This is not a vent board or any other kind of therapy. Before you hit the POST button, ask yourself if your contribution will add to the level of discussion going on.
Important notes on articles:
- Please do not copy entire articles into your post; rather, provide links to them.. We are now links-only for ALL Internet publications. If only a small portion of the article pertains to your post, Fair Use allows you to copy those one or two paragraphs, provided you cite the author's name and the publication for which he writes. Otherwise, put a link in the HTTP Link box.
- Even if you're copying a reference to an article, provide a link to the page from which the article came. We're trying to cut down on duplicate topics, and the posting process will check the link to your article to see if it's already being discussed on this board. At the very least, you'll save yourself some grief on the boards.
- If your first reaction after reading the article you're going to share is the author is uninformed / stupid / a jerk / all of the above, it's not worth sharing with anyone. Not every article needs to be discussed. The more the hair-pulling articles are discussed (e.g. ESPN Page 2), the more the authors will write hair-pulling articles.
Post being replied to
It’s not really an issue of test accuracy by carroll2005
It’s an issue of people not understanding statistics, prevalence, and pre-test probability.
Basically - the prevalence in most communities makes pre-test probability extremely low, so even a test with 99% sensitivity/specificity will end up having a false positive nearly as often as a true positive.
Extreme example of the statistics concept: you have a pregnancy test that is 99.9% accurate. But you are administering it to a man. Pre-test probability of a true positive test is 0, so it doesn’t matter how accurate your test is.