Customer Reviews

Nonparametric Statistics for The Behavioral Sciences

5 star:		(4)
4 star:		(1)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

Speaking as an MPH level student, (i.e. not a real mathematician) this is about the only usable book on nonparametric stats I have encountered, so I ended up buying it despite the rather high price for a not terribly large book. But, as happens frequently in healthcare and social sciences, when faced with data that can't be analyzed with the normal mean and standard deviation stuff (i.e. survey answers, etc.) this book offers a lot of possibilities beyond the standard chi square test, and more importantly, is clear about what test is appropriate, and how to apply it.

In the 1960s Siegel's book was the most popular and the most often cited. This is because except for Fraser it was the only useful test available to researchers. The book was written in a somewhat non-technical manner in order to be accessible to social scientists. At the time it became the standard book for all researchers. Theoretical books such as Hajek and Sidak's "Rank Tests" Came out at the end of the decade and the other good statistical books such as Hollander and Wolfe; Conover,; Lehmann; and Randles and Wolfe didn't come out until the 1970s.

So Siegel's book has historical significance but now the pratitioner and the theorists have many other good books to choose from. The text has been revised many times presumably to keep up with the research advances that have practical use for social scientists.

This is an excellent first book for nonparametric statistical methods. It is a cookbook, but is a good introduction to the many nonparametric techniques for assessing data. These are oftentimes much better suited for your data than the standard stuff you get in intro to statistics. The book by David J. Sheskin or by Conover should your next book.

For a non math major (or stats major) user, this book offers an easy way to have works done quickly. But be cautious, an first-class cookbook does not necessarily yeild a first-class meal on your table.

This is (together with Empirical Methods for Artificial Intelligence by Paul R. Cohen) the best of the statistics books I read.