## Introduction

Welcome to ECNU3ED–Using Economic Data for Causal Inference a `Nobel-Prize-Winning endeavour™’. This is a rather unusual module that will introduced you to three things:

- The workhorse model of Econometrics (OLS)
- Thinking about causality using Econometrics
- Statistical Programming using Stata

This class is supported by DataCamp, an incredibly intuitive platform to learn R. They support education around the world for free via this initiative. Find out more here .

## Module Learning Outcomes

On successful completion of the module, you should be able to:

- analyse economic data using regression analysis and statistical inference, and critically interpret regression output;
- synthesise and communicate canonical causal research designs, with an emphasis on plausibility and application;
- demonstrate a professional working proficiency in data collection, cleaning, and analysis in using reproducible statistical programming packages;
- design and execute a causal evaluation of a real-world policy intervention and communicate the findings to a wide audience.

## Detailed Description

Over the last 30 years, economics has undergone a credibility revolution. This moment has sought to move economics from theory to empirics. This work is grounded on the need for economic theory to be evidence based and consequently economists have sought to use causal inference fuelled by experiments and quasi-experimental methods to analyse policies. This module will arm you with the toolkit of the modern economist.

This module is in two halves. In the first, you’ll be given a grounding on linear regression with a strong emphasis on practical interpretation and use of ordinary least squares, the workhorse model of this module. In the second half of the module we focus on causality, and how this may be inferred using observational data.

The methods covered include:

directed acyclic graphs;

instrumental variables;

regression discontinuity design;
differences-in-differences.

The module also develops your skills by teaching statistical software. Many economics graduates and employers report that skills in data analysis are among the most important and regularly used of those needed by economists working in industry, or government. This module differs to others in as much as Statistical programming is not assumed to be mana from heaven and instead data cleaning and working with data are explicitly taught.

## Course Outline

This section gives a flavour for the reading list and pacing of the module each heading is covered in a single week. The suggested readings for each week are displayed below.

#### Stata Crash Course

Over the first few weeks we will be doing a Stata Crash course, Stata (and to a lesser extent R) is the software we will use to do econometrics. But first we need to learn to use it.

#### Asking Causal Questions

- Wooldridge 1
- Mastering Metrics 1
- Mixtape introduction
- The Effect 5

#### Correlation & Bivariate OLS

- Seeing Theory 6
- Wooldridge 2
- Mixtape 2
- The Effect 13

#### Multiple regression

- Wooldridge 3
- Mastering Metrics 3
- Mixtape 2
- The Effect 13

#### Statistical Inference in Regressions

- Wooldridge 4
- The Effect 13

#### Functional Form & Dummy Variables

- Wooldridge 7; 6.1, 6.2, 9.1
- The Effect 13

#### Violations of Gauss Markov & Careers talk

- Wooldridge Chp: 8, 9.4, 12
- The Effect 13

#### Group Project Presentations & Directed Acyclic Graphs

#### Instrumental Variables

- Mastering Metrics 3
- Wooldridge 15
- Mixtape 7
- The Effect 19

#### Regression Discontinuity

- Mastering Metrics 4
- Mixtape 6
- The Effect 20

#### Differences-in-differences

- Mastering Metrics 5
- Mixtape 9
- Wooldridge 13
- Effect 16-18