International Journal of Regional Development issn 2373-9851 2018

International Journal of Regional Development 
ISSN 2373-9851 
2018
 
54 
hypothesized model. As convergent validity provides evidence regarding the accuracy of 
indicators of a specific construct, to be attained, it is vital that multiple indicators of a single 
construct share a high proportion of variance (Hoyle, 2012, Kline, 2011). Also construct 
reliability (composite reliability) in this study is greater than 0.7 which is higher than the 
recommended one as shown in Table 2.0. In studies, recommended construct reliability 
should be 0.7 or higher. Others recommend that the value between 0.6 and 0.7 can also be 
acceptable. The Average construct variance also proved to be acceptable as exceeded 0.5 as 
rule of thumb. 
We also checked the nomological and face validity. The face validity was assessed to ensure 
that the content of the measured items is correct and is well measured in line with the 
proposed characteristics or attribute. It also checked that the content of each item is unique 
and does not overlap with other measured items which do not belong to them. On the other 
hand, nomological validity in forms whether there is logical relationship between the 
constructs and it makes sense according to the theory.
Whereby, discriminant validity was assessed by calculating the square root of the AVE of 
each construct. In Table2.0 AVE scores were greater than the cut off value of 0.50. And as 
presented, the value of the square root of AVEs along the diagonal on Table 1.0 exceeded the 
corresponding latent factor correlations in the same row and column. Thus, according to the 
method, each construct in the hypothesized model was distinct from one another.
Finally, we checked on items’ loading pattern so as provide evidence of the discriminant 
validity of the constructs. There were no cross-loadings of measured items on the latent 
constructs. Each measured variable well represented single latent factor and not multiple 
latent factors. These CFA results suggest that the measurement model provides a reasonably 
good fit and thus it is suitable to proceed for further examinations.
Figure 2. CFA diagram with standardized loading estimates