1

I have a question about mediation using process analysis.

In Baron and Kenny's (1986) mediation, the 4 steps (below) have to be met to show mediator mediates the IV and DV relationship, right?

Step 1: Significant total effect: IV affects DV Step 2: Significant causal relationship between IV and mediator. Step 3: Significant causal relationship between mediator and DV. Step 4: Direct effect of IV on DV controlling for Mediator should be zero correlation or not significant.

For my case, my total effect is not significant, but my indirect effect and direct effect are both significant.

enter image description here

enter image description here

Does my mediation effect not make sense since the total effect is already not significant? One thing worrying me is that the correlations between variables are quite low too.

Nick Cox
  • 48,377
  • 8
  • 110
  • 156
Shirley
  • 41
  • 1
  • 6
  • I would prefer to write this as a comment but since my reputation does not allow it you can have it as an answer This has been [answered here](http://stats.stackexchange.com/a/186110) I had come across a similar situation and it seems that this phenomenon is not uncommon. Anyway the news are good and it seems that Yes you can still imply indirect effects while the total is not significant – Lefkios Paikousis Mar 01 '17 at 06:52

0 Answers0