Table 6 shows for a group of companies the relation between their profits and some explanatory variables.

a. Run a multiple regression with ANOVA table in all three independent variables, and using the global F-test, test simultaneously the hypothesis that all the slope coefficients are jointly equal to zero: H0 : b1 ¼ b2 ¼ b3 ¼ 0:

b. Conduct a test of hypothesis for the individual regression coefficients and then, of these, drop the one which is not appearing statistically significant using the ANOVA table.

c. Run now a new regression and compare the ANOVA table versus the one of point

a. d. As we have seen the assumptions for regression also require that the residuals remain constant for all predicted values Yb of the dependent variable. This condition is called homoscedasticity. A way to check for this is to plot and evaluate the behavior of residuals versus the predicted values. Do this for the second regression in two independent variables and draw conclusion about the homoscedasticity.

Transcribed Image Text:

TABLE 5 Year Quarters Parent Revenues Branch Revenues 1 132 62 Y1 Y2 2345678 150 63 135 60 156 66 155 70 145 67 139 65 152 68 9 178 74 10 160 65 Y3 11 168 70 12 180 72