PSYC3010 Lecture Notes - Lecture 9: Regression Analysis, Sobel Test
Document Summary
Lecture 9 moderation, mediation and indirect effects. Mediatio(cid:374): iv (cid:862)(cid:272)auses(cid:863) dv through (cid:272)ha(cid:374)ges i(cid:374) the (cid:373)ediator (cid:894)(cid:862)(cid:271)e(cid:272)ause(cid:863)(cid:895) Dv (path a) (path c) (path b. c) (path c. b. Iv is related to dv: mediator is related to dv when effect of iv is controlled for. Iv is no longer related to dv when effect of mediator is controlled aka path (cid:272)(cid:859)(cid:895) R2 change and coefficients (b or ) for. Step 3: conduct a sobel test or bootstrapping (e. g. , online) if sig, there is an indirect effect of iv via mediator on dv. Complexities of mediation: sometimes people show path a (iv to the mediator) just with a correlation. The zero-order association (and direct effect in block 1) can be ns because the iv is raising the dv by one process at the same time as it is lowering it by another. If you have both mediators you can easily show this.
