Page 429 - ISC PROCEEDINGS 21.4
P. 429

2.4.2. Analysis software
                        Given the exploratory nature of the integrated PPM framework and its focus on
                  novel relationships, PLS-SEM was deemed more appropriate than CB-SEM (Hair et al.,
                  2022).
                        3. Results and discussion
                        3.1. Results
                        Prior to SEM analysis, the study assessed scale reliability and validity as follows:
                        (1) Cronbach's Alpha ≥ 0.7, Composite reliability (rho_a ≥ 0.7 and rho_c ≥ 0.7)
                  confirmed that all indicators met reliability standards.
                        (2) Average variance extracted (AVE) ≥ 0.5 demonstrated convergent validity across
                  scales.
                        (3) The Fornell-Larcker criterion and Heterotrait-Monotrait (HTMT) Correlation
                  Ratios demonstrate that most construct pairs exhibit HTMT values ≤ 0.85, confirming
                  discriminant validity. Although one pair (EE-PE) shows HTMT ≥ 0.85, it remains within
                  acceptable thresholds (≤ 0.90).
                        Thus, all measurement scales in this study satisfied reliability, convergent validity
                  and discriminant validity criteria. Consequently, these scales were retained for
                  subsequent partial least squares structural equation modeling (PLS-SEM) analysis.
                        The results of the research model regression are presented below:
                                              Table 1. Bootstrapping Test Results

                                    Original       Sample        Standard
                                   sample (O)     mean (M)       deviation     T statistics    P-values

                   MOOR -> PR         0.807         0.806          0.022         37.065          0.000


                   MOOR -> SC         0.917         0.917          0.009         107.526         0.000


                   MOOR -> SI         0.877         0.877          0.013         69.667          0.000


                    PULL -> EE        0.865         0.865          0.014         60.209          0.000



                    PULL -> PE        0.900         0.901          0.010         86.229          0.000


                    PULL -> PI        0.739         0.738          0.031         23.675          0.000


                   PUSH -> CM         0.863         0.862          0.014         61.251          0.000



                    PUSH -> SA        0.917         0.917          0.007         125.293         0.000

                                                                  Source: Research team survey data, 2025
                        The results indicate that all path coefficients (β) for hypothesized relationships in
                  the research model exhibit positive directions, demonstrating positive impacts between
                  observed variables and the Push-Pull-Mooring (PPM) construct groups. Additionally, all p-
                  values = 0.000 (< 0.05). Consequently, all proposed hypotheses linking observed variables
                  to the Push-Pull-Mooring groups are supported.


                                                                                                      428
   424   425   426   427   428   429   430   431   432   433   434