David Kaplan (Ed.), The SAGE Handbook of Quantitative Methodology for the Social Sciences.

David Kaplan (Ed.), The SAGE Handbook of Quantitative Methodology for the Social Sciences, Thousand Oaks, CA: Sage Publications, 2005. $125.00 hardcover.

The SAGE Handbook of Quantitative Methodology for the Social Sciences was written to introduce quantitative social scientists, applied statisticians, and graduate students to a broad range of state-of-the-art quantitative methods applicable to the social sciences. The Handbook was written by a distinguished group of contributors from psychology, education, statistics, and other related disciplines, and virtually all contributors are established authorities in their particular areas. It shows the breadth and depth of advanced quantitative methods used by social scientists from numerous interrelated disciplines, it is rich with examples of real-world applications of these methods, and it provides suggestions for further readings and study in these areas.

The Handbook is organized into six sections. The first two pertain to measurement. Specifically, the first section, Scaling, includes chapters on dual scaling; multidimensional scaling; and principal components analysis for ordinal and nominal data. The second section, Testing and Measurement, includes chapters on advances in reliability and validity theory; test modeling (item-response theory in particular); differential item functioning; and computerized adaptive testing. Sections three through five detail advances in statistical methodology. Specifically, the third section, Models for Categorical Data, includes chapters on trends in categorical data analysis; ordinal regression models; latent class models; and discrete-time survival analysis. The fourth section, Models for Multilevel Data, includes chapters on growth modeling; multilevel models for school effectiveness research; hierarchical models for analyzing data from experimental and quasi-experimental designs; and meta-analysis. The fifth section, Models for Latent Variables, includes chapters on determining the number of factors in exploratory and confirmatory factor analysis; experimental, quasi-experimental, and nonexperimental design and analysis with latent variables; dynamic factor analysis; and latent variable growth mixture modeling for longitudinal data. Finally, the sixth section, Foundation Issues, provides discussion of the major philosophical issues that underlie quantitative methods. Specifically, it includes chapters on probabilistic modeling with Bayesian networks; null hypothesis testing; exogeneity; objectivity in science and structural equation modeling; and causal inference.

The Handbook provides an excellent introduction to a broad range of state-of-the-art quantitative methods applicable to the social sciences. But why is it important for quantitative social scientists to spend the considerable time required to understand and use these methods? These methods are important for a number of reasons. They allow us to study complex social phenomena that cannot be represented accurately with simple methods (for example, methods that do not take into consideration measurement error, multilevel data structures, or complex sample structure). They also allow us to better handle methodological challenges posed in applied research (for example, new methods for analyzing data from quasi-experimental designs and for handling missing data). Finally, these methods may stimulate the formulation of research questions in new and potentially important ways (e.g., latent class models allow us to think about person-centered research questions in addition to traditional variable-centered research questions).

This Handbook does an excellent job in covering a broad range of state-of-the-art quantitative methods applicable to the social sciences. However, there are a few topics that I wish had have been covered in separate chapters, not just as part of existing chapters (and most of these topics are indeed covered at least in part in existing chapters). These include chapters on recent advances in methods for handling missing data, methods for analyzing data from complex samples, methods for determining statistical power for many of the methods described in sections three through five, and perhaps nonlinear regression methods and methods for modeling count data.

Although some of the chapters in the Handbook are mathematically challenging, the Handbook is successful in providing an introduction to a broad range of state-of-the-art quantitative social science methods to quantitative social scientists, applied statisticians, and graduate students. It is well worth reading cover-to-cover, and it is a very useful addition to the reference libraries of all quantitative social scientists, applied statisticians, and graduate students.

John G. Orme

University of Tennessee