Reference no: EM132507577
Use GRETL and the data BEAUTY (uploaded in the campus) extracted from a famous paper by Hammermesh and Biddle (1994) that relates U.S. workers' wages with their beauty. See the description of the variables at the end of this document.
The qualitative variable looks ranks individuals' beauty from 1 to 5 being 1 the worst look and 5 the best look. Use this variable to construct the following dummy variables: belavg (below average) which will be equal to 1 if looks is lower or equal than 2 (and 0 otherwise) and abvavg (above average) which will be equal to 1 if looks is greater or equal than 4 (and 0 otherwise). Thus, the reference category is average looks (when looks = 3)
a) Find the proportion of workers in the sample that are classified as having above average looks and below average looks. In addition, find the proportions separately for men and women. Interpret.
b) Estimate the following model:
log(wage) = β0 + δ0belavg + δ1abvavg + u
separately for men and women, and report the results in equation form. In both cases, interpret the coefficient on belavg. Test whether belavg is statistically significant at 5% significance level in each of the two equations.
c) Is there convincing evidence that women with above average looks earn more than women with average looks? Explain.
d) For both men and women, add the explanatory variables educ, exper, union, goodhlth, black, married, south, bigcity, smllcity, and service to the regression in part (b). Estimate the same model using all the observations (all men and women). Present the estimation results in a Model Table [Hint: you should get three columns: ALL, WOMEN, MEN]
(i) Interpret the estimated coefficients on educ and exper in the pooled regression with all the observations.
(ii) Use the pooled regression to test whether the return to experience is statistically the same as the return to education [Hint: test on linear restrictions:
βedUc = βexper]. Interpret the result of the test.
(iii) Do the effects of the "looks" variables change in important ways in these two separate regressions with respect to the coefficients you got in part (b)? Interpret.
e) Use the Chow test at 5% confidence level using the general model in (d) to check for a structural break by gender. Interpret the result.
f) Use step-wise regression to find the best model to explain log-wages in this sample.
Attachment:- ACTIVITY for PARTICIPATION.rar