Differences between moderator and mga

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
MariaLunde
PLS User
Posts: 11
Joined: Wed Dec 19, 2018 2:23 pm
Real name and title: Maria Lunde

Differences between moderator and mga

Post by MariaLunde »

I would be very grateful if someone could help me with an issue in PLS-SEM.

I have conducted an MGA where I have found differences between group A and group B. I want to check if gender affects the results between group A and group B.

I have done this in two different ways:

1: Use of moderator: Using a moderator on the MGA, I found that gender moderates the results of hypothesis 1. Hypothesis 2 is not affected by gender.

2: Use of MGA: I divided the original sample into women and men. Note! The sample is so large that even such a division gives several thousand respondents. Using MGA for men and women respectively, on group A and group B, I found that gender affects the results of hypothesis 2. Hypothesis 1 is not affected by gender. As we only have two groups, no Boniferoni correction has been applied.

The use of a moderator shows that gender affects the results of hypothesis 1, while the use of MGA (gender) on the MGA analysis shows that gender affects the results of hypothesis 2. How is this possible? What to trust?
Post Reply