4.3. Factor analysisAn exploratory factor analysis was conducted to determine a reduced set of underlying factors. With he rule counting the eigenvalues greater than 1, the lower bound is four factors (n=4, eigenvalue= 1.02). In order to find the upper bound, the Maximum Likelihood method was used to test the hypothesis: n factors are sufficient (vs. the alternative hypothesis: more factors are needed). At n=15, it has XZ=284, p=0.0222. At n=16, it has Xz=244, p=0.0843. Therefore, the upper bound is set at 16 factors. Within the upper and lower bounds, this research used the SMC (Squared Multiple Correlation) as the prior communality estimate, the Principal Component method for extracting factors, and both the OrthogonalVarimax and Oblique Promax as the rotation methods to try different cases. The final factorstructure is shown in Table 7 with n----7 and Oblique Promax. It satisfies several criteria - total variances explained as 96.8%, variances explained by each factor greater than "average variable" (1/38), and scree test criterion. These seven aggregated factors are named as the following: