Mediation Analysis - VAF or Path Significance

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
PhilipBeu
PLS User
Posts: 15
Joined: Tue May 09, 2017 8:01 am
Real name and title: Philip Beuße

Mediation Analysis - VAF or Path Significance

Post by PhilipBeu »

Hello everyone,

is just performed mediation analysis using SMARTPLS 3.

My question is: Which indicator is relevant for partial / full mediation?

In this post: viewtopic.php?f=5&t=15489&p=25496&hilit=mediator#p25496

JMBecker states, that if the direct effect turns insignificant it is a FULL mediation.

However bootstrapping and the following calculation of VAF only puts out a VAF-Value of 77% which would be a PARTIALmediation. See: http://www.ww.uni-magdeburg.de/marketin ... uery2.html

Could you clear things up?

Best regards

Philip
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:

Re: Mediation Analysis - VAF or Path Significance

Post by cringle »

You may want to follow the procedure on testing mediating effects as explained in the book on PLS-SEM: https://www.smartpls.com/documentation/ ... on-pls-sem

First, you test the significance of the indirect effect. If it is sign. and also the direct effect is sign., you have a partial mediation; if the ind. eff. is sign. and the direct eff. is not sign., you have a full mediation.

Best
Christian
PhilipBeu
PLS User
Posts: 15
Joined: Tue May 09, 2017 8:01 am
Real name and title: Philip Beuße

Re: Mediation Analysis - VAF or Path Significance

Post by PhilipBeu »

Dear Prof. Ringle,

Thanks for the reply!

I refer to the procedure you mentioned.

If I follow this procedure, I would not need to take VAF into account, right?

This is still confusing me, because (as mentioned in my original post), testing significance of paths vs. VAF lead to different results.

To clear things up from my point of view:

Based on the procedure you just described:

I. Indirect Effect is significant.
II. Direct effect not significant. --> Full mediation.

Based on Hair et al. 2014:

I. Direct effect (without mediator) significant.
II. Including mediator.
III. Indirect effect significant.
IV. Access VAF.
V. VAF > 20 %, but < 80 %. --> Partial Mediation.

These a two different results and I´m not sure on which procedure I should rely on. (Sample size and multicollinearity do not play a role in may analysis)

Best Regards

Philip
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:

Re: Mediation Analysis - VAF or Path Significance

Post by cringle »

We updated the procedure described in Hair et al. (2014) in the second edition of the book: Hair et al. (2017). Please check the full chapter on mediation. Only if your results is a complementary partial mediation, the VAF may make sense; otherwise, you do not need that criterion.

Best
Christian
PhilipBeu
PLS User
Posts: 15
Joined: Tue May 09, 2017 8:01 am
Real name and title: Philip Beuße

Re: Mediation Analysis - VAF or Path Significance

Post by PhilipBeu »

Dear Prof. Ringle,

thank you very much - I will take a look in to the 2017 Version.

For now, in my case I can rely on the path significances, which means, I do have a full mediation?

Once again thanks for taking the time to answer all questions in this forum!

Best regards

Philip
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:

Re: Mediation Analysis - VAF or Path Significance

Post by cringle »

Yes! Step 1: Is the indirect effect significant? --> Yes --> Step 2: is the direct effect sign.? --> No; Results: full mediation.
See attached figure.

Good luck with your research!

Best regards
Christian
Attachments
mediation.jpg
mediation.jpg (178.66 KiB) Viewed 19721 times
PhilipBeu
PLS User
Posts: 15
Joined: Tue May 09, 2017 8:01 am
Real name and title: Philip Beuße

Re: Mediation Analysis - VAF or Path Significance

Post by PhilipBeu »

Dear Prof. Ringle,

Thank you very much!!!

Best Regards

Philip
PhilipBeu
PLS User
Posts: 15
Joined: Tue May 09, 2017 8:01 am
Real name and title: Philip Beuße

Re: Mediation Analysis - VAF or Path Significance

Post by PhilipBeu »

Dear Mr. Ringle,

I looked up the 2nd Edition of the book, which was very helpful.

I still have one last question:

I included a Mediator although the direct path was not significant (like Hayes 2009 states not a neccessary condition). Following the procedure above I would have the result of full mediation. But there is a sing change and VAF is above 100% so the included variable is a supressor variable.

How can I be aware of this for future research? Should the figure above be updated, so young researches don´t run into assuming a full mediation, because I read the full chapter on mediation and VAF is not mentioned.

Best Regards

Philip
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:

Re: Mediation Analysis - VAF or Path Significance

Post by cringle »

Please take a look at
Zhao, X., Lynch, J. G., & Chen, Q. (2010). Reconsidering Baron and Kenny: Myths and Truths about Mediation Analysis. Journal of Consumer Research, 37(3), 197-206

There, you find some additonal notes on supressor effects.

Bottom line: You would not come to that conclusion since you only analyze the model including the mediators when following the updated procedure by Zaho et al. In a full mediation, the VAF is not of interest (i.e., 100% since the direct effect is not significant). The VAF only works in a complementary partial mediation and not in a competitive partial mediation.

Best
CR
Yuna
PLS Junior User
Posts: 2
Joined: Fri Jul 07, 2017 11:02 am
Real name and title: Ms. Yuna

Re: Mediation Analysis - VAF or Path Significance

Post by Yuna »

Dear All

I am using PLS 3 to test for mediation. In the latest PLS book (2017), there is no mention of VAF. I have read Zhao et al. (2010) and again there is no mention of VAF.

I have a partial mediation (complimentary mediation) with VAF=18% which according to Hair et al. (2010) is a nearly zero mediation. Is it important to report the VAF in this case or should I only mention that there is a complimentary mediation.

Thank you.

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

Re: Mediation Analysis - VAF or Path Significance

Post by jmbecker »

Current textbooks and journal articles usually do not mention the VAF, because it has been critized in the past.
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
Yuna
PLS Junior User
Posts: 2
Joined: Fri Jul 07, 2017 11:02 am
Real name and title: Ms. Yuna

Re: Mediation Analysis - VAF or Path Significance

Post by Yuna »

Can anyone recommend good journals that have criticized VAF?

Thank you.
ilma
PLS Junior User
Posts: 1
Joined: Tue Feb 28, 2023 4:35 am
Real name and title: ilma gulshan

Re: Mediation Analysis - VAF or Path Significance

Post by ilma »

Hi
Please tell me where is path coefficient values after running bootstrpping in smartpls so to get VAF
Thanks
Post Reply