Page 562 - Ebook HTKH 2024

value: Chi-square = 270.67, degrees of freedom = 142, P value = 0.000. According to
Hair et al. (2010), Multivariate Data Analysis, 7th edition, CMIN/df = 1,450 ≤ 2 is good;

CFI = 0.949 ≥ 0.9 is good; GFI = 0.924 ≥ 0.9 is good; RMSEA = 0.052 ≤ 0.08 shows
that the model is consistent with market data. The above results confirm the
unidimensionality of the scales of product, price, promotion, place and purchase
decision.
Table 6. Standardized regression weights
Estimate

GP3b <--- GP3 0.741
GP3a <--- GP3 0.783
GP3c <--- GP3 0.733
GP3d <--- GP3 0.685

GP1b <--- GP1 0.793
GP1c <--- GP1 0.751
GP1d <--- GP1 0.667
GP1a <--- GP1 0.708
GP4c <--- GP4 0.758

GP4b <--- GP4 0.803
GP4a <--- GP4 0.692
GP4d <--- GP4 0.657

GP2b <--- GP2 0.764
GP2a <--- GP2 0.725
GP2c <--- GP2 0.751
GP2d <--- GP2 0.639
DB3 <--- DB 0.803

DB2 <--- DB 0.741
DB1 <--- DB 0.766
CR 0.736
AVE 0.731

MSV 0.73
The value of Standardized Regression Weights of all items in the questionnaire is
greater than the minimum value of 0.5, so all items are kept. Beside, in AMOS, there is
one more concept to confirm the reliability of the scale, which is the concept of
Composite Reliability (CR) with an assurance level of 0.7. “Hair, J., Black, W., Babin,
B., and Anderson, R. (2010). Multivariate data analysis (7th ed.): Prentice-Hall, Inc.

Upper Saddle River, NJ, USA.” also points that Average Extracted Variance (AVE) is
used to evaluate convergent validity and Maximum Shared Variance (MSV) is used to
measure discriminant validity.
According to the author's calculation, CR = 0.736 > 0.7, so the overall reliability
is guaranteed. AVE = 0.731 > 0.5, showing that the observed variable is correlated with
other variables in the same factor, that is, the latent variable is well explained by

554