Statistics in language research: Analysis of variance (review)

The use of parametric statistics—especially analysis of variance (ANOVA) and linear regression— has played an essential role in language research in a number of disciplinary areas for decades, including psycholinguistics and phonetics. There has been a significant increase, however, in the use of such analytic techniques in other areas of linguistics as well, including syntax, semantics, sociolinguistics, and dialectology. The need for these analytic methods often arises whenever the researcher is analyzing language data (whether collected through surveys, in experiments, in field studies, or in language corpora). The authors’ goals are to provide a comprehensive description of ANOVA and instructions for organizing the data, to conduct analyses in the Statistical Package for the Social Sciences (SPSS), and to interpret the results. This book is designed for students (and researchers) who already have a solid background in basic descriptive and inferential statistics, but who want to develop extended skills in the use of ANOVA techniques; knowledge of basic precalculus mathematics is expected. The authors are well known in their respective fields and extremely well qualified to write this book. Toni Rietveld has numerous publications related to analytic methodologies and statistical treatment of language data, and Roeland van Hout studies language variation in the Dutch language area, including dialect lexicography (and analysis of language corpora) and dialectology.

The organization of the book is quite straightforward and easy to follow. Each chapter is devoted to a reasonably narrow topic and divided into separate sections, each of which builds upon concepts developed earlier. Included with all presentations of material are illustrative examples related to various aspects of language research—a welcome change from standard textbooks on statistics in which examples are often drawn from the social sciences or agriculture. Each chapter has a ‘preview’ section that outlines what will be covered, the content sections, and a section on ‘terms and concepts’ that provides concise definitions of the most important and relevant terms used in the chapter. At the end of each chapter there is a set of exercises that tests the knowledge gained by the reader, and answers to all exercises appear in an appendix at the end of the book.

Ch. 1, ‘Language research and statistics’ (1–12), is the most elementary chapter in the book and is designed to provide readers with a review of concepts like the nature of data analysis, independent and dependent variables, participants, measurement scales, experimental designs, and between- vs. within-subject factors. The authors have also outlined the specific data formats used in SPSS for these research designs. Anyone who has taught statistics using SPSS (or actually any other statistical package) will appreciate how difficult it is to ensure that students have organized their input data in the proper format in order to successfully complete the analyses they want to perform. Therefore, providing these examples is an especially welcome addition.

Ch. 2, ‘Basic statistical procedures: One sample’ (13–30), and Ch. 3, ‘Basic statistical procedures: Two samples’ (31–48), review basic inferential statistical procedures. They cover the material quickly (in only thirty-seven pages), but the topics include some of those most important to the understanding of hypothesis testing, research design, data analysis, and the interpretation of test results, such as sampling variability (e.g. sampling distribution, standard error, and the sample distribution of the means), Type I (α) and II (β) errors, statistical power (1 – β), sample size, and simple t-tests and z-tests (the former will become important when completing post hoc analyses after a significant effect is found in the ANOV Aanalysis). R&H also give a very readable discussion about how to determine the size of the sample that will be needed in order to achieve a given level of statistical power in an inferential test. This is useful when presenting a rationale to the local Human Subject Review Board for why N participants are needed for a study.

Ch. 4, ‘Principles of analysis of variance’ (49–73), describes the principles of ANOVA in...