Tuğkan Baş Yes, it is possible for the total effect to be smaller than the indirect effect. The total effect is defined as the sum of the direct effect (which is negative in your case) plus all indirect effects (positive in your case). As a result of the opposite signs of the effects in your analysis, the sum is a smaller number than the indirect effect itself.

It is also very well possible for the indirect effect to be positive when the direct effect is negative (and vice versa). The indirect effect is the product of two or more regression coefficients. When the indirect effect is given, for example, as the product of two negative regression coefficients, it will be a positive number (whereas the direct effect may be negative).

Alternative effect size measures for communicating the meaning and practical significance of indirect/mediated effects that may be more useful in your case have been proposed for example by Preacher and Kelley:

Preacher, K. J., & Kelley, K. (2011). Effect size measures for mediation models: Quantitative strategies for communicating indirect effects. Psychological Methods, 16(2), 93–115.

Article Effect Size Measures for Mediation Models: Quantitative Stra...

Dear Tuğkan Baş

Yes, An indirect effect is calculated by multiplying the paths that make up the effect. The magnitude of the indirect effect indicates the amount of mediation through the relevant mediator variables. - Mediation can be either complete or partial.

Have a look: http://public.kenan-flagler.unc.edu/faculty/edwardsj/MEDIATE.NOT.pdf

Best Regards

Dr. Fatemeh Khozaei

Yes it is possible as the total effect remains the sum of the indirect and direct effects. As your direct effect is negative but insignificant and indirect effect is significant you have complete mediation. You still may want to think about the face validity of the mediator in terms of whether it is measuring another latent dimension as the endogenous factor rather than both being one and the same. Good theory should account for that as to how those two constructs differ and that the measurement items capture or reflect those differences.

Need some advise please on how to interpret of this one. I have performed mediation analysis using WarpPLS and the result was little bit weird. Is it possible to have insignificant indirect effect even both path a and b were significant? The a path coefficient is 0.13 pvalue 0.009 while b path is 0.12 pvaliu 0.013. Now, Indirect effect is 0.017 with 0.335 which is not significant.