*…show more content…*

First Iteration

The statistics for regression between satisfaction level and usage level are as follows

From the above tables we can see that the r2 value is 0.505. This value shows that with the current regression equation we can estimate 50.5% total variation in the dependant variable.

The total sum of squares is equal to 72.466 which denote the squared error that would occur if used only the mean of the independent variable was used to predict satisfaction level. Using usage level we are able to reduce the this error by (36.602/72.466 = 0.5005) 50.05 %.

The collinearity statistics provide a perspective on the impact of collinearity on the independent variable in the equation. The tolerance value is the amount of an independent variable’s predictive capability that is not predicted by the other independent variables in the equation. Here the tolerance level of 1 shows that the it is totally unaffected by the other independent variables.

Second iteration

To identify the next variable that can be entered into the equation we calculate the partial correlations of the remaining variables with the independent variables taking the variable already in the equation as control variable.

From the above table, looking down the first column we see that the variable manufacturer image has the maximum correlation at the