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.
@Kevin Indirect effect is nothing but the product of path a (X-->M) and b (M -->). Higher magnitude of regression coefficients for paths a and b lead to significance of indirect effect. In your case, coefficients for path a and b are .13 and .13 and thus there product accounts to .016 which is almost negligible hence leading to insignificance.
@Imran Anwar So, therefore, nothing is wrong? It's just path a at path b coefficients were just too low and not enough to have a significant product. At first, all I thought when paths a and b were significant, their indirect effect would also be significant automatically.
I have the same situation and cannot find the logic to interpret them. So in my case, path a is Attitudes on Behavioral Intentions- significant. Path b Behavioral Intentions on Behavior also significant but ab insignificant. How should I interpret this?