Page 469 - ISC PROCEEDINGS 21.4

Table 3. Fornell-Larcker Criterion
Constructs AT IT PBA PBE PS PC

AT 0.784

IT 0.488 0.823

PBA -0.195 0.064 0.756

PBE 0.324 0.348 0.065 0.828

PS 0.336 0.338 0.071 0.382 0.801

PC 0.094 0.195 0.061 0.167 0.136 0.786

Source: Research team (2025)
4.3. Structural Model Assessment
Assessment of multicollinearity using VIF
Before examining the structural relationships in the model, multicollinearity among
latent variables was assessed using the Variance Inflation Factor (VIF). Multicollinearity
occurs when three or more independent variables are highly correlated, which can distort
regression estimates and lead to unstable or misleading results. The results show that the
Inner VIF values for the paths between latent variables were between 1.000 and 1.191.
Since all values remained below the threshold of 3, the model is considered free from
significant multicollinearity issues.
Structural Model Results and Hypothesis Testing
To evaluate the hypothesized relationships in the model, bootstrapping with 5,000
resamples was conducted. The significance of each relationship was assessed using path
coefficients (Original Sample) and p-values, with a conventional significance level of 0.05.

Figure 2. Bootstrapped PLS-SEM Path Model
Source: Research team (2025)
The results indicate that perceived susceptibility does not have a statistically
significant effect on attitude toward health behavior (β = 0.034, p = 0.292). Therefore, H1
is not supported. This suggests that perceiving oneself as vulnerable to health risks alone

468