2. STATA Exercise: The data set "NSW.dta" on Courseworks contains observations from The National Supported Work Demonstration (NSW), which was a temporary employment program designed to help disadvantaged workers lacking basic job skills move into the labor market by giving them work experience and counseling in a sheltered environment. We will work with variables re75 - real yearly earnings in 1975, and education - years of education, both measured for a worker in the sample.

(a) (4 points) Suppose you were interested in measuring the effect of education on real yearly earnings. Before running a regression, what do you expect the sign of the coefficient 1 of the regression to be? How about the sign of 0?

(b) (6 points) Estimate the relationship between years of education and real yearly earnings in 1975. Show the STATA command and its output. Interpret the coefficients of this regression.

(c) (4 points) Calculate a 99% confidence interval for the effect of education on real yearly earnings in 1975.

(d) (4 points) Add a quadratic term to the regression above. That is, create a new variable educ2 which takes the value of the squared years of education and add it to the regression. Show the STATA command and its output.

(e) (6 points) What is the marginal effect of education on real yearly earnings in 1975? That is, what is d y , where y is de

the yearly earnings and e is years of education? What is the value of this derivative when workers have 10 years of education? How about when they have 16 years of education?