5. Results
Table 2 shows that the Cronbach's alpha coefficients (Becker, G. ,2000) for all constructs are above 0.9, indicating a high level of internal consistency reliability. The Cronbach's alpha coefficient of the overall scale is also high, at 0.953, indicating the reliability of the measurement results.
Table 3, shows the basic statistical characteristics of each variable. The minimum, maximum, mean and standard déviations are given. In data preprocessing, descriptive statistical analysis is performed on the data using SPSS. This helps to understand the basic characteristics and distribution of the data.
Table 4 presents an overview of the participants' basic information. Including age, gender, and marital status. In terms of age distribution, the proportion of participants aged 25-34 was the highest, reaching 41.73%; In terms of gender, 56.2% are women; Married accounted for 58.27% of the marital status
The Kaiser-Meyer-Olkin (KMO) (Cerny, C.A., & Kaiser, 1977; Shrestha, 2021). measure assessed sample readiness for factor analysis. Values closer to 1 indicate high variable correlations, indicating factor analysis suitability. This study's KMO value was 0.939, which is suggesting that the data is suitable for factor analysis.
Table 5 presents the factor loading, reliability and average variance extraction of each variable, which helps to evaluate the validity and reliability of the variable measurement.
Table 6 shows the correlations between the main components. The correlation between Spendception and Impulse Buying is 0.626, and CPB is 0.559, however correlation between Impulse Buying and CPB is 0.54, indicating that there is a positive correlation among components (
Table 6).
Table 7 presents the common degree value of each variable, which is used to measure the degree to which the variance of a variable can be explained by common factors. The initial common degree of all three variables is 1, and the common degree of extraction is between 0.6481 and 0.798. It shows that these variables can be effectively explained by common factors to some extent.
Table 8 evaluates the goodness of fit of structural equation models. Several fit metrics such as CFI, NFI, IFI, TLI, GFI, RMSEA, Chi-square and
were used (Rigdon, E. E. ,1996). The values of CFI, IFI and TLI were 0.95, 0.950 and 0.942, respectively, meeting the good fitting criteria of 0.9. NFI was 0.878, GFI was 0.858 and RMSEA was 0.061, which met the corresponding fitting criteria. Chi-square is 1.605, which also meets the good fitting requirement of 3, indicating that the overall fitting effect of the model is good. This indicates that the research is reasonable and effective in setting the relationship between the variables related to consumer behavior. Based on this model, the internal relationship among Spendception, Impulse Buying and Consumer Purchase Intention can be further analyzed and explained.
Table 9, presents the evaluation metrics for the model's performance on both the training and test data sets. The Mean Squared Error (MSE), Mean Absolute Error (MAE), and R-squared (R²) values are provided to assess the accuracy and generalization ability of the model. As shown in
Table 8, the model demonstrates excellent predictive power with R-squared (R²) values of 0.9792 for both the training and test sets, indicating that the model explains approximately 97.92% of the variance in both datasets. The Mean Squared Error (MSE) for the training set (0.0092) is slightly lower than that for the test set (0.0123), which is typical and suggests the model's good generalization to new data. Similarly, the Mean Absolute Error (MAE) values for the training (0.0293) and test (0.0344) sets are close, supporting the conclusion that the model does not overfit and maintains accuracy across different data subsets. The low MSE and MAE values indicate that the model's predictions are close to the actual values, with only a slight increase in the error when applied to unseen test data, suggesting good generalization."
Figure 2 shows a direct and indirect path relationship model between Spendception, Impulse Buying, and Consumer Purchase Behavior.
Table 10 is presenting the direct effects of independent variables on dependent variable.
H1: Spendception → Impulse Buying, the path coefficient (β-value) is 0.47, and the p-value is ***. The result is Accepted. This suggests that Spendception has a significant positive impact on Impulse Buying
H2: Impulse Buying → Consumer Purchase Behavior, path coefficient is 0.544 and p-value is 0.029, and the result is accepted. This suggests that Impulse Buying has a significant positive impact on Consumer Purchase Behavior.
H3: Spendception → Consumer Purchase Behavior, the path coefficient is 0.15, and the p-value is ***, and the result is accepted. This indicates that Spendception has significant positive impact on Consumer Purchase Behavior.
Table 11, shows Spendception
IB
CPB (indirect effect) path coefficient is 0.252, p-value is***, the result is accepted. This suggests that Impulse Buying plays a partial mediating role between Spendception and Consumer Purchase Behavior.
Direct Effect of Interaction term on IB is not significant. This implies that male has no direct moderating influence on the association between IB and Spendception. This
Table 12 also examines how various degrees of male affect the indirect impacts of Spendception on CPB through IB:
Low male: The indirect impact is 0.087 with a SE of 0.086 at low levels of male, and it is significant at the p < 0.001 level (shown by ***). This indicates that when male is low, the relationship between Spendception, IB, and CPB is robust and substantial.
Medium male: The indirect impact rises marginally to 0.158 with a SE of 0.051 at medium levels of male, and it is still significant with a p-value of 0.003. This implies that, despite being weaker than at low male, the mediation pathway is nonetheless important even at medium male.
High male: The indirect impact is 0.23 with a SE of 0.099 and a p-value of 0.099 at high levels of male, which is insignificant. This suggests that the mediation pathway weakens and moderation is start playing its role however, Index of Moderate Mediation with p-value of 0.578, the SE is 0.035, and the Moderated Mediation Index is -0.003, all of which are not significant. This implies that this model does not support the idea of moderated mediation.
With an estimate of 0.034, SE of 0.014, and a critical ratio of 2.527, with a p-value of 0.012, the interaction between female consumer and Spendception has impact on Impulse Buying. This implies that there is a direct moderating influence when IB is decreased and both female consumer and Spendception are high.
Conditional Indirect Impacts:
The
Table 13 also examines how varying degrees of female consumer affect the indirect impacts of Spendception on CPB through IB:
Low Female Consumer: The indirect impact is 0.055 at low levels of female consumer, with a SE of 0.023. It is significant at the p < 0.001 level, as shown by ***. This indicates that when female consumer level is minimal, there is a strong and significant pathway from Spendception to IB and ultimately to CPB.
Medium Female Consumer: With a p-value of 0.001, the indirect effect is still significant at medium levels of female consumer, rising to 0.77 with a SE of 0.047. This implies that the mediation pathway is still important at medium female consumer levels, but more so than at low risk.
High Female Consumer: The p-value is 0.71, the indirect impact is 0.087, and the SE is 0.099 for high levels of female consumer. This suggests that the mediation pathway become insignificant and it suggest moderated mediation.
Index of Moderate Mediation: The p-value is 0.019, the SE is 0.035, and the Moderated Mediation Index is -0.009. This index's significance suggests that there is moderated mediation, which means that the degree of female consumer influences the association between Spendception and CPB through IB.
The presence of moderated mediation suggests that Spendception have a direct impact on CPB, but that this influence is also influenced by female consumer, which in turn influences IB.