This is a path model showing only the structural model components.
Collinearity is assessed using the variance inflation factor (VIF). VIF should be < 5, ideally \(\leq\) 3.
VIF is above 3 for some constructs, specifically lower-order constructs. But it is not above 5 for any constructs.
In-sample predictive power is assessed using variance explained R². R² \(\geq\) 0.75 indicates substantial in-sample predictive power, R² \(\geq\) 0.5 moderate and R² \(\geq\) 0.25 weak in-sample predictive power. R² \(\leq\) 0.10 indicates a lack of model predictiveness.
R² is below 0.25 for the Threat Beliefs lower-order constructs, but not below 0.1. All other R² values are above 0.25, indicating weak in-sample predictive power at a minimum. For Response Beliefs and Behavioral Intention, R² is above 0.5, indicating moderate in-sample predictive power.
Effect size f2² measures the impact of a predictor construct on an endogenous construct. f² \(\geq\) 0.35 indicates a large effect, f² \(\geq\) 0.15 a medium and f² \(\geq\) 0.02 a small effect.
The effect sizes of the relationships including many lower-order constructs are below 0.2. Benevolence has only an effect on Perceived Response Costs and Perceived Severity. Competence has an effect on all its descendant lower-order constructs except Perceived Susceptibility. Integrity only has an effect on any of Perceived Response Efficacy. Knowledge has a discernible effect only on Perceived Response Efficacy, Threat Beliefs and Perceived Severity. From the lower-order constructs, only Perceived Response Efficacy and Perceived Response Costs have a discernible effect on Behavioral Intention. Apart from that, there is only a negligible effect of Threat Beliefs on Behavioral Intention. If this is mirrored in non-significant and irrelevant path coefficients, these structural relationships can be removed. Other than that, all effects are small at a minimum. The effects of Perceived Response Efficacy and Response Beliefs on Behavioral Intention are medium. The effect of Distrusting Beliefs on both Threat Beliefs and Response Beliefs is large.
Out-of-sample predictive performance has to be evaluated using a simplified model without HOCs as an approximation as seminr as of now does not support plspredict for HOC. If the root mean square error (RMSE) or the mean absolute deviation (MAD) of the naive LM model is below the ones for the PLS model, this indicates a lack of predictive performance. However, a lack of predictive performance does not imply a lack of explanatory power.
Only COTB1, COTB3 and COBI3 are predicted better by the model than the LM in terms of both RMSE and MAD.
Significance is denoted by t-test and p values. A t of \(\geq\) 1.65 signifies significance at the 10 % level, t \(\geq\) 1.96 at the 5 % level and t \(\geq\) 2.57 at the 1 % level.
Concerning higher-order and regular constructs, only the path from Threat Beliefs to Behavioral Intention is non-significant. With lower-order constructs, non-significant paths are those from Knowledge to Perceived Response Costs and Perceived Self-Efficacy, all from Benevolence except those from Perceived Response Costs and Perceived Severity, the path from Competence to Perceived Susceptibility, all from Integrity except to Perceived Response Efficacy, all from the Threat Beliefs lower-order constructs and the one from Perceived Self-Efficacy,and both the paths from Descriptive and from Injunctive Norm to Behavioral Intention.
Concerning higher-order and regular constructs, the path from Threat Beliefs to Behavioral Intention is non-significant and the effect size is negligible. Therefore the path and the Threat Beliefs construct can be removed. All other paths are significant and the effects, with the exception of the effect of Knowledge on Response Beliefs, are at least weak. Concerning the remaining lower-order constructs, Benevolence only has a significant path with a discernible effect to Perceived Response Costs, and Integrity only to Perceived Response Efficacy. Knowledge only has a significant path with a discernible effect to Perceived Response Efficacy. Perceived Self-Efficacy has no significant path to or discernible effect on Behavioral Intention. Therefore, Perceived Self-Efficacy could be removed as a lower-order construct. Neither Descriptive nor Injunctive Norm have a significant path to or a discernible effect on Behavioral Intention, though Subjective Norm has both. As Injunctive Norm is the comparatively weaker construct, it should be removed for the next model iteration and Descriptive Norm should replace the Subjective Norm HOC to assess wether it has a significant effect on its own. To differentiate the influence of Distrusting Beliefs and Knowledge on the Response Beliefs lower-order construct, an alternative model with the Response Beliefs lower-order constructs split up should be estimated.