The material for this part of the class is not covered in the
textbook.
Additional material
This section discusses (i) interpreting regressions under
unconfoundedness but with treatment effect heterogeneity and (ii)
alternative estimation procedures that may work better if there is
treatment effect heterogeneity.