Discriminant Validity not met - Yet valid answers?

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
gera
PLS Junior User
Posts: 1
Joined: Sat Feb 20, 2021 10:40 am
Real name and title: Hermann Raab

Discriminant Validity not met - Yet valid answers?

Post by gera »

Hey Community,

i'm writing my thesis on technology acceptance and wanted to test my hypothesis through pls modelling.

After checking the discriminant validity, i noticed that the thresholds for the endogen variables NV (actual usage behavior) and NA (behavioral intention) are not met.
Image

My sample has a high amount of people actually using the technology (76%), which suggests, that there needs to be a high correlation between intention and usage.

So my question is:
Can my thesis deliver valid results by explaining the lack of discriminant validity through the sample composition?
If no, how can i proceed?
Henseler, Ringle, Sarstedt (2015) suggested to increase monotrait-heteromethod by deleting indicator with low correlation within the construct or decrease heterotrait-heteromethod correlations by deleting indicator with high correlation to items of the opposite construct.
Nothing of this worked for me.

Literature:
Henseler, J., Ringle, C. M., and Sarstedt, M. (2015). A New Criterion for Assessing Discriminant Validity in Variance-based Structural Equation Modeling. Journal of the Academy of Marketing Science, 43(1): 115-135.
Post Reply