PSYC 406 Lecture Notes - Lecture 19: Type I And Type Ii Errors, False Positives And False Negatives, Jargon
Document Summary
Everything in psychology (at least the stuff we care about) is fairly rare. Test for a disease, (cid:498)positive(cid:499) means you have the disease, (cid:498)negative(cid:499) means you don"t have the disease. If a test is 80% accurate, there are four possible outcomes For every 100 healthy people, the test will say. 20 are diseased (false positive; type i error) For every 100 diseased people, the test will say. 20 are healthy (false negative; type ii error) Therefore, test is 80% correct whether you are healthy or have the disease. But, that"s assuming the base rates are the same (cid:523)(cid:883)(cid:882)(cid:882) healthy, (cid:883)(cid:882)(cid:882) diseased(cid:524) Test can also assess things are not disease is (e. g. , who is likely to go to graduate school) Example for an 80% accurate test (which is very good) Most tests are way less than 80% accurate. Even if we made the test incredible 99% accurate: But no test is 99% accurate (lower reliability and validity)