Need help to interpret this PLS-MGA result

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
PLS Junior User
Posts: 1
Joined: Tue Jan 10, 2023 9:21 pm
Real name and title: Jonathan Paris, PhD student

Need help to interpret this PLS-MGA result

Post by jchambersparis »

I perform a PLS-MGA and I need help to interpret one of the result.
Note that sample sizes for each subgroup were considered sufficiently large when i looked at statistical power requirements and all results for the model met the criteria to indicate partial measurement invariance.

Here is the result i need help to interpret.

For a relationship (for example X on Y), PLS-SEM result are :
Group A (β = 0.191, p < .01)
Group B (β = 0.025, p = .594)

Concerning PLS-MGA result (A compare to B), the result is (β diff = .166, p diff < .05) , so, parameter estimates are statisticly different.

My thesis supervisor told me that i don't have to interpret a result when the relatioship for 1 group is insignificant, even if the PLS-SEM reveal a significant difference. I'm a little bit confuse here....
I though this result mean that we ca confirm a difference for this relationship for these 2 groups because PLS-MGA reveal that parameter estimates are statisticly different.

I'm thinking that because I have a similar result for another group but PLS-MGA did'nt show statisticly different parameter estimates.
Group D (β = 0.347, p < .01)
Group E (β = 0.165, p = .070)
PLS-MGA = (β diff = .182, p diff = .103)
For this one, i though : even if the PLS-SEM result show that the relationship is significant for group D and not for group E, PLS-MGA reveal that these results are not enough different.

Am i wrong? Thank for your help!
Post Reply