Discriminant Validity through Cross Loadings
According to Gefen and Straub (2005), “discriminant validity is shown when each measurement item correlates weakly with another construct excepts for the ones to which it is theoretically associated”.
In cross-loadings, the researcher examines the various items to identify those that have high loadings on the same construct and those that load highly on multiple constructs. Thus, to establish discriminant validity at the item level means there is a high correlation between items of the same construct and a very weak correlation between items of a different construct. As simple as this approach is, it has no theoretical justifications or empirical proof (Henseler et. al., 2015).
We refer our reader here on how to find cross-loading. In the table below, the bolded items (in blue) represent the factor loadings for each construct and the cross-loading are those in red for the same construct. You can see the cross-loading for each construct is very low indicating good discriminant validity.
References:
Henseler, J., Ringle, C.M. & Sarstedt, M. A new criterion for assessing discriminant validity in variance-based structural equation modeling. J. of the Acad. Mark. Sci. 43, 115–135 (2015). https://doi.org/10.1007/s11747-014-0403-8
Gefen, D., & Straub, D. W. (2005). A practical guide to factorial validity using PLS-Graph: tutorial and annotated example. Communications of the AIS, 16, 91–109.