Quote:
Originally Posted by Melkerson
I just added the repeat test on the HIV, because that is a fairly routine course of action. You can ignore that if you want. Even if there is reason for you to get an HIV test, we can still find reason for you to get an unnecessary one (e.g. you just had one a couple of days ago). But never mind, as I said it's a bad example.
Nah. I'm good. Immediately after the test, I do the right thing and tell my wife who freaks out and stabs me for cheating on her. I get the second test in the morgue.
Quote:
I did see your point, but that's why I keep trying to focus on what you would do, or what I would do. Surely we can predict our own behavior better than any test? But I guess you will continue to hide behind "Well, we can never really predict what anyone will do, can we?"
I'm not hiding behind it. The hot-cold empathy gap is a fact of human existence.
Even when it is not in play, we are **** at predicting our own behavior. The nutshell version is that we are incredibly bad at bayesian analysis. How many logs have authors who believe that they will not be in the 95% (a made up statistic, but probably close to accurate) who will absolutely fail at their simple fitness goals?
Quote:
So, lets move to a different example. We're going to give you an unnecessary test for color blindness. For the record, there are color blindness tests which have a reported zero percent false positive rate. So what risk are you at if we give you an unnecessary color blindness test?
Why go with something difficult? A much better example that "proves" the uninteresting case is whether we should give a battery of blood tests to a brain-dead individual and then tell them the results.
Wait. Your clever point is that there are uninteresting and contrived cases in which it isn't 100% true beyond a doubt that there could possibly exist a test for which no one could be at risk? We already know that. Literally no one doubts that.
I could have given quite a few better examples yesterday. I could have tested every human being for their toenail growth rate and then used an extremely obscure Latin name to obfuscate the results and to not give a reference range along with the test, or put a gold star and smiley face next to the results. Or tested brain-dead people. Or prisoners who are perfectly immobilized.
Now that we have that out of the way, what does this have to do with testing non-symptomatic adult males for testosterone level? You don't actually believe that you can extrapolate anything useful about testosterone testing from this discussion, right? By "anything" I literally mean "anything at all, even with the wildest stretch of the imagination, that is relevant in even the remotest sense including things that do nothing more than rhyme with 'testosterone', 'adult' or 'male' that is actually true."