Page 543 - ISC PROCEEDINGS 21.4

4.3.2. Hypothesis testing and discussion
Hair et al. (2022) explain that path coefficients (β) reflect the magnitude and
direction of links between constructs, whereas t-statistics and p-values assess statistical
significance. The findings support all eight hypotheses at standard significance thresholds
(p < 0.05). Table 4.5 and Figure 4.1 present the PLS-SEM results and hypothesis tests,
derived from PLS-SEM with bootstrap resampling. These outcomes affirm the theoretical
model and the importance of all hypothesized relationships.
Table 8. Hypotheses testing results
Original Sample Standard
Sample Mean Deviation T Statistics P Values Result
(|O/STDEV|)
(O) (M) (STDEV)
H1: BU -> SI 0.197 0.196 0.068 2.896 0.004 Accepted
H2: BU -> MA 0.509 0.510 0.051 9.910 0*** Accepted
H3: MA -> PP -0.572 -0.571 0.055 10.498 0*** Accepted
H4: PP -> SI -0.192 -0.197 0.077 2.502 0.013 Accepted
H5: BU -> SA 0.683 0.683 0.042 16.291 0*** Accepted
H6: SA -> SI 0.180 0.183 0.085 2.121 0.034 Accepted
H7: FL -> SI -0.271 -0.271 0.065 4.195 0*** Accepted
H8: FLxBU-> SI 0.088 0.085 0.024 3.647 0*** Accepted
Source: Author

Figure 2. PLS-SEM analysis results
Source: Author
Hypothesis 1 (H1) was supported, showing that BNPL positively influences shopping
intention (SI) (β=0.197, t=2.896, p=0.004). BNPL's installment feature alleviates short-
term financial pressure, aligning with hyperbolic discounting theory and prior studies
(Pamungkas & Putri, 2024; Kumar et al., 2024; Ahmad, 2025).
H2 was strongly supported, with BNPL positively related to mental accounting (MA)
(β=0.509, t=9.910, p<0.001), as no-interest installments create affordability illusions,
complementing Kumar et al. (2024).

542