Assessment of multicollinearity among predictor variables
Multicollinearity makes it difficult to determine which predictor variable is important in a regression equation. Therefore,before further analysis,multicollinearity among predictor variables was assessed. Pearson's correlation was computed among variable in the model to determine: 1)the linear relationship among predictor variables and self-management,and 2)multicollinearity among the independent varibles.The correlations among variables were shown in Appendix H. The results showed that correlations among the predictor variables ranged from .o2 to .61, indiceting low to moderate multicollinearity. Furthermore, two indices for diagnosing multicollinearity,namely,the tolerances and variance inflation factors (VIF) were tested. Tolerance is define as "the proportion of variability of that is not explained by its linear relationships with the other independent variables in the model".Tolerance values can range from 0.00
(perfect multicollineaity)to 1.00 (no multicollinearity). It was suggested that multicollineraity existed if a predictive variable has a tolerance of 0.1 or less. In this study, tolerance values in the regreession equations among predictor variables range from 0.46 to 0.94 as shown in Appendix I indicating no problem with multicollinearity. Variance inflation factor (VIF) is defined as the