Collinearity Coefficient in Business Models: Assessing Variable Interrelationships in the Digital Economy Coefficients Statistics ModelB (Constant) Std. Error 24.160 3.519 Brand Image (X1) -.049 .132 Customer Satisfaction (X2) .024 .139 Beta t Sig. Tolerance VIF 6.865 .000 -.042 -.372 .711 .913 1.095 .019 .170 .865 .913 1.095 a. Dependent Variable: Switching Barrier (Y) Conclusion of Multicollinearity Test above: X1 tolerance value 0.913> 0.100, X2 tolerance value 0.913> 0.100 and for X1 VIF value 0.913 < 10.00 and X1 VIF value 0.913 < 10.00. 10.00, meaning there are no symptoms of multicollinearity. The basis for decision making for the heteroscedasticity test is scatterplots: According to Iman Ghozali (2011: 139) there is no heteroscedastistas, if there is no clear pattern (wavy, widening then narrowing) in the scatterplots image, and the points spread above Conclusion of the heteroscedasticity test above: Based on the graph above, there is heteroscedastistas, because the dots spread then narrowed. Another alternative can be used with the Glejser test. Durbin Watson autocorrelation test decision basis. Model Summaryb Adjusted R ModelR 1 .051a R Square .003 Square -.020 Std. Error of the Estimate DurbinWatson 2.91979 1.777 a. Predictors: (Constant), Customer Satisfaction (X2), Brand Image (X1) b. Dependent Variable: Switching Barrier(Y) K = 2, n = 90 with sig. 5% du = 1.703; durbin Watson = 1.777 4-du = 2.297 Conclusion: Du 1.703 < durbin Watson 1.777 < 2.297 4-du, then there are no auto-correlation symptoms.