Exercise

Practice Using IV to Solve Noncompliance in the OHIE

In an earlier course, you analyzed data from the Oregon Healthcare Insurance Experiment (OHIE) where you assumed that everyone who was offered Medicaid coverage ended up getting covered. Now let us relax this assumption and instead assume that not everyone who was offered Medicaid coverage actually enrolled. This is a more realistic model of what actually happened, and it plays into the strengths of the Instrumental Variables method.

So we will use assignment to the treatment group as an instrument, and we will check to see if that instrument shows us whether Medicaid had any effect on one of our health outcomes of interest, the number of people who were identified as having depression during a health screening.

Instructions

100 XP
  • 1) Explore the data on your own, using functions like head(), str(), and summary()
  • 2) Compute the Intention-to-Treat effect of being offered insurance on having a positive depression screening
  • 3) Estimate the 2SLS effect of taking up the offer on positive depression screening using two separate calls to lm()