Estimating causal effects using non-experimental data is always challenging due to the presence of cofounders: common factors that cause both the treatment and outcome of interest. Like regression modelling, matching is a technique designed to account for the existence of observed confounders. It can be used on its own or in conjunction with other statistical methods. Whilst it is popular in fields such as epidemiology, political science and medicine; it is rarely taught in other social science disciplines where there is also a lower emphasis on causal inference.

This workshop will provide a crash course in matching and how to implement it in R for social science researchers. It will also briefly introduce the topic of causal inference and its relationship to research design.

This course is designed for those who are interested in causal inference regardless of their level of statistical knowledge.

Attendees will be able to learn the following

• Understand the purpose of matching and how it is used in research (e.g. journal articles)
• Be able to implement commonly used matching algorithms in R
• A short introduction to causal inference.