Suppose that you want to estimate the causal effect of a treatment on Income. There are two periods. Individuals in the treatment group are not treated in 1997 (when P ostt = 0) and are treated in 1998 (when P ostt = 1).
statistics
Description
Suppose that you want to estimate the causal effect of a treatment on Income. There are two periods. Individuals in the treatment group are not treated in 1997 (when P ostt = 0) and are treated in 1998 (when P ostt = 1).
a. Use the following sample data to fill out all 9 cells of the table below with E(Incomei
|P ost, T reat)
and then write out the equation Incomei,t = α + βP ostt + γT reati + δP ostt ∗ T reati + εi,t,
filling in the coefficient and intercept values.
b. What is the estimator ˆδ in a. called?
c. You obtain data on Income in the years leading up to 1997 for your two groups and
plot it in the graph below. What about the graph suggests that δ is not valid (i.e. leads to
biased estimates)?
