Leave the variables as they are, don't standardize. (You never need to
standardize variables for moderation, in fact you're often better not
standardizing.)
Jeremy
On 26 November 2012 08:24, A.L.Lynch <[log in to unmask]> wrote:
> Hey everyone,
>
> I have conducted a hierarchical regression analyses to see whether masculine
> ideology moderates the relationship between 2 IVs and a DV - 4 variables
> altogether. Therefore, I have 3 models including a three way interaction.
>
> 1. As both IV's are dichotomous variables, coded as either 1 or 2, do I
> still need to standardise them or do I leave them as they are?
> 2. How do I go about simple slopes tests if the moderator is found to be
> having a significant effect? Which macro or program is recommended? I'm a
> little rusty as I haven't done these in a while.
>
> Also...
> When conducting moderation mediation using macros provided by Andrew Hayes,
> what is the best model to use for 3 IVs (one of which also acts as a
> moderator), 1 mediator and 1 DV?
>
> Any help on this would be much appreciated.
>
> Thanks,
>
> Amy
|