4.3.3. Structural Model

The structural model was evaluated by assessing path co-efficient, the explaining power of the model, and explained variance. Figure 2 has shown the path co-efficient value with PLS results. Explained variable tested based on the R2 value of the endogenous variable. According to research R2 value is weak when the value is 0.02, moderate when the value is 0.13 and value is substantial when the value is equal and higher than 0.26. Figure 1 in this research shows the R2 value of an endogenous variable (performance) is 0.531, which indicates that the endogenous construct explains 50% of the variance. It is higher than the value of 0.26, which means the model is considered very satisfactory.

**Figure 2.** The path coefficient value. \* Significant at 0.05 level and \*\* Significant at 0.01 level.

In this research GoF is the global criterion of the goodness of fit used to assess the performance of the path model in both structural and measurement models. GoF (0 ≤ GoF ≤ 1) index is obtained as a geometric mean of the average communality index and average R2 value. According to Ringle et al. [101] based on the commonality indexes and R2 values calculated with the average communality index average R2 value was calculated as 0.5013 and 0.531, respectively. Thus,

$$\begin{array}{l} \text{GoF} = \sqrt{\text{Comm multiplicity} \times \text{R}^2} \\ = \sqrt{0.531 \times 0.5031} = 0.516 \end{array}$$

In this study, the GoF value obtained was 0.516 which is higher than the accepted value of 0.36 for huge effect sizes of R2. To compare with the baseline value suggested by Akter et al. [98] where GoFsmall = 0.1, GoFmedium = 0.25, and GoFlarge = 0.36. The result of this GoF value confirms that PLS path modeling has explaining power is substantial. Based on the guideline by Wetzel et al. [102] for the adequate support of GoF, it has also been reported that this PLS model is valid globally.
