*4.1. Measurement Model*

The criteria for validity and reliability are used to check the fit of the measures. Reliability represents the ability of a measuring instrument (items) to consistently deliver the same result. Validity represents the degree to which a measuring instrument (items) accurately measures a concept [41].

The reliability of the subscales of the model was adequate, with alpha coefficients above 0.7 [42].

The factor loadings, composite reliability and mean variance extracted were assessed for convergent validity, as suggested by Hair et al. [42]. The loadings of all items exceeded the recommended value of 0.6, and the composite reliability and the average variance extracted in all cases exceeded the recommended values of 0.7 and 0.5, respectively, as shown in Table 3.


**Table 3.** Reliability and validity of measures.

Table 4 demonstrates the discriminant validity between the constructs, as the mean variance estimates extracted, which exceeded 0.5, were higher than all the phi-squared correlations between the constructs [43].


**Table 4.** Discriminant validity.

#### *4.2. Structural Model*

The structural model was estimated (see Figure 3). The recommended fit indices in assessing model adequacy [44,45] were above the recommended range [46,47], indicating a respectable fit: χ2 = 392.822; df = 163; (*p* < 0.01); comparative fit index (CFI) = 0.968; Tucker–Lewis index (TLI) = 0.963 and root mean error of approximation (RMSEA) = 0.054 with a 90% confidence interval of 0.048 and 0.061. Table 5 shows the standardized path coefficients of the relationships of dimensions.

**Figure 3.** Estimation of model.


**Table 5.** Structural model results.

All standardised path coefficients were positive and significant at 99% confidence, except for the relationship between satisfaction and normative commitment (standardised path coefficient = −0.305; *p* < 0.01). These results support the hypotheses stated, except for H6 ('Student satisfaction positively influences normative commitment'), in which, despite being significant, the opposite effect is observed, with a negative relationship between satisfaction and normative commitment, consequently rejecting H6.

Among the different dimensions of experience, the hypotheses (H1, H2, H3, H4) are corroborated, with cognitive–affective experience having the greatest weight (standardised path coefficient = 0.736; *p* < 0.01) and formative experience being the one that, surprisingly, contributes the least (standardised path coefficient = 0.59; *p* < 0.01). Experience has a strong influence on student satisfaction with the BL model (standardised path coefficient = 0.804; *p* < 0.01), such that as experience increases, so too does satisfaction. Finally, increasing satisfaction has greatly increased affective commitment (standardised path coefficient = 0.875; *p* < 0.01), supporting H5.
