Part 3 Other strategies beyond the adjustment formula
Parts 1 and 2 of this book focus on understanding and applying the adjustment formula. But this formula requires you to know a lot about confounders. Sometimes, figuring out all the confounders can be challenging. When that happens, you may need to use different strategies.
A simple yet effective trick is the instrumental variables method. This approach, as explained in chapter 9, derives causal estimations by taking advantage of the presence of an independent source of variation.
Chapter 10 introduces the potential outcomes framework, another way to think about cause and effect. This approach is important because many methods are based on it, especially those we discuss in chapter 11. In this chapter, you’ll learn about techniques for time series data. These techniques don’t require you to know all the confounders because they work on stronger assumptions. Specifically, you’ll learn about synthetic controls, regression discontinuity design, and differences in differences.