Applied Data Analysis

Catalog description

This course will examine how and when data can be used specifically to infer whether there is a causal relationship between two variables. We will emphasize (a) the critical role of an underlying economic theory of behavior in interpreting data and guiding analysis, as well as (b) a range of advanced techniques for inferring causality from data, such as randomized controlled trials, regression discontinuity, difference-in-difference, audit study (mystery shopping) approaches and stock-market event studies. The issue of causality, and the relevance of thinking about models and methods for inferring causality, is just as central and important for "Big Data" as it is when working with traditional data sets in business and public policy. The emphasis will not be on proofs and derivations but rather on understanding the underlying concepts, the practical use, implications and limitations of techniques. Students will work intensively with data, drawing from examples in business and public policy, to develop the skills to use data analysis to make better decisions. All analysis will be conducted using R. The goals of the course are for students to become expert consumers able to interpret and evaluate empirical studies as well as expert producers of convincing empirical analysis themselves.