Thank you in advance for you help. I am just learning moderation and mediation analysis. In my reading about mediation analysis I never found anything that says the mediator and IV must predict the DV in the same direction. That makes sense in hindsight. If that is the case, I don't know what I have. Here is my issue: a - IV, b - moderator/mediator, c - DV b is not a moderator. There is no significant interaction between a and b. a signifcantly negatively predicts c b significantly positively predicts c a significantly negatively predicts b a + b significantly positively predict c (in this model, a no longer significantly predicts c at all) It works in Baron and Kenny's analysis of mediation and meets the Sobel test as well. But since a and b don't predict c in the same direction is it still mediation? Can it be suppression even though a predicts b (suppression makes strong theoretical sense)? Or is this nothing? Sorry, just learning all of this and have clearly made a rookie mistake.