Why anecdotal evidence is better than studies – Grey Enlightenment
Skip to content
Everyday we’re bombarded with information, such as studies and advice. So what can be trusted or dismissed? Most studies are bullshit–just Google "replication crisis". But a possible exception is when the science agrees with empirical evidence, such as on Reddit and other social media. Reddit IS the science. There is no need for studies. Anecdotal evidence on Reddit is better than studies.
So when a fitness influencer says protein is better for weight loss or more satiating than carbs, I say “nope—that is not what the evidence shows.” Indeed, anecdotally, it makes no difference, and this also backed by studies. On popular dieting "subreddits," low-carb dieters in the aggregate are NOT more successful than the high-carb dieters; they equally struggle at weight loss. There is no specific food or nutrient that is better than the other. It’s all highly individualized. For some individuals, keto works great; for others, not so much.
Very seldom is there a good study that also goes against the empirical evidence. This is why “bro science” is ahead of the studies. For example, many people report diminishing returns to cardio and weight loss:
This is corroborated by studies, such as Herman Pontzer’s research on the so-called “constrained total energy hypothesis,” which posits that the body becomes more thrifty at burning calories when exercise is increased. So again, studies only affirm what people have long experienced/observed. For example, fitness communities were the first to experiment with Retatrutide via compounding pharmacies, and noted its effectiveness. This was around 2024, in agreement with studies published in 2026 noting its efficacy at weight loss.
Another reason to trust anecdotal evidence is there no conflict of interest. What motive is there for an anonymous person to misrepresent their personal experience on Reddit, for example, when there is no obvious incentive. It’s not like careers or reputations are on the line, compared to academic researchers who need to please deans, secure funding, and bring attention for their findings, leading to sensationalist headlines and the misrepresentation of results.
This is not to say all studies are garbage or should be dismissed. A good heuristic is, "Who stands to gain from this being true? What is at stake?" Studies purporting the predictive power of IQ vs life outcomes, I am at least more inclined to trust because there is nothing financially gained from this knowledge. Trillions of dollars of federal funding hinges on the misguided assumption that people can be molded under some image of equality despite evidence that individuals differ substantially in ability and outcomes.
You cannot create a government program or sell an expensive course to raise IQ if it’s innate and stable between individuals. (This is not the same as raising national IQ, such as through improved nutrition, in which the low hanging fruit has mostly been picked.) Many people can relate to growing up and noticing that some individuals simply operate at a "higher level" than others, and that this gap cannot be closed through “better teaching” or “more enrichment.” It often seems intrinsic to the individual, and can even be observed among siblings raised in the same environment.
Likewise, research showing the near 100% long-term failure rate of dieting, runs counter to the interests of the multi-billion dollar fitness/dieting industry, whose business models depend on the belief of dieting and exercise as effective for weight loss. This research agrees with empirical evidence online of people consistently struggling to lose weight and relapsing into their old habits. This is not to say this heuristic always works, but between competing narratives "dieting is really effective" vs "it does not work well," I am more inclined to trust the latter, as there is much less at stake in it being true.
Another problem with studies is that they often aren’t applicable to individuals. Many rely on mice or rats, or on large groups of people lumped together, where it’s hard to ascertain what the effect, if any, actually is for a given person. Fitness studies will often define very low thresholds for what is considered "expert," that is not applicable to someone with more gym experience. Or muscle hypertrophy is conflated with increased blood volume. Or "lean mass" is conflated with muscle. A longitudinal study widely cited by the media involving thousands of subjects may report that drinking one or two glasses of wine per day reduces life expectancy by 0.2 years. How meaningful is such a finding in practice? What is one to make of this? Who knows.
This is again why anecdotal evidence s better, because can find people whose profile or habits matches mine more closely. If, suppose, I have 10 years of gym experience lifting heavy, I can find people online with similar experience lifting similar weight and see...