Microcredential komex A Hands-on Introduction to Longitudinal and Panel Data Analysis

This five-day in-person course introduces you to longitudinal (panel and repeated cross-sections) analysis using Stata & R focusing on research design, handling & visualization of longitudinal data, and widespread methods (fixed and random effects & multiple variants).

What Is This Course About?
This in-person five-day course gives an accessible introduction to two widely used types of longitudinal analyses in political science and related disciplines: panel data analysis (analyses of repeatedly observed individuals) and analyses of repeated cross-sectional data (same questions repeatedly asked to different individuals of the same population). On the one hand we will discuss advantages of longitudinal analysis from a research design perspective. On the other hand, we will look into how to handle and visualize longitudinal data and learn different models to analyse longitudinal data (so-called fixed and random effects models and variants thereof) using Stata and R.

Learning Goals

• Research Design: Longitudinal analysis & causal inference; between and within variance;
• Data Handling: Merging of longitudinal data sets; Long and wide data structures; (Re-) coding over-time change
• Visualization: Visualize (individual or group) outcome trajectories or treatment status over time;
• Methods and Analysis: the basic fixed-effects and random-effects models. “Hybrid” between-within models.
• Potential advanced topics: interactions, Diff-in-Diff basics, interactive fixed effects / individual slopes, treatment-timing issues
• Application to participants’ research interests / research projects

Assignments for the Course

• Attendance (not graded)
• Several formative in-class assignments (e.g. exercises) (not graded)
• 1-2 after-class assignments (i.e. homework; preparation for the next day) (not graded)
• Graded written assignment with feedback: short paper including own implementation of a longitudinal analysis

Schedule

• 09.00-10.30 – Course (e.g. lecture)
• 10.30-11.00 – Break
• 11.00-12.30 – Course (e.g. lab)
• 12.30-13.30 – Lunch break
• 13.30-14.30 – Course (e.g. supervised small group work)
• Day 1: Getting started: Research design, data structure & visualization
  Longitudinal research designs and causal inference
  Between and within variance
  Longitudinal and multilevel data structures
  Examples for available longitudinal and panel data sets in political science and the social sciences more broadly
  Getting started: Using and visualizing longitudinal data

• Day 2: The basics: fixed effects and random effects
  Preparing longitudinal analyses: data-handling and recoding
  Introduction to random-effects and fixed-effects models
  Implementation of random-effects and fixed-effects models for longitudinal data

• Day 3: More complex issues and methods for longitudinal data (I)
  Longitudinal data as multilevel data
  Between-within “hybrid” models
  Interactions

• Day 4: More complex issues and methods for longitudinal data (II)
  Diff-in-Diff basics
  FEIS & violations of the parallel trends assumption
  (Treatment) Timing

• Day 5: Own Application
  Application of course content to participants’ research questions (incl personal and group discussion)
  Discussion of Open Questions

Recommended Readings for the Course

• Bell, Andrew, Malcolm Fairbrother, and Kelvyn Jones. 2019. “Fixed and Random Effects Models: Making and Informed Choice.” Quality & Quantity 53 (2): 1051–74.
• Brüderl, Josef, & Ludwig, Volker (2015): “Fixed-Effects Panel Regression.”, in: Regression Analysis and Causal Inference. Ed. Best, Henning & Wolf, Christof, p. 327–57. Sage.
• Ruspini, Elisabetta. 1999. “Longitudinal Research and the Analysis of Social Change.” Quality and Quantity 33 (3): 219–27.

Who Is Your Instructor?
Nadja Wehl is a PostDoctoral Researcher at the Cluster of Excellence “The Politics of Inequality” of the University of Konstanz working in the project “Students' Perceptions of Inequality and Fairness (PerFair)”.
In her research research she is particularly interested in the long-lasting effects of early socialization experiences on individuals’ political attitudes and behaviour. Methods-wise she applies methods for causal inference in observational data, including cross-sectional data (matching, weighting, etc) and longitudinal data (various methods based on fixed effects) to disentangle the effects of contemporaneous socio-economic circumstances from experiences in childhood and adolescence.
X: @Na_Wehl
Bluesky: @na-wehl.bsky.social
Website: https://sites.google.com/view/nadjawehl