Please use the data file called "PS3 Wage Data" (available in .csv, .txt, or .gdt, depending on which program you choose to use) on Smartsite. This file contains data on wages (w), years of schooling (ed), and years of work experience (ex) for a random sample of 1,289 individuals, taken from the 1995 Current Population Survey by the U.S. Census Bureau. The variables in this data set include, in order by column:

w (wage)

fe (female indicator variable: 1 if person i is female and 0 if male)

nw (nonwhite indicator: 1 if nonwhite and 0 if white)

ed (years of schooling)

ex (years of potential experience)

Assuming that the multiple regression assumptions MR1-MR5 hold, estimate a multiple regression model of the following form:

1. Please estimate the equation and report your regression results (parameter estimates, standard errors, and t-statistics) in a table.

2. Based on this regression, what is the marginal return to schooling? To experience?

3. Your dropout friend claims that a year of schooling couldn't have more than a $1.25 effect on wages. Test this null hypothesis at the 95% confidence level. Will this be a one or two-tailed test? Why?

Back in 1974, an economist named Jacob Mincer proposed the following wage equation:

The Mincer equation has since become one of the most widely used equations in econometrics.

4. Which variables will you need to transform to estimate this equation? Write the transformations you will need to do to those variables.

5. Estimate the Mincer wage equation, and report your results in table form.

6. In this specification, what is the interpretation of b1? Is it different than what you got in Question 1? Why or why not?

7. Suppose you think there might be discrimination with respect to gender and race/ethnicity in the U.S. labor market. Write a model that you can use to test whether wages are lower for women and/or non-whites, controlling for education and experience.

8. Estimate your model and report your results in a table.