Designing and conducting A/B tests

A deep dive into A/B testing and experimental design helps us establish a credible link between cause and effect. As you know by now, at the core of causal inference is the challenge of distinguishing between simple association and actual causation, which requires both statistical skills and an acute, insightful mind. This section will guide you through this process step by step.

Concepts

First, let’s understand a few key concepts:

• Potential outcomes framework: This framework posits that for any individual and any intervention, potential outcomes correspond to each possible action. The causal effect is the difference between these outcomes, typically unobservable for the same unit under both states.
• Randomization: In this setting, randomization is an important concept. By randomly assigning subjects to treatment or control groups, randomization ensures that confounding variables are evenly distributed across groups, isolating...