Why Vague Advice Feels Personal: The Barnum Effect

In September 1995, three researchers at the American Institutes for Research delivered a report that ended a program the United States government had been quietly funding for roughly two decades. The program was about psychic spying. But the report is not really about psychics, and that is why I keep coming back to it.

The program asked whether a person sitting in a room in Maryland could describe a place they had never visited — a submarine pen, a hostage location, a piece of hardware behind a fence. It was called remote viewing. It ran through the 1970s, 80s and into the 90s, and before the intelligence community handed it to a new sponsor, they commissioned an outside autopsy. Michael Mumford, Andrew Rose and David Goslin assembled a panel that deliberately included one statistician sympathetic to the paranormal and one psychologist famous for his skepticism, gave them roughly eighty publications, and asked them to settle it.

What the panel found is genuinely strange, and it is the most useful thing I have read all year about your wellness feed.

A Real Signal That Was Completely Useless

The lab results were not nothing. Across 154 experiments and more than 26,000 trials at SRI International, the statistical departure from chance was so large that the report describes odds against coincidence on the order of one in 10²⁰. The panel conceded the point plainly: "A statistically significant laboratory effort has been demonstrated in the sense that hits occur more often than chance."

And yet the same report recommended shutting the whole thing down. Not because the numbers were fake. Because of what happened when anyone tried to use the output.

The operational reviewers interviewed the end users — the analysts who actually received the three- and four-page reports the viewers produced. Five findings emerged, and the first one is the whole story: "The information provided in the reports was stated in broad, vague terms." The reports were most accurate about "general stereotypical characteristics." They were inconsistent across independent viewers. They contained large amounts of irrelevant material. And the burden of sorting signal from noise fell entirely on the reader.

Read that again. The hits were not found. They were manufactured, in good faith, by intelligent professionals who wanted the information to be useful. The vagueness was not a flaw in the reporting. The vagueness was the engine. Give a motivated reader something broad enough, and the reader will do the work of making it fit — and then experience the fit as evidence.

The panel's final verdict was that the information was "too vague and ambiguous," and that in no case had it ever been used to guide an operation. Twenty years, a real statistical anomaly, and zero actionable output. Because you cannot act on there is water nearby.

The Study That Named the Effect

Psychology had already documented this in a classroom, decades earlier and for free.

In 1948, Bertram Forer gave 39 of his students a personality test, then handed each of them what he described as an individualized profile. He asked them to rate how accurately it described them on a scale from zero to five. The average rating was 4.26. Nearly perfect.

Every student had received the exact same paragraph, and Forer had assembled it from a newsstand astrology book. It contained lines like "You have a great deal of unused capacity which you have not turned to your advantage," and "At times you are extroverted, affable, sociable, while at other times you are introverted, wary, reserved." He published it under the title The Fallacy of Personal Validation. Paul Meehl later attached the showman's name to it, and it has been the Barnum effect ever since.

The finding has held up across seventy-odd years of replication, including Ross Stagner's 1958 demonstration that seasoned personnel managers rated a generic profile as an accurate read on their own personalities. Two conditions reliably strengthen the illusion: when the person believes the description was prepared for them specifically, and when the content skews flattering. Both conditions are standard operating procedure in the wellness industry.

Why Your Brain Does This

The mechanism is not stupidity. It is a set of ordinary cognitive processes doing exactly what they evolved to do, pointed at an input they were never designed for.

You are the most available data set you own. When you read "you sometimes struggle to switch off at night," your memory does not run a fair audit of every night of your life. It surfaces the three nights you lay awake. The claim gets tested against a sample the claim itself selected.

Confirmation bias only has to search, not weigh. Finding one matching memory feels like verification. Finding forty non-matching ones requires effort nobody spontaneously does.

Self-relevance is a spotlight. Information framed as being about you gets processed more deeply and remembered better than the same information framed generally — a robust effect in memory research. "People with high cortisol feel wired but tired" lands differently than "you feel wired but tired."

