LPO 8851: Regression Analysis I

Course Description

This PhD-level course is structured to provide doctoral students with a foundation in regression analysis that will prepare them for subsequent applied regression and causal inference courses. Topics include the following: hypothesis tests and confidence intervals for population variances; simple linear regression model; testing hypotheses about regression parameters; goodness-of-fit tests; multiple regression model; omitted variable bias; measurement error; and limited dependent variable models (e.g., logit and probit models). An introduction to panel data may also be included, if time permits. The course will make use of Stata, a statistical package, to provide examples of topics learned through lectures.

Learning Objectives

1. Formal understanding of econometric theory
   • Identify and explain the assumptions underlying regression analysis
   • Identify situations when assumptions do not hold and explain how to best address these situations
2. Ability to apply regression techniques with a competency sufficient for academic level, empirical research

Practical Skills and Knowledge

1. Read and interpret regression output
2. Understand how functional form decisions affect the interpetation and magnitude of regression statistics and parameter estimates
3. Distinguish between best linear predictor (BLP) and conditional expectation function (CEF) models
4. Understand which variables should (and should not) be included when specifying regression equations
5. Write programs to implement regression analysis
6. Conduct simulations to test regression assumptions