Hi everyone I am doing a moderated mediation test using R Lavaan. I have a binary X, continuous M, Y and moderator. I got an insignificant index of moderated mediation which, to my understanding, means no moderating effect. However the indirect effect of X on Y through mediator is significant at lower moderator level, but not at mean or higher level. In other words, for the below parameters, only 'indirect.SDbelow' is significant.
indirect.SDbelow := a1*b + a3*-sqrt(moderator.var)*b
indirect.mean := a1*b + a3*moderator.mean*b
indirect.SDabove := a1*b + a3*sqrt(moderator.var)*b
I am trying to interpret the results. Does this still mean there is difference across different moderator levels, though it does not reach statistically significant level? I wonder whether anyone could provide any thoughts or suggestions. Thanks so much!
Your results shows that there is no significant difference across different moderator levels.
check whether the moderated mediation is significant at 90% CI level.
Then, discuss that even though the moderated mediation is marginally significant, the pattern of your results is consistent with your theoretical arguments and suggest future research to test the hypotheses using different samples or in different contexts to
There is no difference between the levels of moderation and should justify in discussion their significant results. In the free JAMOVI program it is possible to separate the mediating effects by each level or category of the moderator variable. For example:
Article Exploring the Complex Pathways between the Fear of COVID-19 ...
Dear Cristian Ramos-Vera and Tae-Yeol Kim ,
Thank you so much for your reply! I now understand that as the IMM (Index for Moderated Mediation) is not significant, there is no different between the levels of the moderator. May I further ask, about the specific meaning of the significant 'indirect.SD.Below' index? Does it still mean that at a lower level of the moderation, the mediation is significant? And as my moderator is a continuous variable, I wonder that when I calculate this indirect effect at -1SD level moderator, does this significance include the group of cases with exactly [Mean-1SD] score on the moderator, or a range of cases with a moderator score below [Mean-1SD]?
In R Lavaan this index is calculated as:
indirect.SDbelow := a1*b + a3*-sqrt(moderator.var)*b
Thank you again for your help!
Zhaotian
- Badges
- Science topic
- Similar topics
- Engineering
- Thermal Engineering
- Thermofluid
- Heat Transfer