Customer Reviews

Statistical Methods for Rates & Proportions

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

 

 

This book is an excellent alternative to Agresti's "Categorical Data Analysis" for the analyzing of categorical data. The book is comprehensive and less mathematically demanding than Agresti. The book does a tremendous amount with contingency tables and epidemiologic types of data and has many good worked out examples.

There are a couple of features that could be improved in the next edition:
1) Give the statistical tests that are presented names.
Often statisitical procedures/tests are given and they are not given an explicit name. I think that this is an intentional practise of the authors, but I found it very confusing. For instance, there are many chi-square tests out there, each of which can be written in different ways. Why not call them "Pearson's chi-square", "Cochrane's chi-square", etc... so that the reader knows what the authors are talking about?

2) Not enough time devoted to the subject of combining simple proportions (i.e. not fourfold tables) from several samples/studies. Only one method is shown when several others exist.

3) ALL the common methods used to transform proportions should be presented and explained. For instance, the arcsine transformation is left for an exercise.

4) The writing was a bit terse and cryptic at times. I wish the authors had opted for a simpler English sentence structure and aimed to write at as low a level as practically possible. This encourages understanding of the topic rather than the development of eloquent prose.

5) The references were good but the way they were cited was not. In line author names is an outdated reference method that bogs down sentences and clutters pages with useless text. Simple numeric citation markers would have been better.

6) Bayesian methods are developed, but not a much as I might otherwise have hoped.