Hello everybody,
Please correct me if i'm wrong. My model is a reflective second order model. To build the second order model i assigned the same manifest variables of the first order constructs. For instance, there are 5 variables measure INTEGRATION, 3 variables measure DoD, 4 variables measure BOUNDARIES, and 3 variables measure KNOWLEDGE SPANNING. All these constructs are influenced by FRAG. that is the second order construct.
So to run the model, i assigned all the manifest variables (15 variables) to FRAG. It works, but there is one problem related to AVE of FRAG. It is always small because the correlation between these variable is low.
The question is what to do in this case, do i have to eliminate the weak variables from FRAG (loading less than .5). In this case do you think that will effect the whole model? OR do i have to accept the results of AVE as it is?
Thanks
Second order model
- alialashwal
- PLS Junior User
- Posts: 8
- Joined: Wed Jul 20, 2011 9:58 am
- Real name and title: Alashwal, A.M.
Second order model
Ali M. Alashwal, PhD
Faculty of Built Environment,
University of Malaya,
50603, Kuala Lumpur, Malaysia.
Faculty of Built Environment,
University of Malaya,
50603, Kuala Lumpur, Malaysia.
- Diogenes
- PLS Super-Expert
- Posts: 899
- Joined: Sat Oct 15, 2005 5:13 pm
- Real name and title:
- Location: São Paulo - BRAZIL
- Contact:
Hi,
The AVE of the 2nd order LV are underestimated by SmartPLS,
viewtopic.php?t=640
Did you compute it by hand?
viewtopic.php?p=5297#5297
Excluding one LV could jeopardize the content validity.
Best regards,
Bido
The AVE of the 2nd order LV are underestimated by SmartPLS,
viewtopic.php?t=640
Did you compute it by hand?
viewtopic.php?p=5297#5297
Excluding one LV could jeopardize the content validity.
Best regards,
Bido