search

matheusfacure / python-causality-handbook

Causal Inference for the Brave and True. A light-hearted yet rigorous approach to learning about impact estimation and causality.
https://matheusfacure.github.io/python-causality-handbook/landing-page.html
MIT License
2.74k stars 475 forks source link

Chapter 13 #186

Closed jakubkaplan closed 2 years ago

jakubkaplan commented 2 years ago

There is an issue on chapter 13, in the following paragraph

Notice that is the baseline of the control. In our case, is the level of deposits in Florianopolis in the month of May. If we turn on the treated city dummy, we get . So is the baseline of Porto Alegre in May, before the intervention, and is the increase of Porto Alegre baseline on top of Florianopolis. If we turn the POA dummy off and turn the July dummy on, we get , which is the level of Florianópolis in July, after the intervention period. is then the trend of the control, since we add it on top of the baseline to get the level of the control at the period post intervention. As a recap, \beta_1 is the increment we get by going from the treated to the control, is the increment we get by going from the period before to the period after the intervention. Finally, if we turn both dummies on, we get . is the level in Porto Alegre after the intervention. So is the incremental impact when you go from May to July and from Florianopolis to POA. In other words, it is the Difference in Difference estimator.

It should be (change in bold):

Notice that is the baseline of the control. In our case, is the level of deposits in Florianopolis in the month of May. If we turn on the treated city dummy, we get . So is the baseline of Porto Alegre in May, before the intervention, and is the increase of Porto Alegre baseline on top of Florianopolis. If we turn the POA dummy off and turn the July dummy on, we get , which is the level of Florianópolis in July, after the intervention period. is then the trend of the control, since we add it on top of the baseline to get the level of the control at the period post intervention. As a recap, \beta_1 is the increment we get by going from the control to the treated, is the increment we get by going from the period before to the period after the intervention. Finally, if we turn both dummies on, we get . is the level in Porto Alegre after the intervention. So is the incremental impact when you go from May to July and from Florianopolis to POA. In other words, it is the Difference in Difference estimator.