5. Data analysis
As mentioned in the previous section, SPSS is used to analyze data for this research. A stepwise multiple regression analysis was used to prove the significance of the variables. To avoid violating the basic assumptions underlying the method of least squares used by the classical linear regression model, we conducted a P–P plot for assessing the assumption of normality. The plot showed that the quantile pairs fell nearly on a straight line. It is, therefore, reasonable to conclude that the data used in this research are approximately normal. Second, this research used the condition index (C.I.) to assess the multicollinearity among independent variables in the model. The value of 29.44 indicated no severe multicollinearity problem among the regressors. Finally, we used the Durbin-Watson statistic for detecting serial correlation. The value of 1.89 (less than 2) indicated the autocorrelation problem does not exist (Gujarati, 2003).