Logistic Regression Models for Ordinal Response Variables provides applied researchers in the social, educational, and behavioral sciences with an accessible and comprehensive coverage of analyses for ordinal outcomes. The content builds on a review of logistic regression, and extends to details of the cumulative (proportional) odds, continuation ratio, and adjacent category models for ordinal data. Description and examples of partial proportional odds models are also provided. This book is highly readable, with lots of examples and in-depth explanations and interpretations of model characteristics.
About the Author
Statistical methods, with an emphasis on single and multilevel generalized linear models; evaluation of professional development and health/education programs or interventions, particularly for HIV prevention; secondary analysis of large-scale databases; capacity building for community-based organizations; translation of evidence-based interventions.
Table of Contents
List of Tables and FiguresSeries Editor’s IntroductionAcknowledgments1. Introduction Purpose of This Book Software and Syntax Organization of the Chapters2. Context: Early Childhood Longitudinal Study Overview of the Early Childhood Longitudinal Study Practical Relevance of Ordinal Outcomes Variables in the Models3. Background: Logistic Regression Overview of Logistic Regression Assessing Model Fit Interpreting the Model Measures of Association EXAMPLE 3.1: Logistic Regression Comparing Results Across Statistical Programs4. The Cumulative (Proportional) Odds Model for Ordinal Outcomes Overview of the Cumulative Odds Model EXAMPLE 4.1: Cumulative Odds Model With a Single Explanatory Variable EXAMPLE 4.2: Full-Model Analysis of Cumulative Odds Assumption of Proportional Odds and Linearity in the Logit Alternatives to the Cumulative Odds Model EXAMPLE 4.3: Partial Proportional Odds5. The Continuation Ratio Model Overview of the Continuation Ratio Model Link Functions Probabilities of Interest Directionality of Responses and Formation of the Continuation Ratios EXAMPLE 5.1: Continuation Ratio Model With Logit Link and Restructuring the Data EXAMPLE 5.2: Continuation Ratio Model With Complementary Log-Log Link Choice of Link and Equivalence of Two Clog-Log Models Choice of Approach for Continuation Ratio Models EXAMPLE 5.3: Full-Model Continuation Ratio Analyses for the ECLS-K Data6. The Adjacent Categories Model Overview of the Adjacent Categories Model EXAMPLE 6.1: Gender-Only Model EXAMPLE 6.2: Adjacent Categories Model With Two Explanatory Variables EXAMPLE 6.3: Full Adjacent Categories Model Analysis7. Conclusion Considerations for Further StudyNotesAppendix A: Chapter 3Appendix B: Chapter 4Appendix C: Chapter 5Appendix D: Chapter 6ReferencesIndexAbout the Author