If I calculate my results with Model 4 beforehand, all results are traceable and correct. But if I switch to Model 58 and let my moderator act on the paths, my signs become negative. Why do the signs change so much in Model 58?
Thanks
Can you show the outputs? And why did you caculate model 4 in the first place?
P.S.: should have more than 1 indirect effect (in PROCESS 3 as default), since they are conditional indirect effect, conditional on W. Therefore, you should see 3 indirect effects for different levels of W.
We count on Model 4 first to research the effect of emotions on user behavior. Then I want to let a moderator work on Model 58 to check whether the effects are strengthened or weakened (I also checked this on Model 7). First I have x = .31 over Model 4 and x = -. 90 over Model 58. Can it be that the value changes so massively into the negative?
This picture is really not very helpful to assess your results. The output from your whole analyses are needed. Besided that, apparently you are talking about the paths from X-->M which change in presence of moderator. This was not clear from your original post, since you are talking about mediation, therefore the focal interest is typically on the indirect effect, for which there would be 3 by default in the PROCESS model. BUT do I guess correctly that your X and M variable are categorical?? Or why are there so many X and interaction effects in the model?
No chance helping you without seeing the WHOLE output(s). With a multicategorical X (5 groups?) you should have 12 indirect effects in model 58.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling