Does the return to education differ by race and gender? For this exercise, use the file cps4.dat. (This is a large file with 4,838 observations. If your software is a student version, you can use the smaller file cps4_small.dat.) In this exercise you will extract subsamples of observations consisting of (i) all males, (ii) all females, (iii) all whites,

(iv) all blacks, (v) white males, (vi) white females, (vii) black males, and (viii) black females.

(a) For each sample partition, obtain the summary statistics of WAGE.

(b) A variable’s coefficient of variation is 100 times the ratio of its sample standard deviation to its sample mean. For a variable y, it is CV ¼ 100

sy y

It is a measure of variation that takes into account the size of the variable. What is the coefficient of variation for WAGE within each sample partition?

(c) For each sample partition, estimate the log-linear model lnðWAGEÞ ¼ b1 þ b2EDUC þ e What is the approximate percentage return to another year of education for each group?

(d) Does the model fit the data equally well for each sample partition?

(e) For each sample partition, test the null hypothesis that the rate of return to education is 10% against the alternative that it is not, using a two-tail test at the 5% level of significance.