Multiple mediation

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
rathor1072@gmail.com
PLS User
Posts: 16
Joined: Wed May 18, 2016 6:59 pm
Real name and title: shekhu

Multiple mediation

Post by rathor1072@gmail.com »

Hi
Can antone guide to how to calculate indirect effect of each mediator in a multiple mediator model?
I downloaded the sheet given for calculation on the website, but still can’t understand how I can do that.
Please help- rathor1072@gmail.com
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: Multiple mediation

Post by cringle »

The book on PLS-SEM includes a chapter on (multiple) mediation: Hair, J. F., Hult, G. T. M., Ringle, C. M., & Sarstedt, M. (2017). A Primer on Partial Least Squares Structural Equation Modeling. 2nd Ed. Thousand Oaks: Sage.

Best
Christian

Image
j_terstriep
PLS User
Posts: 19
Joined: Sun Jul 31, 2011 3:49 pm
Real name and title:

Re: Multiple mediation

Post by j_terstriep »

cringle wrote:The book on PLS-SEM includes a chapter on (multiple) mediation: Hair, J. F., Hult, G. T. M., Ringle, C. M., & Sarstedt, M. (2017). A Primer on Partial Least Squares Structural Equation Modeling. 2nd Ed. Thousand Oaks: Sage.

Image
I have carefully read the chapter 7 of the above book, which is very quite clear as regards multiple mediation if combined with the Excel sheet you provide for calculating the indirect effects. Nevertheless, to me two questions remained open:
  • Does a multiple mediation only occur if an exogenous latent variable is mediated by at least two mediators, or does one also refer to multiple mediation in the case that one has multiple exogenous variables which are mediated by the same mediator (e.g., IV_1 -> M -> DV and IV_2 -> M -> DV)?
  • Is it common to also conduct a mediation analysis for models where such effects where not hypothesised?
Many thanks in advance for your support.

Best regards,
Judith
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: Multiple mediation

Post by cringle »

Hi
1. Does a multiple mediation only occur if an exogenous latent variable is mediated by at least two mediators, or does one also refer to multiple mediation in the case that one has multiple exogenous variables which are mediated by the same mediator (e.g., IV_1 -> M -> DV and IV_2 -> M -> DV)?
On usually refers to the first case. However, also the relationship between two endogenous latent variables may be mediated by one or more mediator constructs.
2. Is it common to also conduct a mediation analysis for models where such effects where not hypothesised?
Yes, that what you usually do. Otherwise, it's a kind of random search since a more complex model easily includes many mediation model (intended and unintended).

Best
CR
jbielefe
PLS Junior User
Posts: 6
Joined: Wed Jul 26, 2017 3:57 pm
Real name and title: Jonas Bielefeldt

Re: Multiple mediation

Post by jbielefe »

Hi,

referring to the prior question ("is it common to also conduct a mediation analysis for models where such effects where not hypothesised") I wonder, if I also need to report the total effect (indirect + direct) even when I hypothesized no direct effect between a LV and a DV (only indirect effetcs, for example LV -> M1 -> M2 -> DV)? Of course I will report the (specific + total) indirect effect/s...

Moreover, I'm not sure if I need to use the one-tailed or two-tailed test statistic for assesing mediation effects (I used the one tailed statistic for the PLS algorhytm since I assume negative effects on the DV).

Thanks very much!
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: Multiple mediation

Post by cringle »

@1: There is no general rule. It depends on what you like to analyze.
@2: Again, it depends. If only one direction is possible, you do a one sided test, otherwise conduct the usual two-sided one.

Best
CR
jbielefe
PLS Junior User
Posts: 6
Joined: Wed Jul 26, 2017 3:57 pm
Real name and title: Jonas Bielefeldt

Re: Multiple mediation

Post by jbielefe »

Thank you very much!!!
jbielefe
PLS Junior User
Posts: 6
Joined: Wed Jul 26, 2017 3:57 pm
Real name and title: Jonas Bielefeldt

Re: Multiple mediation

Post by jbielefe »

Hi,

sorry for bothering you again, but there has been coming up another question referring to multiple mediations (which I could'nt find in the forum).

Is it possible (and advisable) to interpret a signif. total (indirect) effect even when one specific indirect effects is not significant (and the second one is significant)?

Thank you very much!
jmbecker
SmartPLS Developer
Posts: 1282
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Re: Multiple mediation

Post by jmbecker »

It always depends on the implications you want to draw (and can draw based on the results).
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
jbielefe
PLS Junior User
Posts: 6
Joined: Wed Jul 26, 2017 3:57 pm
Real name and title: Jonas Bielefeldt

Re: Multiple mediation

Post by jbielefe »

Hi again,

I'm still thinking about your answer: Since I'm more interested in the net effect of one LV on the DV instead of the specific mediations as such, am I right that I can still interpret the total indirect effect even though only one (out of 2) specific indirect effects is significant? Thanks very much for your support!
jmbecker
SmartPLS Developer
Posts: 1282
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Re: Multiple mediation

Post by jmbecker »

If you are interested in the net effect, shouldn't you be interested in the total effect and not the total indirect effect?
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
jbielefe
PLS Junior User
Posts: 6
Joined: Wed Jul 26, 2017 3:57 pm
Real name and title: Jonas Bielefeldt

Re: Multiple mediation

Post by jbielefe »

Hi Dr. Becker,

thanks for responding. Yes, I got your point. But due to my model specification, I only hypothesize an effect from A (Complexity) -> B1 (Attitude) and B2 (Subjective Norm) -> C (Intention) (multiple mediation), and no direct effect from A -> C. Am I right, that in this case the total effect = total indirect effect (as the outcomes of the PLS Algorhytm indicates)?

Thank you so much!
jmbecker
SmartPLS Developer
Posts: 1282
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Re: Multiple mediation

Post by jmbecker »

Yes, if you have no direct effect in your model the total effect = total indirect effect.
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
jbielefe
PLS Junior User
Posts: 6
Joined: Wed Jul 26, 2017 3:57 pm
Real name and title: Jonas Bielefeldt

Re: Multiple mediation

Post by jbielefe »

Thank you very much!!!
Victoria01
PLS Junior User
Posts: 6
Joined: Wed Dec 20, 2017 11:33 am
Real name and title: Victoria Oers

Re: Multiple mediation

Post by Victoria01 »

Hi all,

I just wanted to do the calculation of multiple mediating effects in the Excel-Sheet provided online (found it on https://www.pls-sem.net/app/download/15 ... 1485782413), but I just read in another thread that @jmbecker stated that in SmartPLS version 3.2.6, "those specific indirect effects with all necessary calculations" are included.

Now, does that mean that for the analysis of multiple mediation, the SmartPLS output of the specific indirect effects suffices?
Therefore, for an interpretation, I would look at the total indirect effect, check its significance, and then further elaborate on the specific indirect effects, correct?

Thanks a lot!
Post Reply