Combining mediation and moderation in linear model (SPSS, PROCESS)

#1
I want to perform a mediation analysis, but also include a moderator in the relation between the independent variable and the mediator (see conceptual model as attachment). I have been following the steps of Field to do a mediation in SPSS (Hayes PROCESS tool), but now I am stuck, because I am not quite sure how to continue next. Can anyone please help me out?

Also, how would you call a model like this? To be extra precise, I am also including covariates and different measurement levels in the predictors.
- Independent : Categorical
- Mediator, moderator and dependent variable: Numerical
- Covariates: Some categorical, some numerical
 

Attachments

#2
I found an answer to my own question. It's called Moderated Mediation and can be performed with the PROCESS tool with Hayes Model 7. The covariates can also be included in this model.

As for a tutorial about performing this type of analysis, I like this one:
  • Moderated mediation in SPSS using Hayes' Process Macro (Model 7) (
    )

Hope this is of any help to someone :).
 

hlsmith

Less is more. Stay pure. Stay poor.
#3
I would be a little careful in that moderators can be fairly rare in actual practice. So unless you have good knowledge of the context or perhaps your setting is more social science where with groups and behaviors - its discovery could be spurious.

Glad you found what you were looking for. I think in the causal inference world there you can also use a counterfactual model and something called the 'recanting witness' can be an issue.
 
#4
@hlsmith

Thank you for your input. I have never heard of the 'counterfactual model'. Suppose I would like to give this method a try in SPSS, where should I start you think in terms of literature, method?

Also, with recanting witness, do you mean this can be a threat to validity?
 

hlsmith

Less is more. Stay pure. Stay poor.
#5
I would search a combination of mediation and causal inference in order to find some preliminary references!