Q² with Second Order Constructs and moderating Variables

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
MichaelKeuper
PLS Junior User
Posts: 3
Joined: Thu Dec 07, 2006 2:32 pm
Real name and title:

Q² with Second Order Constructs and moderating Variables

Post by MichaelKeuper »

Hello everyone,

maybe some of you can help me to calculate Stone Geissers Q² in this case:

I have a model consisting of two second order constructs (exogenous variable and the endogenous variable) + 5 Moderating Effects + 5 Interaction Terms.

According to Henseler(University Nijmegen), in the case of moderating constructs, singel-item constructs should be used in order to keep the latent variable scores constant when introducing interaction terms.

But the Q²-Value obviously can't be calculated by PLS in that case. The whole table in the pls-graph output-file shows only 0-Values for each iteration and in the end for Q².

Is there any ohter possibility to calculate Q²? Did I make a mistake?

Thanks for your advice!
User avatar
cringle
SmartPLS Developer
Posts: 818
Joined: Tue Sep 20, 2005 9:13 am
Real name and title: Prof. Dr. Christian M. Ringle
Location: Hamburg (Germany)
Contact:

Post by cringle »

Hi Michael,

the blindfolding procedure works fine for 2nd order constructs and moderating variables. In SmartPLS, you get the results that you need. However, you receive values of zero for Q² when you compute the cv-redundancy for latent exogenous variables and when you compute the cv-communality for in single item constructs (regardless whether exogenous or endogenous). Moreover, the number of cases shall not be a multiple of the selected omission distance for the blindfolding procedure.

Those hints should also be relevant for pls graph (even if we would love to see that our colleges at Aachen finally use and cite SmartPLS ;))

Best regards
Christian
MichaelKeuper
PLS Junior User
Posts: 3
Joined: Thu Dec 07, 2006 2:32 pm
Real name and title:

Post by MichaelKeuper »

Dear Christian,

thanks for your help. Especially that your advices do even work with pls-Graph ;-)..

I thougth it could have been a general problem. But now I spotted another Problem with my second order construct (beyond beeing a S.O.-Construct). The calculations work fine now.

Thanks again,
best regards,

Michael
Post Reply