Mediation in PLS SEM

Research topics can be discussed in this area.
Post Reply
PLS Junior User
Posts: 3
Joined: Mon Nov 19, 2018 5:55 am
Real name and title: Ali

Mediation in PLS SEM

Post by ccp9147 » Mon Nov 19, 2018 6:04 am


I am facing small problem and quick help will be highly appreciated

My IV --> MV and IV --> DV both have negative Beta (Original sample mean) however t and p values are significant. But due to negative Beta my alternate hypothesis is not supported

Now with regards to mediation analysis, shall I consider performing mediation analysis knowing that Beta is negative or shall I look for t values for the significance of relationship.

I applied complete bootstrapping with two tail and t values are significant but beta is negative moreover when I am applying Zhao et al,. (2010) decision tree, result is appearing as "complementary mediation"

Help will be highly appreciated

SmartPLS Developer
Posts: 971
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Re: Mediation in PLS SEM

Post by jmbecker » Thu Nov 29, 2018 3:29 pm

Both together are important. T-value/p-value (or better confidence intervals) tell you whether an effect is significant. The parameter estimate of that effects and its sign show strength and direction of the effect. Only both together make sense.
Dr. Jan-Michael Becker, University of Cologne, SmartPLS Developer
Researchgate: ... v=hdr_xprf
GoogleScholar: ... AAAJ&hl=de

Post Reply