Discriminant Validity not established

Frequently asked questions about PLS path modeling.
Post Reply
amoghta
PLS Junior User
Posts: 5
Joined: Sun Mar 08, 2015 5:37 pm
Real name and title: Amogh Talan, Assistant Professor

Discriminant Validity not established

Post by amoghta » Sun Mar 29, 2015 7:47 pm

Hello all,

In a structural model if discriminant validity is not established between two or more constructs, what would you recommend before carrying on the analysis of R square f square Q square etc and other analyses to see the predictiveness of the model?

Should I remove the constructs which don't qualify to establish the discriminant validity? Or should I just mention that discriminant validity is not established between A and B and just carry on with the analysis for the model for R, Q square etc?

User avatar
Hengkov
PLS Super-Expert
Posts: 1618
Joined: Sun Apr 24, 2011 10:13 am
Real name and title: Hengky Latan
Location: AMQ, Indonesia
Contact:

Re: Discriminant Validity not established

Post by Hengkov » Tue Mar 31, 2015 9:25 am

Hi,

Discriminant validity if not established, I think you have other problems like multicoloniarity which makes the correlation between the constructs to be high, so HTMT, Cross-Loading or Fornell-Lacker criteria can not be met.
In addition, it may be included in the model constructs are irrelevant, or errors in data collection (common method bias).

Greetings,

Post Reply