Should I mean center a dummy IV and dummy moderating variable on a continuous DV?
I am asking because I have a dummy IV, dummy moderator and a continuous DV (mediating) variable in a moderated mediation model (hayes, model 7). I am getting very different effects from dummy IV on continuous DV, (p=0.98 vs p=0.03). So I was wondering what I should report for direct effects, since the interaction effects should be reported according to non-standardized because then the IV and moderating variable take on their normal dummy values. Is this right?
Is it possible to report direct effects according to mean centering and indirect effects according to non-mean centered IV and moderatorr?
Dear Max,
I have a very similar question as you. I am also confused about whether to report mean-centered results or not.
I also used Model 7. My IV is dichotomous, with moderator being multicategorical with 3 levels, mediator and DV being continuous.
Have you got the answer to your question and could you help me with mine?
Thank you in advance.
Best,
Shirley
Generally, in all analyses, the report has to be based on the original data. The transformed data do not make sense generally. I suggest you report only the results of the original data, or retransform the analyses to the original form and then report and interpret them.
- Badges
- Science topic