@Marcus You won't be able to check the significance moderators beforehand as you have mentioned that you want to moderation effect in a mediation model (moderated mediation). Therefore, run multiple models in PROCESS Macro for SPSS.

Imran Anwar Thank you for your quick feedback! Given that I use indicator coding in Process and I have 4 groups in my categorical IV, I would need to estimate the model three times anyways to get all comparisons across groups. Together with three moderators which I have to run seperately, I therefore calulcate the model 3x3=9 times? Do I see that correctly?

Is that common?

Thank you so much

I recommend using the free program JAMOVI, it is just as practical as PROCESS, you can include multiple moderating variables and review to determine which of the routes has a significant effect, these effects can be factored similar to ANOVA, so it can provide results for each category of the independent variable.

I do not recommend to use separate models, since you will not catch the possible interdependences between the moderators and interaction terms. Even if you are not interested in the higher order interactions, the two-way interactions are possibly not independent and therefore should not be inspected separately (same logic as bivariate correlation/regression vs. multiple regression). PROCESS is a nice tool, but you should not limit yourself, just because your are used to PROCESS and it is handy. If the tool your are using is not capable to estimate your model you should change the tool and not the model!! It is the model you want to test (presumably because you have specific hypotheses) and not the tool itself.

Threrefore, switch more flexible programs (I do not know how JAMOVI will handle this) but any program thay can handle structural equation modeling should do it. Very popular is the lavaan package in R for example. Here you can create complex models quite easily and calculate any indirect effects, which you can specify manually at will.