Statistics Concepts and Controversies |
||
(Illus.) W.H. Freeman & Co. 1979 350pp. 0-7167-1717-4 (paper) |
The author's stated purpose for this text is to present statistics to nonmathematical readers as an aid to clear thinking in personal and professional life. The suggested use of this book is as a text in a general education course for students in humanities, social or behavioral science, or in education. An alternative would be as a supplement to a more traditional statistics course. In either of these settings, the book could have tremendous value.
If you are interested in a statistics book that is written for readers concerned with ideas rather than techniques, then you should give this one serious consideration. However, if you are interested in such topics as error analysis, propagation of uncertainties, least squares fitting and so forth, then skip this one. As an example of the book's mathematical level, we are on page 172 before we are introduced to the mean, median, and mode.
Many of the author's examples deal with the use and misuse of statistics in advertising and in political rhetoric. An interesting example of an incorrect method of presenting data concerns a bar graph used in an advertisement for Time magazine. The bar graph is presented such that it leaves the impression that Time has a four-to-one lead over Newsweek in advertising sales, while actually its lead is somewhat less than two-to-one. A second example of poor presentation of data is that of a graph used by the Associated Press to show how local and property taxes have varied over the years. The graph is presented such that its time axis is vertical while dollar amounts are horizontal. As a result, where the graph is flat, it implies that taxes were not increasing, while actually they were thundering upward.
One would hope that the readers would resolve never to commit the kinds of errors and abuses presented in this book. However, I must admit that I could not escape the mental picture of some Madison Avenue executive reading this book and thinking to himself, "Hey, there is a good misuse of statistics and data that I had not thought of before."
The basic topic outline is: (a) collecting data, including sampling, experiment design, and measurement, (b) organizing data, including tables, graphs, and descriptive statistics, and (c) drawing conclusions from data, which include the study of randomness and a discussion of formal statistical reasoning. Each chapter ends with a list of suggested readings and with a set of excellent exercises. However, it should be repeated here that the approach taken in all of the above is directed more toward the social and behavioral sciences than to the discipline of physics.
The book is interesting to read if you enjoy such pieces of trivia as the statistical proof that the numbers racket has a higher payout than does a typical state lottery. However, I doubt if there are many teachers of physics who will find use for this book beyond the "good-read" level.
--Reviewed by Jerry Faughn in The Physics Teacher, 24 (October 1986), pp. 444-445.