Econometrics

Lab 3.

Part 1.

Open the data set mhctuition from the courses/mirobins/econ320 folder. This data set contains information on Mount Holyoke Net Tuition and GDP for the U.S. running from 1947 to 2000. If you think about Mount Holyoke sales as being primarily determined by national income you might estimate a regression of net tuition on gdp. Estimate this regression.

1. How much of variation in Mount Holyoke’s net tuition revenue is explained by GDP?______

2. If the economy enters a recession and real gdp declines by 5 percent, what would happen to net tuition revenue at Mount Holyoke?______

In the late 1960’s and early 1970s many of the formerly male colleges went coed, by 1975 most of the formerly male colleges were coeducational. It seems likely that during the relationship between income and net tuition revenue for Mount Holyoke would be affected by this change of events. Create a coeducation dummy variable that is equal to 1 for the years 1976 to 2000 and 0 otherwise. To do this use the transform -- recode – into different variables option. Choose the variable year to recode and give SPSS a name for your new variable. Use the old and new values button to create the recoding. It probably makes sense to use the minimum through and value through maximum options. After you create the new variable rerun the regression you ran before including the coeducation dummy variable.

3. What happened to net tuition revenue after coeducation?______

The regression you just ran implies an intercept shift, but it is possible that both the intercept and slope are different. Use the data tab, split file, compare groups option to split the file by your coeducation variable. Once the file is split every analysis you do will be done for both groups. Rerun the regression of net tuition on real gdp. Be sure and remove the coeducation dummy from the regression. Use the results of these two regressions, along with the results of the first regression you ran to test the hypothesis that the regression line is different after 1975 from before 1975. Note: You will need the Sum of Squared Residuals from both of these regressions and from the original equation you estimated.

F-Statistic______

Is this significant?______

Part B. Use the SPSS dataset beauty.sav for this problem. Estimate the following model.

Lwage=f(exper, goodhlth,educ,union) then a second model also adding (looks, female, married).

The first model contains only productivity variables, while the second model contains both productivity and personal attributes. Do the joint F-test of whether the additional variables in the second model are significant as a group. You can get SPSS to do this automatically by putting the added variables in block 2 using next block of the regression and under the statistics button asking for change in R squared.

1. wage hourly wage

2. lwage log(wage)

3. belavg =1 if looks <= 2

4. abvavg =1 if looks >=4

5. exper years of workforce experience

6. looks from 1 to 5

7. union =1 if union member

8. goodhlth =1 if good health

9. black =1 if black

10. female =1 if female

11. married =1 if married

12. south =1 if live in south

13. bigcity =1 if live in big city

14. smllcity =1 if live in small city

15. service =1 if service industry

16. expersq exper^2

17. educ years of schooling

F and p-value for the test of whether the extra coefficients are significant.

B. Do looks matter the same for everyone? Estimate the simple model log(wage)=f(looks).

Then pick one other dummy variable that you think might influence the effect of looks on wages. Split the file by this variable (Data, Split File, Compare Groups) and rerun the regression for the two groups you have chosen. Briefly explain you made the choice you did and list the F-Stat for your test and give the estimated coefficients.

Group 1 ______Looks Coeff.______

Group 2 ______Looks Coeff.______

F-Statistic.______(is this significant)______

Discussion.