Causal Analysis with Event History Data Using Stata

Hans-Peter Blossfeld, Prof., Dr. rer. pol. Dr. h. c., has been Emeritus of Excellence at the Graduate Centre Trimberg Research Academy (TRAc) at the University of Bamberg in Germany since April 2020. He held the Chair of Sociology I at the Faculty of Social Sciences, Economics and Business Administration at the University of Bamberg and was Professor of Sociology at the European University Institute in Florence, Italy.Götz Rohwer was Professor Emeritus of Methods of Social Research and Statistics at Ruhr-University Bochum in Germany. He passed away in March 2021.Gwendolin J. Blossfeld is a Postdoc at the Faculty of Social Sciences, Economics and Business Administration at the University of Bamberg in Germany.

• 1 Introduction 11.1 Causal Modeling and Observation Plans 1.1.1 Cross-Sectional Data1.1.2 Panel Data 1.1.3 Event History Data1.2 Event History Analysis and Causal Modeling 1.2.1 Causal Explanations1.2.2 Transition Rate Models 2 Event History Data Structures 2.1 Basic Terminology 2.2 Event History Data Organization3 Nonparametric Descriptive Methods 3.1 Life Table Method 3.2 Product-Limit Estimation 3.3 Comparing Survivor Functions 4 Exponential Transition Rate Models 4.1 The Basic Exponential Model 4.1.1 Maximum Likelihood Estimation4.1.2 Models without Covariates4.1.3 Time-Constant Covariates4.2 Models with Multiple Destinations 4.3 Models with Multiple Episodes5 Piecewise Constant Exponential Models 5.1 The Basic Model 5.2 Models without Covariates 5.3 Models with Proportional Covariate Effects5.4 Models with Period-Specific Effects 6 Exponential Models with Time-Dependent Covariates 6.1 Parallel and Interdependent Processes6.2 Interdependent Processes: The System Approach6.3 Interdependent Processes: The Causal Approach6.4 Episode Splitting with Qualitative Covariates 6.5 Episode Splitting with Quantitative Covariates6.6 Application Examples 7 Parametric Models of Time Dependence 7.1 Interpretation of Time Dependence7.2 Gompertz Models 7.3 Weibull Models 7.4 Log-Logistic Models7.5 Log-Normal Models 8 Methods for Testing Parametric Assumptions 8.1 Simple Graphical Methods 8.2 Pseudoresiduals9 Semiparametric Transition Rate Models 9.1 Partial Likelihood Estimation9.2 Time-Dependent Covariates9.3 The Proportionality Assumption9.4 Stratification with Covariates and for Multiepisode Data9.5 Baseline Rates and Survivor Functions9.6 Application Example10 Problems of Model Specification 10.1 Unobserved Heterogeneity 10.2 Models with a Mixture Distribution10.2.1 Models with a Gamma Mixture10.2.2 Exponential Models with a Gamma Mixture 10.2.3 Weibull Models with a Gamma Mixture10.2.4 Random Effects for Multiepisode Data10.3 Discussion 11 Sequence Analysis Brendan Halpin11.1 What is Sequence Analysis? 11.1.1 Sequence Data 11.1.2 The Value of a Holistic View 11.2 Defining Distances 11.2.1 Hamming Distance 11.2.2 Optimal Matching Distance 11.2.3 Other Distances 11.2.4 Determining State Distances11.3 Doing Sequence Analysis in Stata .11.3.1 Example Data 11.3.2 A First Look at the Data 11.4 Unary Summaries11.5 Intersequence Distance 11.6 What to Do with Sequence Distances? 11.7 Optimal Matching Distance11.8 Special Topics 11.8.1 Other Distance Measures11.8.2 Ideal Types 11.8.3 Multichannel Sequence Analysis 11.8.4 Dyadic Analysis 11.9 Conclusion Appendix: Exercises References About the Authors