Moderating effects

This forum is closed, and read-only.
Locked
nndu
PLS Junior User
Posts: 1
Joined: Wed Jan 26, 2011 6:10 pm
Real name and title:

Moderating effects

Post by nndu »

The moderating effects has been added in the new released version, however, it is only possible to have the multiplication of one moderator and one predictor. What about more moderators? Perhaps Smart PLS should support the multiplication of several moderators and predictor.

Best,
Nina
User avatar
Diogenes
PLS Super-Expert
Posts: 899
Joined: Sat Oct 15, 2005 5:13 pm
Real name and title:
Location: São Paulo - BRAZIL
Contact:

Post by Diogenes »

Hi Nina,

Is it a higher order interaction (x1*x2*x3)?

To include many interactions terms (pairs of variables) is already Ok.

Another idea is implement a question in the “moderation box” if the LV is formative?
Or use the LV score of the formative LV to compute the interaction terms.
Today the SmartPLS make the multiplications (interaction terms) regardless if the LV are formative or reflective.


Another ideas were made here:
Between-group analysis
viewtopic.php?t=348

Using pls on a split dataset
viewtopic.php?t=357

Moderator-Variables and Interaction Indicators
viewtopic.php?t=146

Multi-Group-Comparison
viewtopic.php?t=210

Best regards,

Bido
Peter1988
PLS Junior User
Posts: 2
Joined: Tue Jun 10, 2014 3:41 pm
Real name and title:

Re: Moderating effects

Post by Peter1988 »

Dear all,

I am currently facing the same problem. I would like to create a joint moderating effect of gender*age*experience as was done in UTAUT 2.
Gender is coded as dummy and the other two variables are integers.

How should I create the variable gender*age*experience?
--Should I standardize the values for gender, age and experience and then multiply them?
--Or use latent variable scores for gender, age and experience and multiply them?
--Or use original values for gender, age and experience and multiply them?

Your help would be highly appreciated.

Best,
Peter
jmbecker
SmartPLS Developer
Posts: 1282
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Re: Moderating effects

Post by jmbecker »

Are age, gender and experience single indicator or multiple indicator variables in you model?
I would guess single indicator variables. In that case the latent variable score is simple the standardized score of your single indicator variable.

I would usually use mean-centered variables for the product variable. However, the decision will affect your result interpretation, so you should look in the standard literature on moderated multiple regression. The issue is well discussed in the literature.
Dr. Jan-Michael Becker, BI Norwegian Business School, SmartPLS Developer
Researchgate: https://www.researchgate.net/profile/Jan_Michael_Becker
GoogleScholar: http://scholar.google.de/citations?user ... AAAJ&hl=de
Locked