Counterfactuals, Treatment Effects, and Identifying Assumptions
Does X cause Y?
“The fundamental problem of causal inference” (Holland 1986).
D_i = \begin{cases} 1 & i \text{ received treatment (e.g., aspirin)} \\ 0 & i \text{ received no treatment (e.g., no aspirin)} \end{cases}
Y_i^D = \begin{cases} Y_i^1 & \text{outcome if } i \text{ received treatment} \\ Y_i^0 & \text{outcome if } i \text{ did not receive treatment} \end{cases}
\delta_i = Y_i^1 - Y_i^0
ATE = \frac{1}{N}\sum_{i=1}^{N}(Y_i^1 - Y_i^0) = \mathbb{E}[Y_i^1] - \mathbb{E}[Y_i^0]
ATE = \mathbb{E}[Y_i^1] - \mathbb{E}[Y_i^0]
ATE = \mathbb{E}[Y_i^1 \mid D_i = 1] - \mathbb{E}[Y_i^0 \mid D_i = 0]
(Y^1, Y^0) \perp D
| \(i\) | \(D_i\) | \(Y_i\) | \(Y_i^1\) | \(Y_i^0\) | \(\delta_i\) |
|---|
| \(i\) | \(D_i\) | \(Y_i\) | \(Y_i^1\) | \(Y_i^0\) | \(\delta_i\) |
|---|---|---|---|---|---|
| 1 | 1 | 2 | 2 | 5 | −3 |
| 2 | 1 | 4 | 4 | 7 | −3 |
| 3 | 0 | 1 | 0 | 1 | −1 |
| 4 | 0 | 3 | 2 | 3 | −1 |
Causal inference is about identification strategy and research design, which depends on treatment assignment. Not fancy machine learning methods.