What’s it good for?

• Prediction: For example, predict future vote (with some uncertainty!) given the economy.

• Exploring Associations: For example, the slope above summarizes the impact of the economy on election outcomes. A 1 percent improvement in the economy results in (on average) a 3% improvement in the vote for the incumbant.

• Extrapolation: For example, pre-election polls can be adjusted for factors like party identification to extrapolate to the voting population at large. (Chapter 17.1)

• Casual Inference: Estimating treatment effects. For the election example, we can imagine looking at the effect some law (for example tax cuts) on election outcomes. (Chapter 19)