My friend @ChrisHeery's tweet struck a chord a while back. Most intros to Bayesian stats use similar examples which implicitly assume doctors order tests using uninformative priors and no subjective information.
Maybe that's bad pedagogy?
...I've never been tested for Ebola.https://twitter.com/ChrisHeery/status/1048963583516631040 …
-
-
Replying to @generativist
Yes, it belittles the whole principle of only ordering tests when probability of the disease entity is high.
1 reply 1 retweet 1 like -
Replying to @ChrisHeery
Yea, and considering I am, 1) Statistically well-trained; and, 2) Unfortunately familiar with medical testing, it's really weird that I never read any of the articles like the one you quoted and thought, "um, no, you're ignoring a pretty import condition."
1 reply 0 retweets 1 like -
Replying to @generativist
Hey man, that’s my job! I take offense to someone ignoring such an important aspect of medical decision making. It’s as though the term differential diagnosis has never entered the author’s mind.
2 replies 0 retweets 1 like -
Replying to @ChrisHeery @generativist
The best way to know an example is poor is to use basic common sense. This example is clearly far from meeting that threshold.
1 reply 0 retweets 1 like -
Replying to @ChrisHeery
The one you cited is particularly pathological. Doctors order tests for patients; observe the results; and, follow the cases all the time. It's kinda what they do. They don't need to break out Bayes -- their estimates are experiential.
1 reply 0 retweets 1 like
Thanks for tweeting this weeks ago :)
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.