In Section 17.1 we estimated a regression model that tested for gender and age discrimination in salaries. We found that the number of years of experience x and the gender dummy d 1 were significant in explaining salary differences; however, the age dummy d2 was insignificant. In an attempt to refine the model explaining salary, we drop d2 and estimate three models using the data from Table 17.1, where y represents annual salary (in $1,000s).

Model 1. y = β0 + β1 x + β 2d 1 + ε

Model 2. y = β 0 + β 1 x + β 2 x d 1 + ε

Model 3. y = β0 + β1 x + β2 d 1 + β3 x d 1 + ε

a. Estimate and interpret each of the above three models.

b . Select the most appropriate model, based on an objective model selection criterion.

c. Use the selected model to predict salaries for males and females over various years of experience.