Different results testing multiple regression analysis and multiple regression via mediation testing via Hayes Process Macro

Hi,

For writing my thesis I tested the relationship between attitudes toward organizational change and job satisfaction. As visualized in my conceptual model below, attitudes toward change have been set as a mediator, therefore I tested all my relationships via the Hayes Process Macro V4.2 (Hayes, 2022).

When testing this relationship between attitude toward organizational change and job satisfaction, no significant relationship between these two variables in the mediation model was found.
However, when testing the relationship via multiple regression in SPSS the relationship was found to be significant.

My questions are:

  1. How is it possible that two similar analyses give different results
  2. What result should I use? It would be convenient to accept the hypothesis
  3. How to report, in case I can report both (it was found in earlier studies that the relationship is very ambiguous)

reshaped conceptual model.drawio

Regards, Floris, MBA Student Groningen NL