Methodological texts fascinate me. I am always intrigued by the decisions authors make for presenting content that some would mistakenly perceive to be straightforward, such as the extent to which the text (a) takes a conceptual versus technical/procedural approach, (b) is meant to be read as a course textbook versus a reference for independent researchers, and (c) covers basic versus more advanced methods.

Although existing texts in applied linguistics touch all ends of these continua, this volume stands out on several accounts. First, I found Cantos’s book to provide exceptionally clear, concise, and accessible explanations of basic concepts and descriptive statistics (Chapter 1), which form the foundation and set the tone for the rest of the book. The clarity of explanations, aided by numerous graphic displays, follows throughout the entire text. The introduction to multivariate statistics (Chapter 4), for example, which can intimidate even the most quantitatively oriented among us, portrays this set of procedures in a way that gets at the heart of the matter: reducing the complex nature of our reality—and our data—into terms, models, and sets of relationships that we can understand. The author accomplishes this while managing to avoid sensationalizing these techniques; the focus is on their utility, a feature less likely to prompt methods-driven research.

The text’s breadth is impressive as well, ranging from descriptives and univariate analyses to multivariate techniques (e.g., factor analysis). Furthermore, a third of the book’s chapters describe quantitative techniques for analyzing corpus data, an area often overlooked and left for specialized volumes. This variety of techniques provides a much-needed first exposure for novice researchers; for more experienced ones, a much-needed refresher.

At the same time, I was not entirely satisfied with the book’s scope, which may have been overly ambitious. To be clear, I appreciate the author’s inclusion of multivariate and corpus-based statistics. However, I might have preferred to exchange breadth and range of techniques for depth in carrying out those techniques. That is, in order to carry readers through to a workable knowledge of the techniques that are covered, providing procedural instructions using one or more software packages (e.g., SPSS) is critical. As is, readers need to consult an additional manual to reconcile software procedures with the conceptual presentations provided in this text. I should mention, though, that this is the case in nearly all statistics texts in applied linguistics. (Two notable exceptions are Larson-Hall, Reference Larson-Hall2010, for SPSS, and Gries, 2013, for R.) The book’s usefulness, especially as a course text, would have also been enhanced by providing sample data for students to practice running some of the analyses described.

Another limitation of the text is the complete omission of a number of very basic but equally important concepts such as confidence intervals, statistical power (coincidentally, many of the book’s sample analyses appear to be lacking power), and the relative merits of statistical versus practical significance. For example, not introducing the Cohen’s d effect size along with z scores and other descriptive statistics is a sorely missed opportunity. The field’s consideration of these statistics, largely absent until recent years (see Plonsky & Gass, Reference Plonsky and Gass2011), represents an important shift in statistical literacy; their exclusion from a new text is unfortunate. Other less weighty oversights include discussion of covariates, commonalities among the family of statistics within the general linear model (e.g., ANOVA and multiple regression), and the value of “to what extent”-type research questions (and their corresponding analyses) compared to the common but less informative “yes/no” variety.

Yes, I have reservations related to the book’s content and utility as a course text, some of which I admit reflect my interests and priorities as much as limitations to Cantos’s work. Nevertheless, a thorough read of this book would benefit any applied linguist conducting quantitative research, regardless of orientation and background, but especially those interested in analyzing corpus data. Few methods authors in the field have ventured into the statistical territory mapped out in this book. Even fewer have succeeded, as this book has, in presenting the material in such an approachable manner.