And the fit itself feels like evidence. Marks and Kammann called this subjective validation: the sensation of recognition is generated inside you, but arrives labeled as though it came from outside. That labeling error is the entire business model of a great deal of what gets sold as personalized health.

Five Signatures, Borrowed From a Government Report

The AIR reviewers were not writing about supplements. But their five observations about remote viewing reports are the cleanest diagnostic checklist for Barnum-grade wellness advice I have found. Hold any health claim up against them.

1. It is broad and general. "Your gut is inflamed." "Your nervous system is dysregulated." "You have a slow metabolic type." True of nearly everyone, or unfalsifiable, or both.
2. It is not consistent across independent sources. Three practitioners, three different root causes, each delivered with equal confidence. In the remote viewing program, three viewers describing the same site disagreed with each other — which is a fact about the method, not about the site.
3. It lacks specifics that match known facts of your case. It never says anything checkable that it could have gotten wrong.
4. It contains a large amount of irrelevant material. Volume reads as thoroughness. A forty-page report and a nine-hundred-marker panel feel rigorous in a way that a single well-chosen blood test does not, regardless of which one changes your decision.
5. The burden of interpretation falls on you. This is the tell. If you have to work to make it fit, you are the analyst forcing the hit.

Where This Shows Up If You Are Not Looking For It

Horoscopes are the easy target, and almost nobody reading this is being harmed by horoscopes. The expensive versions are better dressed.

Quiz-gated supplement funnels that ask nine questions and return a "protocol formulated for your profile." Body-type and metabolic-type systems with no reliable classification behind them. Diagnostic labels that sound clinical and aren't — adrenal fatigue is the durable example, endorsed by no endocrine society and rejected in a systematic review of 58 studies, yet it survives because its symptom list (tired, foggy, crave salt, hard to get going) describes essentially all of modern adult life. Wearable "readiness" scores presented as insight when the underlying signal is thin. And increasingly, AI health chat, which is unusually fluent at producing confident, personalized-sounding, entirely generic prose.

The common structure: a description broad enough to always fit, a delivery specific enough to feel bespoke, and a product at the end.

What Actually Passes

I want to be careful here, because "everything general is a scam" is its own kind of lazy thinking. Some of the most important health advice in existence is extremely broad: sleep enough, move most days, eat mostly whole food, stay connected to people. Those are general and they are true.

The difference is not breadth. It is falsifiability and the direction of the burden. "Regular resistance training preserves muscle mass with age" is a general claim that has been tested, could have failed, and tells you precisely what to do. "Your energy is blocked" is a general claim that cannot fail and tells you only to buy something.

So the practical move is not more skepticism about everything. It is redirecting the skepticism at the moment where you feel most seen. That feeling of how did they know is not a reason to trust a source. Under laboratory conditions, it is what generic text produces on purpose.

Two places I put this into practice. When I buy a supplement, I am buying verification rather than a story about my type — a [AFFILIATE:THORNE] product earns its price because there is third-party testing and a published certificate of analysis behind the label, which is a checkable, falsifiable claim in a category that mostly sells unfalsifiable ones. And the underlying skill — catching your own mind in the act of supplying the fit — is metacognition, which is trainable. Ten minutes of noticing thoughts arrive without automatically believing them, whether through [AFFILIATE:HEADSPACE] or a free timer and a chair, is closer to intellectual self-defense than most people expect.

Twenty years, an unambiguous statistical anomaly, and a stack of reports about water being nearby. The program did not fail because the people in it were foolish. It failed because vagueness is comfortable, interpretation is generous, and the human mind will always rather force a hit than sit with a miss. Your feed knows this. It has known it since 1948.

Related Reading

• Why Health Experts Disagree — And How to Read the Science Yourself — the filters I run on any study
• The Science of Altered States of Consciousness — the other half of this literature, taken seriously
• The Science of Inner Speech — the voice that supplies the interpretation
• The Social Comparison Trap — why the feed is optimized to feel personal
• Conscientiousness and Longevity — what personality research looks like when it's done properly
• Why Forgetting Is a Normal Function of Memory — how selective recall actually works