No. Most nutrition studies are “natural experiments,” simple observational studies of populations. That is, they correlate some measurement of people’s health with their reported diets.
Less common are studies where diets are actually manipulated, because of practical and ethical issues, but that happens also.
But rather than argue “more or less” just know that researchers often use observational studies but these are most often used as exploratory strategies. As in, “a low-carb high-fat diet seems to lead to better health in Americans, let’s do a more targeted study, an experiment using restricted diets in a carefully measured group.”
These kinds of observational studies are great at generating hypotheses. The reason that the scientific community at least doesn’t tend to go overboard on publishing the results, however, is that they are not empirical. They can yield helpful data, but do not establish causality.
This is contrasted with an experimental study design where you have two groups of people with consistent habits (say, for example, a high-cholesterol diet). They can take one of the groups and begin limiting its cholesterol intake, observing the changes it experiences compared to the group still on the same meal plan. By deliberately introducing and studying these kinds of variables, we can conclusively arrive at provable and testable statements like “reducing cholesterol improves heart function,” which is a correlation that might have been found by looking at healthy people, but could not be proved until statistical controls were in place.