Understanding Contamination Bias

Problems and solutions of linear regression with multiple treatments

In many causal inference settings, we might be interested in the effect of not just one treatment, but many mutually exclusive treatments. For example, we might want to test alternative UX designs, or drugs, or policies. Depending on the context, there might be many reasons why we want to test different treatments at the same time, but generally, it can help reduce the sample size, as …

bias causal inference data science econometrics editors pick python understanding