Talk:Screening test fallacy
This page's example is vague and poorly written... I drew a conditional probability tree, having just covered this subject in Statistics a little while ago, and could not solve the problem. It should be told in the article whether Ralph and/or his sister have the disease. Otherwise, what is the point of showing the probability and a bunch of numbers, when the problem can't be solved anyway? Or, they should refer to the formula for conditional probability, P(A | B) and all of that stuff. For example, why would people "succumb" to the belief that Ralph has an 80% chance of being sick? The only time 80% appears in the article is the chance of a healthy person testing negative for the disease, but Ralph tests positive... I am genuinely confused, and I think some math or further explaining should accompany this example. LockeShocke 23:30, Feb 8, 2005 (UTC)
Accuracy - the original author's intent is not clear
[edit]It's not obvious from the text of the article, what kind of misuse has author in mind. There are 2 possibilities:
- False positives like that in bayesian inference article. This is not a misuse in the correct sense of the world, but rather a property of bayesian interpretation of probabilities, which cannot accurately convey how much information you have. If this is the case, then there should be explained this connection and also better example with apriori probabilities given.
- Or the author just wanted to point out forgetting of prior probability or equating conditional probability with unconditional one, as would lines like
- Many people succumb to the belief that Ralph has an 80% chance of having the disease and his sister has a 10% chance of having the disease. Some people may decide that Ralph has 81.8% chance of having the disease and his sister has 11.1% chance of having the disease.
- suggest.
So unless this is decided, it should be considered inaccurate. Samohyl Jan 20:06, 11 Feb 2005 (UTC)