Page 223 - Ebook HTKH 2024

Multiple regression analysis: To assess the suitability of the model, we use the
multiple linear regression method with 4 independent variables analyzed. The adjusted

R-squared value is 0.664, which is greater than 0.5, indicating that the 4 independent
variables included in the regression account for 66.4% of the variation in the dependent
variable, with the remaining 33.6% attributed to variables outside the model and random
error (see Table 12).
b
Table 12. Result of multiple regression analysis: Model Summary

Source: Data analyzed in SPSS 27.0
The Durbin-Watson (DW) statistic is used to test the autocorrelation of adjacent
errors. From the result above, DW = 1.976 falls within the range of 1.5 to 2.5,
indicating no first-order autocorrelation. The significance level of the F-test (Sig) is 0,
which is less than 0.05. Therefore, the multiple linear regression model is appropriate

for the dataset and can be used (see Table 13).
Table 13. Results of multiple regression analysis ANOVA

Source: Data analyzed in SPSS 27.0
From the correlation coefficient analysis results, the significance level (Sig.) is <
0.05, indicating that the independent variables are linearly correlated with the

dependent variables with a confidence level of 100%. With the data presented in this
table, the linear regression model provided is suitable for the data and can be used.
Therefore, the independent variables in the model are related to the dependent variable
MX. The Sig. coefficients of the independent variables are all less than 0.05, indicating
that these independent variables are all significant in explaining the dependent

variable. The influence of green promotions on online shopping decisions of
consumers in Vietnam has an impact with Beta coefficients in decreasing order: GX
(0.438), KMX (0.215), SPX (0.193), and PPX (0.147). Since all Beta coefficients are
positive, these independent variables have a positive effect on the dependent variable.
The Variance Inflation Factor (VIF) of all variables is less than 2, so multicollinearity
does not occur (see Table 14).

215