Hi,
recently I run Pearson's correlation analysis of variables X and Y and I found not significant result.
I run mediation analysis using X as an independent variable and Y as an dependent variable, and Z as possible mediator. Here I found a significant effect of X on Y (significant total effect).
I don't want to ask you about mediator Z here.
I wonder does it make sense that X has a significant effect on Y when X and y are not correlated?
I know that two analyses are different, but I'm not sure whether significant correlation between X and Y is required for mediation analysis.
Thanks for your help in advance.
Bests,
Inseon
Hello In-Seon,
I I can reassure you that yes, it is possible to find a significant mediation effect in the absence of a significant correlation from X to Y. However, note that different approaches to mediation and moderation provide different confidence intervals. It is worth thinking about the assumptions underlying the approach you have chosen before proceeding to interpreting it.
Hello,
Can anyone recommend an article that states the above? I have the same issue regarding no correlation but have a direct effect and would like to read more on this.
Thanks for your help in advance,
Tracey
- Badges
- Science topic
- Similar topics
- Psychology