Mediated regression (parallel): Opposing effects → how to interpret total effect?


New Member
Hi everyone!

In my student assistant job I had to to test a model of mediated regression, with various variables mediating the effect of A on B. I used the SPSS process tool.
Considering that the mediating variables were conceptually related I tested the model with parallel mediation, meaning I did one test in which I tested the mediation of all possible mediators.

I found three variables to significantly mediated the effect. Some with a positive effect, some with a negative effect. My supervisor was now wondering whether it is possible to sum these effects up to draw a conclusion on the direction of the overall indirect effect?

Thank you for your help in advance!