Exercise

Oregon Health Experiment Data: Finding an ATE

Now that we've checked for balance between our treatment and control, let's check to see the effect of Medicaid on some of our medical outcomes of interest, like the patients' blood pressure. A person's blood pressure is measured by two numbers, e.g. 114/71, where the first number is the "systolic" blood pressure and the second is the "diastolic" blood pressure. These are counted in the OHIE dataframe through two variables: systolic pressure in bp_sar_inp and diastolic pressure in bp_var_inp.

Now let's check the treatment effect on systolic blood pressure (bp_sar_inp) by comparing systolic blood pressure in the treatment and control groups. To check for statistical significance, let's look at p values of the t-test. A p-value below .05 suggests that a result is statistically significant.

Instructions

100 XP
  • 1) Manually calculate the average treatment effect on systolic blood pressure by subtracting the mean value for systolic blood pressure in the treatment group minus the mean value for systolic blood pressure in the control group.
  • 2) Use a t.test to determine whether the treatment and control groups have significantly different average values for systolic blood pressure (variable bp_sar_inp).
  • 3) After looking at the p-value of the t-test, can we say the t-test results are statistically significant? Answer "yes" or "no".