Multiple Linear Regression Models: Exploring the Relationship Between Service and Customer Satisfaction Levels Regression : Variables Entered/Removeda Variables Variables Model Removed Entered 1 Method . Enter X4, X3, X2, X5, X1b a. Dependent Variable: Y b. All requested variables entered. ModelR Model Summary Adjusted R R Square Square a 1 ,616 ,379 Std. Error of the Estimate ,345 3,41908 a. Predictors: (Constant), X4, X3, X2, X5, X1 Description: • R = 0.616 means the relationship between X1, X2, X3, X4, X5 and Y is 61.6%. This means that the relationship is close. The bigger R means the closer the relationship. To ensure the type of relationship between variables can be seen in the following table. Relationship Table between Variables Value Interpretation 0.0 - 0.19 Not Very Tight 0.2 - 0.39 Not Tight 0.4 - 0.59 Close Enough 0.6 - 0.79 Erat 0.8 - 0.99 Very Tight • R square of 0.379 means 37.9% customer satisfaction can be explained by Reliability (X1), Assurance (X2), Tangible (X3), Emphaty (X4), Responsiveness (X5) while the remaining 62.1% can be explained by other factors not examined by this study. • Adjusted R Square of 0.345 means that 34.5% of customer satisfaction can be explained by Reliability (X1), Assurance (X2), Tangible (X3), Emphaty (X4), Responsiveness (X5) while the remaining 65.5% can be explained by other factors not examined by this study. There are two choices here, whether to use R Square or Adjusted R Square. If there are more than two variables, the Adjusted R Squareis used. • Standard Error of Estimated means that it measures the variation of the value that is predicted Standard Error of Estimated can also be called standard deviation. In the example the Standard Error of Estimated is 3.41908 Additional Information: • The table shows that Regression df = 5 means that the number of independent variables is 5, namely Reliability, Assurance, Tangible, Empathy, and Responsiveness. • Total df = 94 means the number of respondents/data is 95 (where df = n-1) while Regression df = 89 comes from 94-5. • Residual: the difference between the actual decision and the predicted decision. Residual = Total - regression. • Mean Square: Sum of Square/ df. In the example mean square regression (127,243) = 636,214/5 • SSE (sum of squares for error) shows the sum of squares for error meaning that if a regression equation is obtained from substituting the values of a and b which are minimizing the SSE, the least squares prediction regression line equation will be generated. =