Interpreting and Comparing Effects in Logistic, Probit and Logit Regression

✍ Scribed by Jacques A P Hagenaars; Steffen Kuhnel; Hans-Jurgen Andress

• Publisher: Sage Publications
• Year: 2024
• Tongue: English
• Leaves: 205
• Category: Library

✦ Synopsis

Interpreting Effects in Logistic Regression and Logit Models shows how to compare coefficient estimates from regression models for categorical dependent variables in three typical research situations: (i) within one model, (ii) between identical models estimated in different subgroups, and (iii) between nested models. Additionally, this volume presents a practical, unified treatment of comparison problems and considers the advantages and disadvantages of each approach and when to use them.

✦ Table of Contents

Cover
PRAISE FOR THIS BOOK
QUANTITATIVE APPLICATIONS IN THE SOCIAL SCIENCES
SERIES: QUANTITATIVE APPLICATIONS IN THE SOCIAL SCIENCES
Half title
Dedication
Title Page
Brief Contents
Detailed Contents
Series Editor Introduction
Preface and Acknowledgements
About the Authors
1 - Introduction
1.1
Purpose
1.2
Content
1.3
Causality
2 - Regression Models for a Dichotomous Dependent Variable
2.1
Introduction
Simulated Data Set University
2.2
Discrete Response Model — DRM
2.2.1
Logistic Regression, Response Profiles, Discrete
(​DC​), and Instantaneous (​IC​) Change Measures
2.2.2
Logistic DRM as a Logit Model: Odds Ratios as Effect Measures
2.2.3
Probit Regression
2.2.4
Linear Probability Model – LPM
2.3
Latent Variable Model — LVM
2.3.1
Logistic Latent Variable Model
Underlying Standardized Effects and Explained Variance of ​​Y​​ ​​
2.3.2
Probit Latent Variable Model
2.3.3
Heteroscedastic Errors, Unequal Thresholds, and Biased Effects
2.4
Inserting Mavericks, “Orthogonal” Independent Variables, Into Equations
3 - Interpreting and Comparing Effects Within One Equation
3.1
Comparing Effects Within a Single LVM Equation
3.2
Comparing Effects Within a Single
DRM Equation
3.3
Causal Interpretations in LVM and DRM Logistic Regression
3.3.1
Causal LVM
3.3.2
Causal DRM: Graphs, SEMs, and Pearl’s Structural Modeling Approach
3.3.3
Causal DRM: Rubin’s Potential Outcome Approach
In Sum
4 - Comparing Subgroups or Time Points: Investigating Interaction Effects
4.1
Interaction Effects in LVM
4.2
Interaction Effects in DRM
4.3
Interaction and Causal Analysis
In Sum
5 - Causal Modeling: Estimating Total, Direct, Indirect and Spurious Effects; Using Effect Coefficients From Different (Nested) Equations
5.1
Introduction
Discrimination Data
5.1.1
General Background: DAGs, Path Models, Residualized Independent Variables
DAGs—Directed Acyclical Graphs
Standard Path Models
Residualized Independent Variables
5.1.2
Comparability and Compatibility Issues
Comparability Problems: Scaling and Collapsing Effects
Compatibility Problem
5.2
LVM
5.2.1
Illustrating the Problems: Different Scaling Factors and Incompatible Assumptions
Effects on D: Creditworthiness
Assessing Underlying Indirect Effects
Evaluating the Scaling Ratio ​​φ​ r​​ / ​φ​ f​​​
5.2.2
Solution: Comparing Standardized Effects
Tackling the Comparability/Scaling Problem
Tackling Both the Comparability/Scaling and the Compatibility Problem
5.2.3
Solution: Residualization and the KHB Approach
5.2.4
Solution: The Path Analytic Approach
5.3
DRM
5.3.1
The Main Problem: Confusing Collapsing, and Mediating
5.3.2
Separating Collapsing and Mediation: The Path Analytic Approach
Linear-Additive Relationships Among the Independent Variables
Systems of Logit Equations
5.3.3
Simulations and Counterfactual Approaches:SIMMAV and SIMNAT
SIMMAV —Simulating Mavericks
SIMNAT — Simulating Pure, Natural Effects
SIMNAT — Direct Natural Effects (OR)
SIMNAT — Direct Natural DC Effects
SIMNAT — Indirect and Total Natural Effects (OR, DC)
Modified Natural/Pure Effects Using SIMMAV
5.4
Causal Modeling
In Sum
6 - Concluding Remarks; Extensions, Effect Measures, and Evaluation
6.1
Polytomous Dependent Variable
6.1.1
LVM
6.1.2
DRM
6.2
How to Measure Effects in Logistic
Regression
6.3
Concluding Remarks
References
Index