Fundamentals of Behavioral Statistics [Hardcover]

In each chapter topics are concisely layed out, followed by numerical examples when applicable. In the final section of the chapters, all discussed issues are put together in practical terms and are made relevant primarily to psychological research, by, for example, showing how the findings of a given test can be reported in a paper.

Being a text-book, each chapter has a number of exercises for the student to solve, and a rather detailed (though selective) answers section. I found the presentation of the definitions of key terms on the margins of each chapter rather handy, though a glossary in the end of the book would have been welcomed. Finally, the book has one of the shortest reference sections I have seen, with only 47 references, half of which are not directly related to statistics.

The text could be roughly divided into two parts: Following conventional wisdom, in the first part (chapters: 1 to 10) the basic concepts are introduced such as the scope of statistics, the notion of probability, measures of central tendency and variability, the normal curve and the z-scores. Welcome inclusions are chapter 3, which deals with the concept and the applications of exploratory data analysis (EDA), by sticking closely to Tukey’s ideas, and chapter 7, which is solemnly devoted to briefly presenting and explaining the most commonly used graphs and tables, and showing guidelines for their proper use. The second part (chapters 11 to 18) deals with inferential statistics. There is a brief introduction to statistical power and effect size (chapter 13), and a fair implementation and explanation of those concepts within the context of a few but not all of the covered statistical techniques.

However, in a rather unjustified move, the authors present the concepts of correlation and regression (chapters 8 & 9) before the introduction of probability and hypothesis testing. As a result, the estimation of the statistical importance of the beta coefficients, for example, is only discussed in reference to the visual inspection of their confidence intervals, since at that stage the concepts of the alpha-level and the t-test have not been introduced yet.

As far as inferential statistics are concerned, the book is heavily orientated towards the t-test and mainly the ANOVA designs (together occupying five chapters and more that a fourth of the text). That said, the only multiple comparison discussed is Tukey’s honestly significant difference, while for any other procedure the reader is referred to an external source. In another example, tests for homogeneity of variance / covariance are absent, with only a single mention of the Fmax statistic.

The authors also chose to play down the importance of nonparametric techniques, by only devoting two short chapters (all in all 26 pages) to the analysis of nominal and ordinal data. That said, the inclusion of a brief presentation of Cohen’s Kappa coefficient in chapter 17 and McNemar’s change test in appendix C are a surprise, especially given that perhaps more commonly used nonparametric techniques, like Kendall’s tau or the Kruskal-Wallis ranks test, are absent from the text.

In conclusion, interested parties should weigh the book’s clear and accessible arrangement against its less than satisfactory handling and coverage of certain subjects before making an adoption decision.