Mediation analysis: 3 binary variables

[Toetanchamon]

Hi,

In my research I want to test a model with 3 binary variables:
-X = year (year x, year x)
-Y = volunteering (no, yes)
-M = caregiving (no, yes)

Since this is not possible with the Process macro, is it possible using the Baron and Kenny method?

Thanks in advance!

 

[spunky]

Hi,

In my research I want to test a model with 3 binary variables:
-X = year (year x, year x)
-Y = volunteering (no, yes)
-M = caregiving (no, yes)

Since this is not possible with the Process macro, is it possible using the Baron and Kenny method?

Thanks in advance!

No. The Baron & Kenny approach needs all variables to be continuous so the properties of OLS regression hold. You'll need to fit this as a Structural Equation Model using Path Analysis.