When I'm doing exploratory analysis I was thinking I could approach it like this:
I have an outcome I'm interested in, in my case unprotected sex, and I find in the multivariate model for that, that condom self-efficacy is significantly associated with unprotected sex.
So My next step is to now do an exploratory model with condom self-efficacy as the outcome and see what is associated with that. Then if I find any variables at that stage, associated to condom self-efficacy, I would then check to see if the was mediation or moderation happening with that variable.
1)Take outcome1,find multivariate model1,find independent1 sign. assoc. with outcome1. 2)take independent1 from model1 as outcome2, what multivariate assoc with outcome2? 3)Now check for mediation/moderation? What if there are multiple variables associated at each step?
the simplest case is when you have one mediating variable that stands between your background/control variables and a single dependent variable. Moderation is a different matter, which basically consists of interaction effects with the moderating variable.
Just in case you don't have it, I have attached a copy of the original Baron & Kenny article on mediation and moderation.
Hi,
newly published papers on mediation and moderation with different methods of statistical analysis are attached. hope they clarify your concern.
best regards
Hi Gabriel. The above resources may be more helpful and already answered your questions. If not, below is listed a very easy to understand resource for mediation and another for moderation. It gives you the step-by-step process of how to do it, the assumptions, and the interpretation. I think there are even links to syntax for spss. Best of luck.
http://davidakenny.net/cm/moderation.htm
http://davidakenny.net/cm/mediate.htm
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling