Causal Effects via Regression

3 Popular Techniques With Python Example Code

This is the 5th article in a series on causal effects. In previous articles, we discussed different ways to compute treatment effects from data. Here, I introduce an alternative approach to estimating causal effects via 3 popular regression-based techniques. I conclude the article with example Python code of how one can use these techniques in practice.

Key Points:

• Regression is a way to learn relationships between variables using data
• 3 popular regression-based approaches for estimating causal effects are: linear regression, double machine learning, and meta-learners

Toy example of causal effect via linear regression. Image by author.

What is regression?

Regression is a way to learn the relationships between variables using data. For example, the relationship between height and weight for adult Huon tree kangaroos in Papua New Guinea.

The output of a regression process is called a model. This is essentially something that we can use to make predictions e.g…