Collinearity Analysis
Collinearity
The final test is testing for collinearity. This tests for variables that can be linearly predicted with high accuracy. In order for the independent variables to pass this test they may not be correlated more than +/- 0.90. If they happened to be correlated this much or great the standard error would be greatly inflated.
Table 2: Collinearity of Variables After evaluating the table, one issue arises. There is a collinearity issue between TV and Internet. The correlation is .91 which shows that the variables are highly correlated. To choose which variable should be kept, two regressions must be run, one without each variable. When evaluating these regressions, the variable that leads to the higher R Square value needs to …show more content…
Table 3 results in an R Square of .569974 whereas Table 4 results in an R Square of .581708. This shows that the independent variable Internet needs to be removed because it results in a higher correlation than TV.
Part C
After careful analysis of the data during part B, the remaining variables are: TV, Gas Price, Ave. # of Sales people, Ave. Rebate, and Corporate Sponsored Sales Promo. Each of these independent variables passed all five hurdles and can move on to the final model. It is still imperative that the P- values be analyzed.
Table 5: Regression Analysis Coefficients Standard Error t Stat P-value Intercept 278.028 10.081 27.580 0.000 TV ($1000s) 0.935 0.178 5.239 0.000 Gas Price ($1s/gallon) (6.283) 3.706 (1.695) 0.094 Ave. # of Sales people (0.351) 1.132 (0.310) 0.757 Ave. Rebate 3.763 0.532 7.078 0.000 Corporate Sponsored Sales Promo 25.072 3.917 6.400 …show more content…
First the intercept coefficient of 264.01 shows that if Tommy does not spend any money on TV advertising, Ave. Rebate, or Corporate Sponsored Sales Promotion that he will still sell 264.01 boxes of brakes that month. The first independent variable is the TV advertising sales, where for every $1,000 Tommy spends on TV advertising he will sell 0.98 more boxes of brake pads. The next variable is average rebate, where for every dollar in planned rebate Tommy will sell 3.38 boxes of brakes that month. Lastly, if there is a corporate sponsored sales promotion that month Tommy will sell 24.53 more boxes of
The final test is testing for collinearity. This tests for variables that can be linearly predicted with high accuracy. In order for the independent variables to pass this test they may not be correlated more than +/- 0.90. If they happened to be correlated this much or great the standard error would be greatly inflated.
Table 2: Collinearity of Variables After evaluating the table, one issue arises. There is a collinearity issue between TV and Internet. The correlation is .91 which shows that the variables are highly correlated. To choose which variable should be kept, two regressions must be run, one without each variable. When evaluating these regressions, the variable that leads to the higher R Square value needs to …show more content…
Table 3 results in an R Square of .569974 whereas Table 4 results in an R Square of .581708. This shows that the independent variable Internet needs to be removed because it results in a higher correlation than TV.
Part C
After careful analysis of the data during part B, the remaining variables are: TV, Gas Price, Ave. # of Sales people, Ave. Rebate, and Corporate Sponsored Sales Promo. Each of these independent variables passed all five hurdles and can move on to the final model. It is still imperative that the P- values be analyzed.
Table 5: Regression Analysis Coefficients Standard Error t Stat P-value Intercept 278.028 10.081 27.580 0.000 TV ($1000s) 0.935 0.178 5.239 0.000 Gas Price ($1s/gallon) (6.283) 3.706 (1.695) 0.094 Ave. # of Sales people (0.351) 1.132 (0.310) 0.757 Ave. Rebate 3.763 0.532 7.078 0.000 Corporate Sponsored Sales Promo 25.072 3.917 6.400 …show more content…
First the intercept coefficient of 264.01 shows that if Tommy does not spend any money on TV advertising, Ave. Rebate, or Corporate Sponsored Sales Promotion that he will still sell 264.01 boxes of brakes that month. The first independent variable is the TV advertising sales, where for every $1,000 Tommy spends on TV advertising he will sell 0.98 more boxes of brake pads. The next variable is average rebate, where for every dollar in planned rebate Tommy will sell 3.38 boxes of brakes that month. Lastly, if there is a corporate sponsored sales promotion that month Tommy will sell 24.53 more boxes of