Experimental and Quasi-Experimental Methods

Introduction

This course offers an introduction to causal inference using experimental and quasi-experimental methods to study marketing phenomena. It should prepare doctoral students to design, critically evaluate, and analyze empirical research that uncovers cause-and-effect relationships in marketing using state-of-the art causal inference methods and paradigms. Students will learn about identification strategies of causal effects in experimental and quasi-experimental designs as well as statistical methods that allow to derive causal relationships from these designs. Through hands-on examples and the discussion of top-tier academic journal articles, students will have the opportunity to learn how to employ field experiments, lab experiments, and quasi-experimental designs and methods—such as difference-in-differences, instrumental variables, matching, regression discontinuity, and double machine learning—to address causal marketing questions. By the end of the course, students should be equipped to make informed decisions about the most appropriate methods and designs for their research inquiries but also to communicate findings effectively to academic audiences. In doing so, they will gain the expertise needed to generate and evaluate high-impact (marketing) research that is grounded in robust empirical evidence.

Course content

The course will roughly center around the following core topics. Details are exemplary and may be subject to smaller changes.

• Foundations of causal inference
  • What is causality?
  • Differences between causal, predictive, and descriptive questions
  • Rubin’s potential outcome framework
  • Different types of treatment effects
  • Basics of causal diagrams (DAGs)
• Experimental designs
  • Field vs. lab vs. natural experiments
  • Between-, within-subject and mixed designs
  • Multi-factorial designs including interactions
  • Internal and external validity of different designs
• Quasi-experimental and observational methods for causal inference
  • Regression adjustment, matching, and weighting
  • Instrumental variables (IVs) and IV-free methods
  • Difference-in-difference, synthetic control, and synthetic difference-in-difference
  • Causal Machine Learning (double debiased ML)