Impact Evaluation Methods

Overview
 

The course introduces students to the modern framework for causal inference in economics and other social sciences. The students will learn about the concepts of research design and identification strategy and how to apply these concepts to answer various research questions. The course starts by introducing the two workhorse models for understanding identification: the potential outcomes model and causal diagrams. We will then cover most commonly used tools for identifying and estimating causal effects: regression, matching, instrumental variables, regression discontinuity, and difference-in-differences, as well as a few recent developments, if time permits. Mastering these tools will allow students to answer their own research questions in academic, public, and private-sector contexts. The course will guide students through the intuition behind the methodology, formal derivations and proofs, as well as practical tools to implement each method. The course content relies on a mix of textbooks, article readings, and practical exercises in R.

Learning Objectives
 

At the end of the course, the students should 

• be able to explain the difference between correlation and causation
• understand the role of research design and identification strategy in identifying causal effects
• be capable of using the potential outcomes model and causal diagrams to come up with credible research designs
• know when to apply different causal inference methods
• be able to perform those methods on existing data sets, interpret the results, and defend their identification strategy

Prerequisites 
 

Knowledge of at least introductory-level statistics and econometrics is required. Knowledge of statistical software, such as STATA, R, or Python, is helpful but not required.

L: Lecture; T: Tutorial

Content questions: Prof. Dr. Lea Cassar (opens your email program)

Organisational and technical questions: Iris Bleisteiner (opens your email program), Vanessa Dietl (opens your email program)