Methods-Workshop

We are pleased to offer you an exciting workshop on "Causal Analysis of Panel Data using Structural Equation Modeling" by Prof. Dr. Jochen Mayerl from TU Chemnitz

Research Interests 

(1) Methods of empirical social research: 

• computer-assisted survey methods (including: Paradata and response reaction time measurements, experiments in online surveys, digital methods and big data in the social sciences) 
• Social Cognition in Surveys (e.g. respondent behavior, response effects, social desirability) 
• Cross-National Surveys (international comparative survey research) 
• Structural equation modeling (e.g. models of causal inference, longitudinal models, measurement equations) (e.g. models of causal inference, longitudinal models, measurement equivalence)

(2) Theory-oriented social research: 

• Attitude research (attitude-behavior research, change and stability of attitudes) 
• Social context effects
• Bounded rationality 
• Content-related applications in areas such as:
  environmental sociology
  health awareness
  ethnocentrism
  technology acceptance
  prosocial behavior 

Workshop Causal Analysis of Panel Data using Structural Equation Modeling 

The analysis of panel data is a central tool for analyzing causal and reciprocal relationships as well as change and stability in the social sciences. Panel analysis with SEM includes the following modeling variants in particular: 

• autoregressive cross-lagged models 
• fixed effects panel models
• latent growth curve models 
• hybrid models: latent growth curve models with structured residuals 

The workshop focuses on appropriate model specification and interpretation of results, but also on methodological pitfalls of panel analysis and typical problems of unobserved heterogeneity, measurement equivalence, time-varying and -invariant predictors, reverse causality and feedback loops as well as measurement error correlations. The workshop introduces the application-oriented statistical analysis of panel data (e.g. with data from the Gesis Panel) with structural equation models (SEM) using typical SEM software (especially R with the lavaan package, but also Mplus and AMOS).