You do not currently have permission to view the full text for this document. Please sign in to an account with an active subscription or contact support.
This book is a serious and thoughtful attempt by Prof. Meehl to consider impartially the pros and cons of an argument which has often been put in a garbled and barbed form through ignorance, intolerance, and partisanship. The argument concerns clinical predictions, and may be stated thus: Who is likely to make better predictions—the clinician who knows his patients through close and personal contact, or the statistician who has only actuarial data at his disposal, but knows how to arrange them in a mathematical formula that will provide a predictive answer?
Few people are qualified to consider both sides of this problem. Prof. Meehl, who describes himself as a 'hybrid working clinician and a rat psychologist', has the necessary qualifications, knowledge, and experience. Unfortunately, he seems to have lacked the time to tidy up the presentation of his thesis in this book; he has merely expanded a course of lectures. As a result, his book is discursive, repetitive, and often difficult to read. It is likely to scare away many would-be readers who would profit by a knowledge of his thesis.
I shall therefore try to give a short account of the essence of that thesis.
To begin with, we can postulate that the statistician and the clinician are to start with the same set of data, derived from case histories, interviews, test scores, and so on. They are, of course, free to disregard some data (i.e. to give them zero statistical weighting).
How then does the statistician manipulate his data to obtain a predictive answer? The data can be classified in various ways. For each classification actuarial tables can be constructed which will contain the frequencies with which different predictive answers occur. Finally, a statistical formula can be evolved which will reflect the contents of the actuarial tables. When all this has been done, the most likely prediction can be mechanically computed, with ease and speed, by any trained clerk. To give a concrete example, let us co