This question is about an issue when working with an exploratory approach in mediation. When we have three variables a, b ,c (where b is the possible mediator, a is predictor, c is the outcome) and have acceptably fit levels on both the partial and full mediation solutions. In my logic when we have partial mediation, we can not have full mediation and I feel that partial model is a more plausible model. However, I wonder whether it is the other way around. I work with AMOS and use bootstrapping. Answers or references on either position would be appreciated.

More Çınar Kaya's questions See All
Similar questions and discussions