendations, etc), and they have more or less found that all the diets they test tend to have about the same effect (moderate improvement) when compared to the "control group." This to me indicates that the control group is not really a control group at all...
So with those two points in mind, when thinking about my own health, I tend to take controlled studies (normally the gold standard of science) with a very large grain of salt. So, how should we make health decisions then? Well, a more holistic approach is called for, that combines information from (in order of broad to specific):
a) Epidemiological research studies (basically looking at populations that already do something different and comparing them to a population that is more like you. Sounds great in theory, but their results are often very general/fuzzy/inconclusive since its impossible to isolate out individual factors).
b) Controlled research studies (but only some of them, and even then, determining their value based on my own opinions about their biases, funding sources, deficiencies, etc).
c) Evolutionary biology (we evolved to be healthy, not sick/full of heart disease/obese. The idea is to find out what conditions we evolved under and replicate them as best we can through diet/exercise/behavior/supplementation, and the so-called "diseases of civilization" will go away as well).
d) Clinical evidence (ie. case studies). If it worked for someone else in a clinical setting, it may work for you... If it works for lots of people in lots of clinical case studies, then its even more likely to work for you, though by no means guaranteed.
e) A study of n=1. The idea of a n=1 study is that the number of participants is 1: yourself. Clearly a study of yourself would not be applicable to anyone else, but ultimately they are all conducting their own n=1 studies and therefore its irrelevant what they are or aren't doing to themselves - the only thing that matters is you and how you respond to something.
Personally, I put by far the most credence in option E.
There is no gospel when it comes to nutrition.
Truth doesn't change, whether you deny it or face it.
There is no universal truth about diet that works for every single person.
There is no "infallible 100% optimal human diet." Trying to find "the optimal human diet" is a fools errand.
Instead, try things out in your own experiment of n=1. Eat more protein. See what it does. Eat less protein. See what it does.
The only thing that matters is how you respond.