What is it, and how does it work?

Photo by Jeremy Thomas on Unsplash

Professor Gary King and his colleagues, Christopher Lucas and Richard A. Nielsen, published a paper titled “The Balance-Sample Size Frontier in Matching Methods for Causal Inference” at the American Journal of Political Science in 2017. In the paper abstract, they wrote: “We propose a simplified approach to matching for causal inference that simultaneously optimizes balance and matched sample size.” This paper caught my eye because it is relevant to the important bias-variance trade-off for matching.

This post aims to provide a non-technical description of the method proposed in King et al (2017). How, and why does the method work? What…


How does it actually work?

Photo by Anne Nygård on Unsplash

My previous post shows how the propensity score matching can be implemented using 13 lines of R codes. Though those codes, you can see that matching is much about data pre-processing. The idea is to find control units that are comparable to the treated units, so we can attribute the differences in the outcome between treatment groups to the treatment with more confidence. Besides, I want to argue that matching is, in fact, simple, so you should not be terrified by it.

Today, I look into a popular alternative to propensity score matching: Mahalanobis Metric (MM) matching. The two matching…


You can manually do matching using 13 lines of R codes

upslash.com

Background

A challenge with many observational studies for obtaining causal effect is self-selection — in many cases, people choose to receive treatment for some reasons, and consequently, the treated people cannot be directly compared with the untreated people. For instance, students from rich families might be more likely to choose to attend private college. If we wish to know the causal effect of attending private college on earnings, we shall rule out (or control) the effect of family background.

Matching estimator has been widely used across disciplines such as statistics, economics, sociology, political science, etc, to estimate causal effects. …


Though terrifying, economists dare to speak its name.

In J. K. Rowling’s series of Harry Potter novels, Lord Voldemort is so feared in the wizarding world that it is considered dangerous even to speak his name. So, nearly every witch or wizard and refers to Lord Voldemort instead with such monikers as You-Know-Who.

Among economists, the word “cause” is perhaps almost as terrifying as the name Voldemort.

Economists nowadays are trained to treat the word “cause” very carefully, and, by training, any causal claim is prohibited unless a great effort of smart work has been dedicated to support the claim. …


A fairly simple and intuitive method for identifying the causal effects

Inverse Probability Weighting (IPW) is a popular quasi-experimental statistical method for estimating causal effects under the assumption of conditional independence. This method can be easily implemented and easy to understand. In this post, I will provide an explanation of this method with minimal R codes.

What is wrong with the Ordinary Least Squares (OLS)?

Consider a case that we are interested in the effect of a binary treatment (labelled as W) on a continuous outcome (labelled as Y). Using data, we could estimate the Average Treatment Effect (ATE), which is the average effect of W on Y over all observation units (both treated and untreated).

OLS has been widely…

Bowen Chen, PhD

Applied Economist & Data Scientist

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store