Hello,
How to interpret a positive indirect effect, in a moderation analysis when path a and path b are negative (and significant)?. Also, the total and direct effects are positive.
Hello Miri,
I guess you mean "mediation analysis"? The indirect effect is simply the product of the direct effects. Hence, if both effects are negative, the result is positive in total. As an example, imagine that the "degree of back pain" reduces wellbeing (negative effect) and wellbeing is again has a negative effect on "complaining behavior" (the better your wellbeing, the less frequently you complain). Hence back pain incrases complaining behavior.
The question, hence, is whether the direct effects make any sense to you. Expecially a nonsensical direct effect can point to unobserved confounding of the mediator-outcome like--hence beware the causal assumptions underlying mediation!
See here more on that:
https://www.researchgate.net/post/I_want_to_know_the_assumptions_of_multiple_mediation_analysis_with_AMOS_output_Can_someone_give_any_guidelines
HTH
Holger
Thank you Holger. Your reply is most helpful (and yes, I meant mediation analysis)
Dear Dr Miri Cohen and Holger Steinmetz ,
Thank you for the good question and very nice explanation.
All the best,
- Badges
- Science topic
- Similar topics
- Correction