An international management consulting company develops multiple-regression models for executive salaries of its client firms. The consulting company has found that models which use the natural logarithm of salary as the dependent variable have better predictive power than those using salary as the dependent variable. * A preliminary step in the construction of these models is the determination of the most important independent variables. For one firm, 10 potential independent variables (7 quantitative and 3 qualitative) were measured in a sample of 100 executives. The data, described in Table 6 , are saved in the EXECSAL file. Since it would be very difficult to construct a complete second-order model with all of the 10 independent variables, use stepwise regression to decide which of the 10 variables should be included in the building of the final model for the logarithm of executive salaries.