4.14 How much does education affect wage rates? This question will explore the issue further. The data file cps_small.daf contains 1000 observations on hourly wage rates.

education, and other variables from the 1997 Current Population Survey (CPS).

(a) Construct histograms of thewAcBvariable and its logaritbn ln(WAGE). Which appears more normally distributed?

(b) Estimate the linear regression WAGE:9r * !2EDUCle arrd log-linear regression LI(WAGE):9r * P4EDUC *e. What is the estimated return to education in each model? That is, for an additional year of education, what percentage increase in wages can the average worker expect?

(c) Construct histograms of the residuals from the linear and log-linear models in

(b), and the Jarque-Bera test for normality. Does one set of residuals appear more compatible with normality than the other?

(d) Compare the Rz of the linear model to the "generalized" R" for the logJinear model. Which model fits the data better?

(e) Plot the least squares residuals from each model against EDUC. Do you observe any patterns?

(f) Using each model, predict the wage of a worker with 16 years of education.

Compare these predictions to the actual average wage of all workers in the sample with 16 years of education.

(g) Based on the results in parts (a)-(0, which functional form would you use?

Explain.