The evidence required for these kinds of studies is far greater than it might initially seem. The sheer number of other variables (environmental factors) makes it very easy to assign causal relationships incorrectly. I mean to say that I’m willing to bet most individual studies involving “32” participants (or on that ballpark) are basically wrong. By individual studies I mean those where the evidence for the phenomena is almost just that study.

Some anecdotal evidence (lacking being able to share the resources that I’ve read in the past) is the “science” in (and media reporting of) the fields of nutrition, psychology, sociology and similar fields where the environment is impossible to fully control. They seems half-baked (or outright contradictory) because the media reports on findings that don’t have very strong evidence behind them (need further research) while making it seem like it is an established fact.

It's far too small of a sample size for something as complex as human diet and exercise. You'd need a much larger sample size with people of varying types of diet and exercise regimes.