*5.1.1 Assessing the measurement model*

The assessment of the measurement model includes the evaluation of reliability and validity [72]. Cronbach's alpha and the composite reliability (CR) are used in the reliability analysis, with acceptable values over 0.70 [71]. To evaluate the validity, the measures of indicator reliability, convergent validity, and discriminant validity are used. Indicator reliability is confirmed if the values of standardized outer loading are above 0.7, while the convergent validity is presented with average variance extracted (AVE), whose acceptable values are higher than 0.5. The results of indicator reliability and convergent validity are presented in **Table 2**.


*The Influence of Employer Brand Dimensions on the Affective Organizational Commitment… DOI: http://dx.doi.org/10.5772/intechopen.112133*


#### **Table 2.**

*Model evaluation measurements.*

As visible from the table, the measures of Cronbach's alpha and composite reliability (CR) are within acceptable ranges, and so are the measures for average variance extracted. On the other hand, standardized outer loadings for several items (ECON2, INT3, APP4, DEV3, DEV4) are below their acceptable value of 0.7, so further analysis was conducted to see if the indicator deletion would have an impact on AVE and composite reliability [69]. As this was not the case, the loadings were retained in the model.

In order to assess discriminant validity, the Fornell-Larcker criterion and the heterotrait-monotrait (HTMT) ratio of correlations were observed [69]. **Table 3** shows the Fornell-Larcker criterion. From the table it is visible that the AVE of each latent variable is higher than the squared correlations of the other latent variables, so the criterion is satisfied as well [71].

Another measure of discriminant validity is the HTMT ratio, shown in **Table 4**. For the model to be valid, the interval of the HTMT data should not include the value of one for all the combinations of constructs, which is also present in this case [69].


#### **Table 3.**

*Discriminant validity: Fornell–Larcker criterion.*


#### **Table 4.**

*Discriminant validity: HTMT ratio.*

The indicated results demonstrate that model is appropriate for further analysis of the structural model assessment and hypothesis testing.

### *5.1.2 Assessing the structural model*

Before analyzing the structural model, the variance inflation factor (VIF) was calculated in order to check for multicollinearity issues [72]. **Table 5** reports such results, which are all within acceptable values of below 5.

The absence of multicollinearity allows the valuation of the structural model, which was derived from the bootstrapping method with 5000 samples [69]. The results are presented in **Table 6**, which also indicates the outcomes of hypothesis testing.

The table shows that the application value positively influences affective organizational commitment (β = 0.361; p < 0.05), which confirms hypothesis H1. The interest value also has a positive influence on affective organizational commitment (β = 0.504; p < 0.01), thus confirming the second hypothesis. Other independent variables did not show a statistically significant or positive influence on affective organizational commitment, so hypotheses H3, H4, and H5 cannot be confirmed.

The coefficient of determination, R2 , is used as a measure of the predictive accuracy of the model [69]. The R2 for affective organizational commitment in this study is 0.462, which can be interpreted as moderate [71], and reveals that 46.2% of the variation in developing affective organizational commitment occurred under the influence of employer branding dimensions [73]. Additionally, predictive validity was calculated through a cross-validated redundancy measure, which is based on Stone-


**Table 5.**

*Multicollinearity—Variance inflation factor (VIF).*

*The Influence of Employer Brand Dimensions on the Affective Organizational Commitment… DOI: http://dx.doi.org/10.5772/intechopen.112133*


*Note: AOC = affective organizational commitment. \*p < 0.10.*

*\*\*p < 0.05.*

*\*\*\*p < 0.01.*

#### **Table 6.** *Hypothesis testing.*

Geisser's Q2 value [70]. The Q<sup>2</sup> value for affective organizational commitment is Q<sup>2</sup> = 0.353, which is interpreted as acceptable predictive relevance, meaning that the model has predictive validity for the endogenous construct [71]. The effect size coefficient is obtained by calculating Cohen's f2 [72]. The effect size is interpreted according to reference values, where 0.02 represents a small, 0.15 is a medium, and 0.35 is a large effect of an individual exogenous construct on an endogenous construct [74]. From the above, it follows that interest value has a large effect on affective organizational commitment (f<sup>2</sup> = 0.412), application value has a medium effect on affective organizational commitment (f<sup>2</sup> = 0.146), while other values have a weak effect on affective organizational commitment (f<sup>2</sup> economic value = 0.062, f<sup>2</sup> development value = 0.009, f<sup>2</sup> social value = 0.003).
