3 ms·
This article is the perfect example of why our higher education system is so dysfunctional: "The obvious initial inference to make about a test that does not p
by rkon 15y ago
This article is the perfect example of why our higher education system is so dysfunctional:
"The obvious initial inference to make about a test that does not pick up a difference where you expect one is that it is not a very good test."
Or maybe the expectation is incorrect because you're a biased, myopic dimwit? Sorry, former Ivy League professor, but you just failed miserably in 'Logic and the Scientific Method 101'. How ironic, considering the fact that he was trying to counter an argument about universities' failure to impart critical thinking skills.
- ahh 15y agoConsider prior probabilities and Bayes' rule.
- kenjackson 15y agoActually I think the professor is right -- that should be your initial inference if you expect it to pick up a difference and it doesn't. For example, if I make a test to tell me which item weighs more between two, and then I weigh myself and my refrigerator and it says I weigh more, then I should look at the test first. Now it may turn out that the test was right, and my fridge actually is a new lightweight fridge, but your expectation is the default hypothesis. In fact the way you typically calibrate tests is to verify that it signals differences that you very strongly expect.