The effect of indicators of lV on DV?!

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
katarina86
PLS Junior User
Posts: 7
Joined: Thu Feb 21, 2019 9:44 am
Real name and title: PhD Katarina Njegic Assistant Professor

The effect of indicators of lV on DV?!

Post by katarina86 »

Hello everyone,

I have tested the model using SmartPLS 3 and I recently got the review from the journal. Besides other variables in the model, I have a latent variable (A) that is composed of three reflective indicators (a1,a2,a3). I have tested the effect of a latent variable (A) on the dependent variable (C). The dependent variable is a higher-order construct. One of the reviewers asked to comment which indicator (a1,a2 or a3) has the most important role in the relationship between A and C. I have explained in my paper the effect of A to C and the implications of that effect, but the reviewer wants me to comment and explain which of the indicators of A contribute the most to the increase of C.

Is there a way I can test that? Or at least give a satisfactory answer... I don't see a way since the latent variable (A) is unobservable and it is just measured through the reflective indicators (a1, a2, a3). How can I test separately the effect of a1, a2 or a3 on DV rather than create a new model... I'm really confused... The reviewer did not asked to create a new model, just to give explanation based on the results I've already presented.

Sincerely,
Katarina
MFW
PLS Junior User
Posts: 2
Joined: Thu Nov 26, 2020 4:56 pm
Real name and title: Marie Freia Wunderlich

Re: The effect of indicators of lV on DV?!

Post by MFW »

Hi,

what about running an Importance-Performance-Map Analysis (IPMA)? I think, with that you will be able to see the effect sizes (=importance) of each item (and if you want also your constructs) on your DV. Furthermore, you will be able to see how respondents are currently performing. It could be that one item/construct is very important in terms of effect sizes, but people score low on this item.

Best,
Marie
Post Reply