Measurement connects theoretical concepts to what is observable in the empirical world, and is fundamental to all social and behavioral research. In this volume, J. Micah Roos and Shawn Bauldry introduce a popular approach to measurement: Confirmatory Factor Analysis (CFA). As the authors explain, CFA is a theoretically informed statistical framework for linking multiple observed variables to latent variables that are not directly measurable. The authors begin by defining terms, introducing notation, and illustrating a wide variety of measurement models with different relationships between latent and observed variables. They proceed to a thorough treatment of model estimation, followed by a discussion of model fit. Most of the volume focuses on measures that approximate continuous variables, but the authors also devote a chapter to categorical indicators. Each chapter develops a different example (sometimes two) covering topics as diverse as racist attitudes, theological conservatism, leadership qualities, psychological distress, self-efficacy, beliefs about democracy, and Christian nationalism drawn mainly from national surveys. Data to replicate the examples are available on a companion website at https://study.sagepub.com/researchmethods/qass/roos-confirmatory-factor-analysis, along with code for R, Stata, and Mplus.
Chapter 1: Introduction
Chapter 2: Model Specification
Chapter 3: Identification and Estimation
Chapter 4: Model Evaluation and Respecification
Chapter 5: Measurement Invariance
Chapter 6: Categorical Indicators
Chapter 7: Conclusion
Appendix: Reliability of Scales
Student Study Site
Example code as well as data (variously in R using lava, Stata, and Mplus) is provided by chapter.