Customer Reviews

Computer-Intensive Methods for Testing Hypotheses: An Introduction

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

 

 

The text is only 92 pages. There are 102 pages of basic programs, 25 pages of statistical tables and 4 pages of references. Some of the references are interesting. The advantage of this book is that it is mostly written on a non-technical level, although some treatment of the bootstrap is a little technical and awkwardly written. It also is nice that it introduces Monte Carlo before the bootstrap so that the bootstrap (actually the Monte Carlo approximation to the bootstrap) can be seen as a special case of Monte Carlo testing of hypotheses. Among the disadvantages are (1) it deals solely with hypothesis testing, many readers may be interested in other applications such as subset selection, bias correction and confidence interval estimation, (2) written in 1989 it misses the many advances from 1989 to 2000 and even avoids some of the more complex but possibly better approaches that were available in 1989, and (3) it lacks depth. Readers interested in more depth without loss of good exposition can turn to Manly (1997) Randomization, Bootstrap and Monte Carlo Methods in Biology. Well written books dealing strictly with permutation methods of hypothesis testing are Good (2000) Permutation Tests A Practical Guide to Resampling Methods for Tesing Hypotheses 2nd Edition and Edgington (1995), Randomization Tests 3rd Edition. Excellent books devoted to bootstrap are Efron and Tibshirani (1993) An Introduction to the Bootstrap, Davison and Hinkley (1997) Bootstrap Methods and their Applications and Chernick (1999) Bootstrap Methods: A Practitioner's Guide.

This is a great foundation book. Separate chapters on approximate randomization, Monte Carlo sampling and bootstrap resampling. Good discussion of strengths and weaknesses of each. Very readable, well paced, good appendices with additional technical detail.

I need to find a way to do this without using Pascal, Fortran or Basic. Surely SPSS, SAS or Excel support approximate randomization, but this book was written too long ago to mention them. It's a great book for the basic theory, but don't count on it to help you actually do the work. Dr. Noreen: please write an update!