The word ‘research’ is doing a lot of heavy lifting in the world right now.
A viral article says ‘research shows this food causes cancer.’ A TikTok claims ‘studies prove this supplement works.’ A news headline declares ‘scientists find that night owls are smarter.’ And your brain goes — okay, cool, that sounds official.
But here’s the thing nobody tells you: a study where someone surveyed 12 of their classmates outside the campus dining hall uses the exact same word as a clinical trial involving 50,000 people across dozens of hospitals. “Research.” Same word. Wildly different weight.
Size actually matters here — a lot. Researchers at the University of Minnesota analyzed brain imaging studies and found that most published studies were done with too few participants to produce reliable findings. They analyzed data from nearly 50,000 people across publicly available datasets to examine how sample size affects reliability. That’s not a quirk of brain science — that’s how statistics work in general. Tiny samples give you noisy, unreliable results, and those results don’t hold up when someone tries to repeat the study.
How small is too small? According to a tutorial published in the Journal of Cognition, even for a simple comparison in psychological research, you typically need more than 50 participants just to have a reasonable shot at detecting a real effect. And that’s the floor — not a great number, just the minimum before your results start meaning something.
So when a study has 12 people in it, the findings might be interesting. They might point toward something worth investigating. But they can’t tell you much with confidence. The sample is just too small for the math to work in your favor.
Your brain doesn’t automatically sort for this, though. It hears ‘research says’ and files it under ‘probably true.’ That’s not a character flaw — it’s just how we’re wired to process authority and expertise. The problem is that the word ‘research’ has been stretched so far that it now covers everything from a Nobel-winning multi-decade study to a guy with a clipboard asking ten people a question.
A few things worth asking next time you see a study cited anywhere:
How many people were in it? A few dozen is interesting. Thousands is meaningful. Tens of thousands is where real patterns start to emerge reliably.
Who were the participants? Twelve college students at one school aren’t a stand-in for all humans. Even 500 people from the same city might not be.
Has anyone else replicated it? One study proving something is a starting point. Multiple independent studies finding the same thing is when you can start believing it.
Who paid for it? A supplement company funding a study on their own supplement is not the same as independent university research. That doesn’t automatically make the findings wrong — but it’s worth knowing.
None of this means you should dismiss every study you see. It means you should ask one quick question before you forward something or change your behavior based on it: how solid is the ‘research’ behind this, actually? Sometimes it’s very solid. Sometimes it’s twelve people and a clipboard. The word won’t tell you which.