📜The ASA Statement on p-Values

A 🔤 p-value is a statement about data under an assumed model: how surprising would results like these be if the model, usually no effect, were true. It is not the probability that the hypothesis is true, not the probability that the result is a fluke, and not a measure of how big or how important an effect is. Decisions should never hang on whether it falls on one side of a fixed line, and it means little without an account of everything else that was tried.

Readings. Five pages, four sittings:

• 10 May. Read twice. Copied the six principles out in my own words.
• 11 May. Third reading. Wrote the literature note.
• 9 June. Whit Monday revision. Recited the six on the estuary loop. Got five, lost number four both times.
• 1 October. Back to it for the autumn reading response, the one about what a p-value is not. Used principles two and five. It took an hour, because the notes from May could be read without the paper.

• Read twice on Saturday 10 May, once at Lantern Coffee and once on the sofa with a pencil, and again on Sunday. A month before the final of 00. Statistics for Product Decisions I Hub.
• Five pages. A professional body of statisticians telling everyone else, formally and for the first time on a point of practice, that the commonest number in science is being misused.
• Six principles. I copied them out by hand in my own words, the way you copy a recipe you intend to actually cook.
• It is a committee document and reads like one: every sentence has been negotiated. That is its strength. This is the minimum that a room of people who disagree about everything else would all sign.
• A week after Ioannidis. He supplies the alarm, and this supplies the wording.