I read the British science journal Nature. In such publications one will often see the statistics associated with a particular experimental result. What this means, imho, is the odds of that result happening by chance alone. Generally this is expressed as P < something, typically a very small number like 0.0001, or one chance in ten thousand this result occurred by accident.

Something is often considered “proven” if P < .05, which is one chance in 20.

Something is often considered “not-proven” if P >= .05. Clearly P = 0.5 is a coin-flip and would not impress any of us. However, suppose P = 0.051? This is what I mean by in between, neither proven nor disproven.

Anti-proof would require an experiment testing for the converse of the original experiment. If in this case we had P < 0.05 we might consider the converse statement proven, or the original statement to be proven false.

Disproof is a bit trickier. In mathematics, a general statement is considered disproved by showing a single counterexample. “All crows are black” is disproven if one crow in seven million has a white feather somewhere. In science, P = 1/7,000,000 would be considered convincing.

Apparently it is even harder to prove that a new medication is better than an existing one. Sadly I no longer have the article I saw this in, I think a psychiatric journal, with a title something about “unicorns are possible”, hinting at the difficulty of finding a result of this nature.

Something to think about when you are using FDA-approved medicines. Is it better? What were the odds in the clinical trial? What are the odds of cure by psychiatry as opposed to by drugs? Do shaman practices work? Are there any examples of faith healing? If P was only 0.051, was the result discounted?

We are told, with confidence, that additives, drugs, procedures, diets, operations, are “proven effective”. Perhaps, if the stakes are high, we should find out what the odds really are.