I am working on a SEM where I hypothesize a causal chain where A influences B, which in turn influences C. The data are from a 2x2 between-subjects experiment, N = 297.
All the goodness of fit indices look very good: chi-square p-value = .07 CFI = .991 TLI = .989 RMSEA = .028, PCLOSE = .992 Both the A>B path and the B>C path are significant, as predicted.
Now, here is my problem. The A>C path is not significant, neither before nor after adding the B variable. Based on this:
- Would it be correct to claim that A has "an indirect effect" on C? Or what would be a more correct description?
- Is this a case of mediation or of some other form of indirect effect?
- Should there always be a A>C path in models with indirect effects of this kind?
Thanks so much in advance for your kind help!