|New Reviews| |Software Methodologies| |Popular Science| |AI/Machine Learning| |Programming| |Java| |Linux/Open Source| |XML| |Software Tools| |Other| |Web| |Tutorials| |All By Date| |All By Title| |Resources| |About| |
Keywords: Statistics, Data analysis, Mathematics Title: Statistics In A Nutshell Author: Sarah Boslaugh and Paul Watters Publisher: O'Reilly ISBN: 0596510497 Media: Book Level: Introductory/Intermediate Verdict: Useful, but not a beginner's textbook |
Billed as a desktop quick reference on statistics, this is a book that is geared to developing statistical reasoning rather than simply listing statistical techniques or tables of figures. Aimed at students taking introductory stats courses, adults learning stats in a work-environment or those who are just interested in learning more about the subject. That's a fairly wide remit, and as we'll see it succeeds more in some areas than others.
The book is organised into four sections: introductory (which starts with basic concepts of measurement, introductory probability, descriptive statistics, research design, critiquing statistical analyses and data management); basic inferential statistics (introduction to sampling and hypothesis testing, the t-test, correlation, categorical data and non-parametric statistics); advanced techniques (the general linear model, analysis of variance, multiple linear and other regression techniques, principal components and other advanced techniques); and finally a section that looks at statistical analysis in different domains (business and quality improvement, medicine and epidemiology, education and psychology). The book rounds off with a series of appendices, including a chapter that looks at the main statistical software tools - from SPSS to SAS, R and even Excel.
While there's no disputing that this is an interesting book (for those who find the subject interesting at least), how well it succeeds in meeting the needs of its intended audience is not so clear. For those approaching statistics for the first time the book works well in the first few chapters but it's not really a text book that's suited for home study. It works better as a supporting text to complement a first course in statistics. What it lacks in basic explanations and extensive worked examples it adds in looking at topics such as data management and research design that most introductory texts will skip or not devote enough time to.
For those readers with some existing background (no matter how rusty), the book provides more value. It's a quick refresher course with plenty of discussion over and above step-by-step algorithmic instructions. It reminds the reader, for example, that statistical significance is not the same as real-world significance, and provides pointers to calculations of power and other ways of attaching real-world importance to results that are significant.
The book also suffers from some poor proofing, and readers are advised to check the O'Reilly site for the list of errata (on the other hand picking up mistakes in the text is a good test of how much you're paying attention!).
Overall, this is a useful read for those who are statistically inclined, it's not such a good book for those learning statistics for the first time.