Part 2 The adjustment formula in practice
You may encounter challenges when attempting to apply the adjustment formula in real-life scenarios. One critical requirement for using this formula effectively is that the positivity assumption is met. This assumption mandates that every version of the treatment is applied across all subpopulations. In chapter 5, you will discover methods to identify when the positivity assumption does not hold and strategies to navigate this issue.
Until this point, our discussion has primarily focused on situations where the treatment or decision variable is binary. Chapter 6 introduces linear models to deal with the effect of continuous variables. You will also discover that linear models provide insights into how correlation and causation propagate through graphs. This understanding will be extremely useful in the following chapter.
Another challenge is figuring out which variables to include in the adjustment formula, especially if your graph is complex. Chapter 7 tackles this issue by explaining the back-door criterion, a graphical criterion to pick the right variables for adjustment.