No mediation, how mentioning this in discussion?

[veggie]

I performed the mediation hypothesis described by baron and kenny and eventough all 4 paths for mediation were significant and the standard regression coefficient reduced (indicating partial mediation), the Sobel test showed that this was not the case.

So my question is, how do I report this in my discussion and conclusion?

Thnx

[veggie]

Thus: how can you explain the lack of mediation? what is exactly going on?

Can you say that the relationship between te dependent variable and the independent variable goes via other variables than the variable you were mentioning?

[Lazar]

This would be easier if you spelled out your findings. My guess is that your paths from x to m and from m to y were small and/or standard errors were large.

You could try the delta method which is slightly more robust or try bootstrapping the standard errors.