Several Mediators

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
laura1291
PLS Junior User
Posts: 3
Joined: Fri May 25, 2018 9:15 am
Real name and title: Laura Braun

Several Mediators

Post by laura1291 »

Hi guys,

at the moment I'm trying to figure out on how to analyze mediation effects in the following model:

Image

(Of course, in general: bootstrapping -> specific indirect effects)

So my questions...
1) Do I simply report A -> M1 -> M3 -> X and so on?

2) How do I get the following "combined" effect: M1 -> M3 -> X? Smartpls does not provide this one, or is it just irrelevant? I'm curious because in some articles there were similar models and they only stated the combined effect M1 -> M3 -> X while ignoring A/B/C -> M1 -> M3 -> X ...

3) Do I need to conduct any further analysis (in SPSS/PLS) regarding this model when analyzing mediation effects?

4) In my theoretical model there is no direct effect from M1 to X or Z. But full mediation is highly unlikely. So in order to analyze indirect effects of M3 to X and Z correctly I should also draw arrows from M1/M2/A/B/C to X/Z, right (if there is a direct effect...)?

Any ideas?
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: Several Mediators

Post by jmbecker »

1) For how to report a mediation analysis a good source is the PLS book: https://www.smartpls.com/documentation/ ... s-sem-book
You will also find some information in our online sources: https://www.smartpls.com/documentation/ ... /mediation

2)
Currently we only report the specific effects from exogenous to all possible constructs. It was a design decision that the thought is plausible from our fields of research, but as many people are requesting also specific indirect effects starting from endogenous constructs we have added that to the new small updated comming soon.

3)
That depends on what you want to report. Usually you get the most important results directly from SmartPLS.

4)
Yes. Testing if there is no direct effect (and therefore including the direct effect into the model) would be best.
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
Post Reply