Block 5: Regression and causation

Week 8 · Prediction II · Imai §4.3 · about 4 min

NoteHow this works

Run each cell with Run Code, look at the output, then answer the question at the end of the block. Every cell is editable; Start Over resets it. Errors are information: read them.

With a binary X, the regression slope is the difference in means. With a randomized treatment this difference is the average treatment effect.

1. A binary X: women as village council heads (India)

2. A categorical X: the GOTV experiment

(300,000 rows: loading takes a few seconds.)

Conclusion: In the GOTV regression, which group is the base level (intercept)?
R uses the first level (alphabetical): Civic Duty. The intercept is its mean (0.3145); the other coefficients are differences from Civic Duty, not from Control. The effect of Neighbors vs. Control is 0.0634 - (-0.0179) = 0.081. With random assignment this is a causal effect.