Page 457 - ISC PROCEEDINGS 21.4

Table 2. Measurement Model Assessment Result
Average
Outer Cronbach's alpha Composite variance
loadings reliability
extracted
GBA 0.707 - 0.822 0.845 0.849 0.618
GCC 0.772 - 0.845 0.811 0.818 0.638
GM 0.775 - 0.833 0.738 0.740 0.656
GPB 0.798 - 0.812 0.731 0.733 0.649
Source: Author’s calculation using SmartPLS 4 software.
Based on the data provided in the second table, the evaluated measurement
framework fully satisfies the essential benchmarks for convergent validity and overall
reliability. Initially, all item loadings are shown to fall between 0.707 and 0.845. Since
these figures comfortably clear the standard 0.70 cutoff, it proves that the manifest
variables appropriately capture their respective latent factors. Furthermore, regarding
internal consistency, Cronbach’s alpha coefficients span from 0.731 to 0.845, while
composite reliability (CR) indices range between 0.733 and 0.849. Both metrics
successfully eclipse the accepted 0.70 baseline. Lastly, the Average Variance Extracted
(AVE) yields values from 0.618 to 0.656, safely surpassing the 0.50 minimum to verify
proper convergent validity. Consequently, this empirical evidence definitively
substantiates the structural soundness and dependability of the model. Moreover, effect
size and discriminant validity were additionally examined using the f² index and the HTMT
ratio, as summarized in Table 3. In particular, the f² statistic was applied to assess the
extent to which each exogenous construct influences the endogenous variables, where
benchmark values of 0.02, 0.15, and 0.35 represent small, moderate, and strong effects,
respectively (Hair et al., 2019). Concurrently, the HTMT ratio was employed to verify
discriminant validity among the constructs, with values below 0.90 indicating adequate
distinctiveness (Henseler et al., 2015). Overall, these measures provide further evidence
of the model’s explanatory strength and confirm the empirical separation of the latent
variables.
Table 3. Discriminant validity and effect size assessment
HTMT criteria
GBA GCC GM GPB
GCC 0.754
GM 0.789 0.761
GPB 0.728 0.831 0.742
f-square result
GBA GCC GM GPB
GBA 0.039
GCC 0.206 0.189

GM 0.190 0.031
Source: Author’s calculation using SmartPLS 4 software.
The results in Table 3 show that the HTMT values range from 0.728 to 0.831, which
are all below the recommended threshold of 0.90 (Henseler et al., 2015). This indicates
that the constructs demonstrate adequate discriminant validity, meaning that each latent
variable is empirically distinct from the others. Regarding the effect size (f²), the results

456