Mediator is negative, but moderator is positive, how?
These were the results of the mediator analysis:
X - M is significant positive
M - Y is significant negative
My indirect effect is negative, so I think that means that when X gets bigger, Y gets lower, because M also gets bigger.
I did a moderator analysis with the same variables, but now the moderating effect is positive.
What does that mean? I thought it meant when X gets bigger/being a student, causes more positive experiences, which leads to more PTSD-symptoms.
Is it not contradictory with the mediating effect?
X = being a student (0= no student, 1 = student)
M = positive experiences because of trauma
Y = PTSD symptoms
Another thing was that when I run the model with all mediators I get these results. Is it right for me to say that, when all mediating variables taken together X has a negative indirect effect on Y which means that increase in X reduces A,B,C which in turn reduces
For mediation, you need to think in terms of the total effect of the variable, and the portion of that effect which is indirect. So, a variable could have a strong, positive direct effect, as well as a smaller, negative indirect effect.
For moderation, there is no necessary relationship between how a variable acts as a mediator and as a moderator. In particular, moderation involves interaction effects, which modify how one variable operates at different levels of a second variable. If both of the variables in your moderation effect are interval level, then you should search for information about: linear by linear interaction effects.
@Silke Indirect effect is calculated by multiplying the effect of X to M (a) and M to Y (b) and in your case, M to Y relationship is negative therefore, indirect is also negative. Moderation effect is calculated by regressing the product variable (X*moderator) and it seem that the correlation between interaction Variable and Y is positive thus the interaction effect is also positive.
David L Morgan Imran Anwar Thank you so much for your answers! However, there is still something I don't understand. If the indirect is negative, does that mean that being a student (X, student =1, non student=0) reduces PTSD-symptoms (Y) through the positive effect being a student has on positive experiences (M). Or does this mean that because a positive mediator reduces the relation between X and Y, a negative mediator increases the relation between X and Y. This would mean that positive experiences increase the predictive value of being a student on PTSS symptoms.
I interpreted the positive moderating effect as follows: being a student with more positive experiences leads to more PTSD-symptoms in comparison with non-students. I think this interpretation is correct.
However, I do not understand how I can incorporate the negative mediating effect and positive moderating effect in my thesis, because it seems contradictory. If the mediator increases the relation, than it would contradict the fact dat positive experiences are a negative predictor right? However, if the mediator decreases the relation between being a student and PTSD-symptoms, it is contradictory with the fact that students have more PTSD-symptoms compared to non-students?
Can someone please explain to me what the negative mediating effect means in this context?
- Badges
- Science topic