Using the Excel data set

CPS08opens in a new tab

(describedopens in a new tab),

run a regression of average hourly earnings

(AHE)

on age and answer the following questions.The coefficient on age shows the

A.

the change in hourly earnings for every 1-year increase in age

B.

the change in hourly earnings for every 5-year increase in age

C.

the change in age for every $5.00 increase in average hourly earnings

D.

the change in age for every $1.00 increase in average hourly earnings

Given the following hypothesis:

H0:

age

= 0 where

age

is the coefficient on age you would:

A.

Not reject

H0

and conclude that the coefficient on age is statistically insignificant at the 5% level.

B.

Reject

H0

at the 5% level because the t-statistic is greater than 1.96

C.

None of the above

D.

Not reject

H0

and because the coefficient on age is so small in magnitude

The 95% confidence interval associated with the null hypothesis:

H0:

age

=0 is

A.

-1.969 - 4.133

B.

0.204 - 0.489

C.

0.039 - 0.605

D.

0.527 - 0. 683

Given the following hypothesis:

H0:

age

=0.50 where

age

is the coefficient on Age you would:

A.

Not reject

H0

because the 95% confidence interval includes 0.50

B.

Reject

H0

because the t-statistic associated with the null is 2.63

C.

Not reject

H0

because the estimated coefficient on age is close to 0.50

D.

There is not enough information to test the null hypothesis

Based on the regression the expected average hourly wage for a person 20 years of age would be:

A.

Would be $21.65 which is obtained by multiplying the intercept by 20

B.

Would be $1.32 which shows how poorly age explains earnings

C.

Would be $13.18 which is plausible for a 20 year old

D.

Would be $14.82 which is more than a person 25 years of age would earn.

Re-run the regression in (1) but estimate it only for women (Hint: the number of observations in the regression should be 3,336). The results indicate that

A.

none of the answers in (a)-(c) are correct.

B.

the coefficient on age is 0.346 but it is not statistically significant at the 95% confidence level

C.

the average hourly earnings of women is $7.27 less than males

D.

the average hourly earning of a woman 30 years of age is $3.46 more than a woman 20 years of age

Re-run the regression in (1) but now estimate it only for men (Hint: the number of observations in the regression should be 4,375). The results indicate that

A.

the expected average hourly earnings for a man 30 years of age is $20.38

B.

the coefficient on age is 0.78 but it is not statistically significant at the 95% confidence level

C.

men actually earn $3.05 less per hour than women

D.

every year of age increases a man's hourly earning by $7.81

Based on the separate regressions for men and women we would conclude that

A.

the expected average hourly earnings for a man 30 years of age is $20.38 whereas the expected hourly earnings for woman of the same age is $15.66

B.

the negative intercept in the regression for males means they earn $3.05 less than women as age increases

C.

women earn slightly more than men for every year of age

D.

the expected hourly earnings of men increases by $0.43 more than women for every one year increase in age

(Challenging question) In Exercise 5.15, Stock and Watson describe a means by which one could statistically test whether the coefficient on age in the female regression is different from the coefficient on age in the male regression. Following their approach, the 95% confidence for the difference of the two coefficients would be

A.

0.588 - 0.281

B.

0.112 - 0.740

C.

0.672 - 0.890

D.

0.238 - 0454

(Challenging question) Using the results in part (9), test the following null

hypothesis,H0:

age,male

-

age,

female =0.

A.

Reject the null because the 95% confidence interval includes zero

B.

Reject the null because the 95% confidence interval does not include zero

C.

Do not reject the null

D.

Do not reject the null because the difference in the two coefficients is only 0.433