Personal experience proves what happened to one person, once. Evidence is a pattern checked across enough cases that chance and individual circumstance have been ruled out. Experience is always a legitimate account of your own life; it becomes evidence only once similar accounts have been gathered, compared and checked for the ordinary ways one person’s story can mislead.
Flow diagram of 4 steps: Read the claim back with "for me" attached, Check whether the claim describes a rate, a typical case, or a cause, Ask whether anyone with the opposite result would be visible to you, Say plainly which one you have.
Personal experience is a report: this happened, to me, in these circumstances. It requires nothing beyond honesty to be true. Evidence is a claim of a different shape entirely: this tends to happen, across enough different people and circumstances that the pattern is unlikely to be chance, coincidence, or something particular to one case. Both can be completely honest. They are not interchangeable, because they answer different questions — "what happened to you" and "what generally happens" are not the same question, even when the same story is used to answer both.
The confusion rarely comes from someone deliberately passing one off as the other. It comes from a true, specific account being read — by the teller or the audience — as if it had already done the work that only a wider pattern can do.
There is a real category of claim where one person’s experience is not just adequate evidence, it is the only evidence that could exist. Nobody needs a study to tell them how something felt, what they preferred, or what they noticed about their own reaction. Experience is the primary source for exactly those questions.
What a personal account can and cannot settle on its own
| Claim | Can experience alone support it? | Why |
|---|---|---|
| "This made me feel calmer" | Yes | You are the only witness to your own state, and no outside check is more authoritative |
| "This is what I noticed changed for me" | Yes | A first-person observation about your own life needs no corroboration to be true |
| "This will calm most people down" | No | A claim about most people requires cases beyond the one you have |
| "This caused the change, and nothing else did" | No | Ruling out other causes needs comparison, which one account cannot provide |
A single experience is missing two things that evidence exists specifically to supply: a comparison, and a check against coincidence.
If you felt better after trying something, you cannot know from that alone whether you would also have felt better without it — time passing, an unrelated change, or simple variation in a bad patch can produce the same result. Evidence answers this by comparing outcomes with and without the thing in question, across enough cases that the difference is unlikely to be noise.
People who try something and it does nothing, or makes things worse, are far less likely to write about it than people for whom it worked dramatically. A single striking account, even a completely honest one, is not a random sample of what tends to happen — it is disproportionately likely to be a case where things went unusually well, simply because that is the kind of case people share.
Before a personal account is used to support a claim, it is worth checking which of the two questions it is actually answering.
If the sentence still means the same thing with "for me" added at the end, it is an experience claim and needs nothing further to be honest.
Words like "usually", "most people", "causes" or "proves" signal a claim that needs more than one account to support.
If people for whom this did not work would likely never mention it, your account is probably not representative of what tends to happen, whatever it shows about your own case.
A specific, honest experience is worth stating as exactly that — it does not need to borrow the authority of evidence to be worth telling.
It is tempting to treat evidence as automatically superior to experience, but they answer different questions and a wider pattern cannot tell you what one specific case felt like from the inside. Evidence about a general tendency does not erase a genuine exception, and a personal account does not become less true just because it does not generalise.
The failure worth avoiding runs in both directions: using one account to make a claim only a pattern could support, and dismissing a real personal account because it is "just anecdotal" when nobody was claiming it was anything more. Precision about which kind of claim is on the table resolves most disagreements that look, on the surface, like disagreements about facts.
No — an anecdote is a true, useful account of one case, and it can be the first sign something is worth investigating further. It stops being useful the moment it is treated as having already answered a question that only a wider pattern could answer, such as how common or reliable an effect is.
There is no fixed number — what matters is whether the cases were gathered in a way that avoids only hearing from people with a particular outcome, and whether they were compared against what happens without the thing in question. A large but self-selected group of accounts can still fall short of being evidence; a smaller, carefully compared set can meet the bar.
For questions about what happened to you specifically — how something felt, what you noticed, what mattered to you — your experience is the authority and no study changes that. For questions about what generally happens or what causes what, a well-run study addresses a limitation, the missing comparison, that personal experience cannot supply on its own.
Evidence is information that makes a claim more or less likely to be true; proof is a much higher bar, typically reserved for logic and mathematics, where a conclusion follows with certainty. Most claims about people and outcomes never reach "proof" — the honest goal is usually strong evidence, not certainty.