Why after I connected another two paths the significant status changed

Before posting, check our FAQ to see if your question is already covered.
Post Reply
PLS Junior User
Posts: 1
Joined: Sat May 27, 2023 9:17 am
Real name and title: Phd

Why after I connected another two paths the significant status changed

Post by JKGG »

Hello all,

I am new to SmartPLS4 and got stuck with one issue.

Based on the hypothetical research model, I wanted to test two indirect effects: A1->B1->C, and A2->B2->C. C is for sure my DV. A1 and A2 are the IVs. However, I don't know in this case whether B1 and B2 also serve as IVs. I didn't expect to test the mediation effects in the original model. (Please see the attachment for the image).

As a result, I collected all my data and ran the PLS-SEM test with the above-proposed model (without mediation). It turned out that all the four paths were significant. But the paths from B1->C and B2->C are just around the critical value ( t values are only around 2.1 something). Then I checked the direct effect table and found that there were significant (though only with t values around 1.988) between A1->C and A2->C. I am thinking perhaps there exists a mediation effect.

The strange thing was that when I added two paths, which are A1->C and A2->C, the two significant paths (B1->C, and B2->C) turned out to be non-significant. And the direct effect from A1->C and A2->C are not significant at all.

I have been so confused about:
1) What is the role of B1 and B2 in my proposed model? I don't think they are mediators but seems like another two IVs...
2) If I go with my original model, all paths are significant - which is good, but if the reviewers ask if there are any direct effects from A->C, how should I respond? Actually, the direct effect table did show the result, but when I connected the arrows, there were no direct effects...

Thank you so much in advance.
path.jpeg (184.38 KiB) Viewed 828 times
Post Reply