- Joined
- Jan 9, 2009
- Messages
- 18,636
- Reaction score
- 34,552
- Points
- 148
When does MODERATION become MEDIATION? Like, when does a factor or variable go from MODERATING the effect of x on y to LARGELY ACCOUNTING for the effect of x on y? Where is the line? Is there a statistical test that delineates moderation from mediation?
Normally when correlation > .5, assuming of course that there is no multicollinearity among x, y, and your unnamed independent variable.