I want to specify a path diagram with 7 (observed) variables. It includes the following two relationships:

b ~ a; c ~ a+b.

Analyzing these three variables separately, it is possible to define a simple moderator mediation model: b is both a mediator and a moderator of the relationship between 'a' and c (or 'a' is a moderator in the relationship between b and c). All variables are continuous.

I would like to include this moderator effect in the overall (bigger) path diagram, with seven variables, ie., I would like to define the interaction variable ab=a*b and to include the following relationship in the path diagram:

c ~ ab

My doubts are:

1) Does this variable/path make sense in the overall path diagram?

2) In the affirmative case, I would like to know whether it is necessary to include the correlations

ab ~~ a; ab ~~ b

and require 'fixed.x=FALSE' in the 'sem' function (as it happens in the simple moderated mediation model, according to https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4616155/pdf/nihms685520.pdf).

Can you please tell me your opinion?

Kind regards,



Your Answer

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

Browse other questions tagged or ask your own question.