Instruction: Describe the potential outcomes framework and discuss how it aids in understanding causal effects.
Context: This question tests the candidate’s knowledge of fundamental concepts in causal inference and their application in modeling.
Official answer available
Preview the opening of the answer, then unlock the full walkthrough.
The potential outcomes framework, often attributed to the work of Neyman in the 1920s and later expanded by Rubin in the 1970s, is a foundational approach for thinking about causal inference. It's a way to conceptualize the effect of an intervention or treatment on an individual or a system by considering all possible outcomes—those that happen as a result of the treatment and those that would have happened in the absence of the treatment.
In practical terms, for each unit (which could be a person, a geographic area, etc.), we imagine two potential outcomes: one if the unit receives the treatment (let's call this (Y_{i}(1))) and one if it doesn't ((Y_{i}(0))). The causal effect of the treatment on the unit is then the difference between these two potential outcomes, (Y_{i}(1) - Y_{i}(0)). However, the fundamental problem of causal inference is that we can never...