1
$\begingroup$

I tested a mediation model and found the total effect from independent variable to dependent variable is significant. But both indirect and direct are insignificant.

Can I still report this model in my paper? What is the potential explanation for this phenomenon?

$\endgroup$
1
  • 3
    $\begingroup$ Perhaps you could expand on your model a little? At least for me, this is somewhat unclear, although I think I know what you mean... $\endgroup$
    – jbowman
    Aug 27, 2018 at 20:22

1 Answer 1

2
$\begingroup$

You can definitely report this. This is likely a power issue. You have enough power to detect the full total effect, but decomposing the effect into its parts yields mediation effects that are too small to detect. It doesn't mean they aren't there, and there must be some reason why the IV and DV are related, but your sample size and the size of effects are not large enough to actually determine how the decomposition might work in the population.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.