Peter Donnelly on how stats fool juries

25 Oct, 2013 at 21:46 | Posted in Statistics & Econometrics | 2 Comments



  1. nice articles

  2. I have some sympathy with those who think that ‘99% accurate’ means that there is a 1% chance that the test result is wrong. To put it another way, suppose that a doctor had the choice of giving you P(disease | test result) or P(test result | disease ), with the former being of direct relevance to you and the latter being of no use (without some other information), and he does not make it clear which he is giving you. Is it unreasonable to assume that he has given you the relevant data?

    Similar for experts and juries, or experts and ‘decision-makers’.

Sorry, the comment form is closed at this time.

Blog at
Entries and Comments feeds.