| Abstract: | Responds to commentary by I. Silverman (see record 2007-08963-001) entitled "Review of Pure types are rare": Comment on book review. I do not think that either one of us will change his position, and therefore it is up to the interested (yet disinterested) reader to adjudicate our dispute by checking Professor Silverman's book and his references. In the process, the reader will be forced to consider the thought-provoking implications for our mental health system of some of the incidents which the author describes so vividly, and that will not be a bad thing. There is one point, however, which is worth pursuing further here, for Professor Silverman persists in an elementary statistical fallacy. Let us take the situation which he cites, where the base rate for diagnosing schizophrenia is 50%. Suppose with the same base rate the degree of agreement is in fact 53%; this is far above chance level. It must be emphasized, however, that the 53% agreement under discussion came not from a study where the base rate for diagnosing schizophrenia was 50%, but from one where the conditions were far more stringent since the base rate for diagnosing schizophrenia was around 20%. (PsycINFO Database Record (c) 2010 APA, all rights reserved) |
