how does PLS test the influence of each IV in a model?

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
irene eleonora
PLS Junior User
Posts: 10
Joined: Mon Mar 26, 2012 3:06 pm
Real name and title:
Location: Università Bocconi

how does PLS test the influence of each IV in a model?

Post by irene eleonora »

Hello.
I have a doubt problem in understanding a reviewer's comment concerning my PLS analyses. I have a model with 3 independent variables (three motivations), some controls and one dependent variable (the adoption of one policy). The reviewer says:
"Considering the fact that you have examined the three motivations individually, how can you be sure that the use of EPMS [i.e. my DV] was driven specifically by this motivation instead of the two other motivations? My point is that the three motivations coexist simultaneously. They are complementary and not mutually exclusive. In other words, given the fact that you did not examine conceptually or empirically the combination of motivations, you cannot conclude that firms actually use EPMS in reaction to stakeholders’ concern. In order to do that, you would need to test the influence of stakeholders’ concern on EMPS while controlling for the other two motivations [....] PLS cannot distinguish the impact of those various combinations. It tests the overall specific influence of each motivation. Other types of tests would be necessary to test those combinations"

Is this true? Doesn't PLS test the influence of each IV in a model by controlling for the other IV in the model? If not, what other types of test should I use, more traditional regression analyses?

Thank you very very much for your help ...
Irene
christian.nitzl
PLS Expert User
Posts: 248
Joined: Sat Jul 25, 2009 1:34 pm
Real name and title:

Post by christian.nitzl »

I believe the review refer to the point that PLS as a SEM cannot say something about casual relation. Therefore you need always a good theory or some experimental studies for deriving a model in SEM. Perhaps you can check for your independent variables the VIF (variance influence factor) with help of the latent construct values. These values you can find in the PLS output. But for calculating the VIF values you have to use another software program like SPSS. Perhaps you can also perform an explorative factor analysis. This is a little bit hands-on but with help of that you can see how many constructs you get with help of this factor analysis when using the items of the independet variables.
Cheers,
Christian
Post Reply