In this section, the analysis of validity includes three parts: validity of CSR measurement scale, validity of brand image measurement scale and validity of brand loyalty measurement scale. Based on the output of SPSS (details in Appendix F), we observed that there are three main steps to examine the validity. First, looking at the values of KMO and Bartlett's Test. Then, based on the Total Variance Explained table to determine the numbers of extracted factors. In the end, based on the Rotated Component Matrix table to observe the value of each variable factor loadings (a rule of thumb that factor loadings should greater than 0.30) to evaluate whether or not the extracted factors are consistent with the dimensions of the scales.
4.5.1 Validity of CSR scale
Firstly, after using the EFA method to analyze the 24 questions of the CSR part, we can see that the results of KMO and Bartlett's Test are shown in Appendix F: Output of validity analysis (CSR part). The KMO value is 0.896 (between 0.8 and 0.9), which falls into the range of being superb. The Bartlett's Test Chi-Square value is 1787.900 (df=276) and the significant level at 0.000, which means the R-matrix is not an identity matrix and. Therefore, these 24 questions suitable for factors analysis and there are some relationships between …show more content…
After extraction, factor 1 accounts for 35.721% of variance (contributes to the most variance, as each successive factor will account for less and less variance), factor 2, factor 3 and factor 4 accounts for 7.759%, 5.660% and 5.342% respectively. Moreover, the eigenvalues of the factors after rotation are displayed at the column called “Rotation Sums of Squared Loadings” that factor 1 only makes up 5.338% of variance (compared to 2.691%, 2.618% and 2.428%
4.5.1 Validity of CSR scale
Firstly, after using the EFA method to analyze the 24 questions of the CSR part, we can see that the results of KMO and Bartlett's Test are shown in Appendix F: Output of validity analysis (CSR part). The KMO value is 0.896 (between 0.8 and 0.9), which falls into the range of being superb. The Bartlett's Test Chi-Square value is 1787.900 (df=276) and the significant level at 0.000, which means the R-matrix is not an identity matrix and. Therefore, these 24 questions suitable for factors analysis and there are some relationships between …show more content…
After extraction, factor 1 accounts for 35.721% of variance (contributes to the most variance, as each successive factor will account for less and less variance), factor 2, factor 3 and factor 4 accounts for 7.759%, 5.660% and 5.342% respectively. Moreover, the eigenvalues of the factors after rotation are displayed at the column called “Rotation Sums of Squared Loadings” that factor 1 only makes up 5.338% of variance (compared to 2.691%, 2.618% and 2.428%