[bdzr]

> I agree. Expecting perfection from humans, even experts, is not reasonable and is frankly counterproductive.

There's a big difference between perfection and "Statistical Literacy Among Doctors Now Lower Than Chance"[1]. I don't think their intentions are bad, but they are woefully incompetent at many basic things.

[1] https://slatestarcodex.com/2013/12/17/statistical-literacy-a...

[bonsai_spool]

> There's a big difference between perfection and "Statistical Literacy Among Doctors Now Lower Than Chance"[1]. I don't think their intentions are bad, but they are woefully incompetent at many basic things.

As it happens, the daily practice of medicine does not require interpretation of p-values. Indeed, medicine existed before the p-value.

The people who create studies that ultimately guide policy decisions are specialized (much like people who write GPU drivers are different from those who run inference)

[bdzr]

> As it happens, the daily practice of medicine does not require interpretation of p-values. Indeed, medicine existed before the p-value.

What are you talking about? Doctors refer people based on test results every single day. From what I've seen, hardly any of them understand the precision/recall of the tests that they then use to refer you (or not) to screening procedures (which are not all harmless).