Using real-world data examples, this authoritative book shows how to use the latest configural frequency analysis (CFA) techniques to analyze categorical data. Some of the techniques are presented here for the first time. In contrast to methods that focus on relationships among variables, such as log-linear modeling, CFA allows researchers to evaluate differences and change at the level of individual cells in a table. Illustrated are ways to identify and test for cell configurations that are either consistent with or contrary to hypothesized patterns (the types and antitypes of CFA); control for potential covariates that might influence observed results; develop innovative prediction models; address questions of moderation and mediation; and analyze intensive longitudinal data. The book also describes free software applications for executing CFA.
Tabela de Conteúdo
1. Introduction1.1 Questions That CFA Can Answer1.2 The Five Steps of CFA1.3 Introduction to CFA: An Overview1.4 Chapter Summary2. Configural Analysis of Rater Agreement2.1 Rater Agreement CFA2.2 Data Examples2.3 Chapter Summary3. Structural Zeros in CFA3.1 Blanking Out Structural Zeros3.2 Structural Zeros by Design3.2.1 Polynomials and the Method of Differences3.2.2 Identifying Zeros That Are Structural by Design3.3 Chapter Summary4. Covariates in CFA4.1 CFA and Covariates4.2 Chapter Summary5. Configural Prediction Models5.1 Logistic Regression and Prediction CFA5.1.1 Logistic Regression5.1.2 Prediction CFA5.1.3 Comparing Logistic Regression and P-CFA Models5.2 Predicting an End Point5.3 Predicting a Trajectory5.4 Graphical Presentation of Results of P-CFA Models5.5 Chapter Summary6. Configural Mediator Models6.1 Logistic Regression plus Mediation6.2 CFA-Based Mediation Analysis6.3 Configural Chain Models6.4 Chapter Summary7. Auto-Association CFA7.1 A-CFA without Covariates7.2 A-CFA with Covariates7.2.1 A-CFA with Covariates I: Types and Antitypes Reflect Any of the Possible Relationships between Two or More Series of Measures7.2.2 A-CFA with Covariates II: Types and Antitypes Reflect Only Relationships between the Series of Measures and the Covariate7.3 Chapter Summary8. Configural Moderator Models8.1 Configural Moderator Analysis: Base Models with and without Moderator8.2 Longitudinal Configural Moderator Analysis under Consideration of Auto-Associations8.3 Configural Moderator Analysis as n-Group Comparison8.4 Moderated Mediation8.5 Graphical Representation of Configural Moderator Results8.6 Chapter Summary9. The Validity of CFA Types and Antitypes9.1 Validity in CFA9.2 Chapter Summary10. Functional CFA10.1 F-CFA I: An Alternative Approach to Exploratory CFA (Sequential Identification of Types and Antitypes)10.1.1 Kieser and Victor’s Alternative, Sequential CFA: Focus on Model Fit10.1.2 von Eye and Mair’s Sequential CFA: Focus on Residuals10.2 Special Case: One Dichotomous Variable10.3 F-CFA II: Explaining Types and Antitypes10.3.1 Explaining Types and Antitypes: The Ascending, Inclusive Strategy10.3.2 Explaining Types and Antitypes: The Descending, Exclusive Strategy10.4 Chapter Summary11. CFA of Intensive Categorical Longitudinal Data11.1 CFA of Runs
Sobre o autor
Alexander von Eye is Professor of Psychology at Michigan State University. He develops, studies, and applies methods for the analysis of categorical data (in particular, configural frequency analysis and log-linear modeling) and longitudinal data. He also works on and with classification methods and conducts simulation studies. Dr. von Eye has published over 350 articles in methodological, statistical, psychological, and developmental journals, and he is the (co)author or (co)editor of 18 books. He is a Fellow of the American Psychological Association and the American Psychological Society, and he was visiting professor of statistics, psychology, human development, and education at a number of universities in Austria and Germany, as well as at Penn State. Patrick Mair is Assistant Professor in the Institute for Statistics and Mathematics, WU Vienna University of Economics and Business. He was a visiting scholar at the University of California, Los Angeles. Dr. Mair’s research focuses on computational/applied statistics and psychometrics, including methodological developments as well as corresponding implementations in the statistical computing environment R. His publications appear in journals of applied and computational statistics. Eun-Young Mun is Assistant Professor of Psychology at Rutgers, The State University of New Jersey. Her research aims to better understand how alcohol and drug use behaviors develop over time, and to delineate mechanisms of behavior change in order to develop effective prevention and intervention approaches, especially for adolescents and emerging adults. She is also interested in extending existing research methodology by integrating and synthesizing distinctive methods together–in particular, pattern-oriented and person-oriented longitudinal research methods–and by disseminating applications. She is coauthor of Analyzing Rater Agreement and publishes articles in developmental, clinical, and methodological journals.