Talk:Information bias (psychology)

Nonwiki
Reads like an op ed. Seems unencyclopediac. Mathiastck 19:27, 25 April 2006 (UTC)

Agree with the previous comment, also other things for consideration:

the test is not useless as you can eliminate one of the possible diseases (even if it won't change the best treatment at the moment).

The test could be useful if you run it multiple times IF the test is independent of previous trials (i.e. if the person has globemia and each time the ET test is done there's a 50% chance of it being positive or negative, then by running the ET test n times all coming out the same way, then there's only a 1/(2^n) chance of it being globemia. If the probability is not independent (i.e. depends on some other factor) then this does not hold.  Biomede 19:13, 4 June 2007 (UTC)


 * Is it possible that the case is just badly explained? The way I read it, the ET scan is not absolutely worthless (and I realize that the article is written to suggest that it is worthless).  It is just that the most cost-efficient and expedient course of action is to first administer the treatment for globemia.  You have 80% of being right and stop the medical expenses right there.  I'm assuming there are no serious side-effects.  (Yes, I'm assuming.  I think we all agree that the current article is quite sketchy.)  If you find that the globemia treatment does not work, then you can run the ET scan to decide between the two other diseases.  Although it is probably still more cost-effective to just try another treatment.  If you run one ET scan before doing any treatment, you gain no new information.  It is true that running it multiple times would help decide what the disease is but then the article mentions the high cost of the ET scan. If somebody who has access to the original case would rewrite this article, that would be nice.  --Kartavyam (talk) 14:00, 22 February 2008 (UTC)


 * Whether encyclopedic or not, lines like "– what is not worth knowing is not worth knowing." & "Information that is not worth knowing is not worth knowing." is just bad writing. That such a thing is used twice is all the more to note of how this article needs to be re-written. Nagelfar 13:09, 27 July 2007 (UTC)

Quality of example
It seems this page has improved somewhat since the above comments were written, but this still doesn't seem like a great example to me. It's quite possible that the test subjects simply weren't equipped to work out the probabilities in each case, and therefore guessed intuitively that the test might as well be administered. A simpler and purer example of the effect would be nice. calr (talk) 03:36, 11 September 2010 (UTC)

The example given isn't completely correct although it has the right outcome. If you were to just take one individual they would have an 80% chance of having globoma and a 20% chance of having popitis or flapemia. The odds of having popitis or flapemia aren't given so let's say that it's 50/50 for 10% each. If you do the ET scan it can't rule out globoma but it can rule out popitis or flapemia. Once you rule out popitis or flapemia it increases the chance of it being globoma by 10% making it 90% likely that they have globoma and 10% likely that they would have popitis or flapemia. This of course leads to the same outcome that the test is wasteful and actually only makes globoma a more likely diagnosis. The only logical course of action in this example is to administer the globoma treatment and if that doesn't help do the ET scan to determine whether it's popitis or flapemia. EAbyss (talk) 02:33, 3 November 2013 (UTC)