Solved – Complete mediation (Preacher Hayes)

Below is the output of my bootstrap model (Preacher & Hayes, 2008). Path a, b and c are marginally significant (p < 0.1), but c' is not significant. I have read that when this is the case, there is reason to believe that there is complete mediation? Is that correct?

Thank you for your help!

Dependent, Independent, and Proposed Mediator Variables:  DV =   Relation  IV =   DummyRDn  MEDS = Boardinv   Statistical Controls:  CONTROL = Numberpa            Priortie            Weighted   Sample size: 101    IV to Mediators (a paths)              coeff        se         t         p Boardinv    1,5302    0,8953    1,7090    0,0907   Direct Effects of Mediators on DV (b paths)               coeff        se         t         p Boardinv    0,1160    0,0672    1,7255    0,0877   Total Effect of IV on DV (c path)               coeff        se         t         p DummyRDn    1,1070    0,5959    1,8577    0,0663   Direct Effect of IV on DV (c' path)               coeff        se         t         p DummyRDn    0,9295    0,5988    1,5523    0,1239   Partial Effect of Control Variables on DV               coeff        se         t         p  Numberpa   -0,2421    0,1090   -2,2206    0,0288 Priortie   -0,2533    0,2173   -1,1656    0,2467 Weighted   -0,7825    0,2688   -2,9107    0,0045   Model Summary for DV Model    R-sq    Adj R-sq         F       df1       df2         p 0,2002      0,1581    4,7547    5,0000   95,0000    0,0006   ************* BOOTSTRAP RESULTS FOR INDIRECT EFFECTS *************   Indirect Effects of IV on DV through Proposed Mediators (ab paths)                Data      Boot      Bias        SE TOTAL       0,1775    0,1796    0,0021    0,1689 Boardinv    0,1775    0,1796    0,0021    0,1689   Bias Corrected Confidence Intervals               Lower     Upper  TOTAL      -0,0197    0,7078  Boardinv   -0,0197    0,7078   ******************************************************************  Level of Confidence for Confidence Intervals:  95   Number of Bootstrap Resamples:  5000  

No, you cannot assume a complete mediation.

For a mediation you would need a significant direct relationship between the IV and the Mediator (a), as well as a significant direct relationship between the Mediator and the DV (b).

In addition, the decrease of the direct relationship between IV and DV without the Mediator (c) and with the Mediator (c') is rather small. In a complete mediation this difference should be larger and c' should be reasonably close to $0$. See this question for more information about testing for complete mediation.

Similar Posts:

Rate this post

Leave a Comment