Moderating effect analysis by comparing group differences
Posted: Wed Jan 18, 2017 2:38 am
I want to conduct a moderation analysis by comparing group differences as the moderators are either categorical (e.g. gender) or ordinal mostly.
My model includes both second-order formative and first order reflective constructs so I used two-stage approaches to get the main model. But I am wondering at which stage I should use the categorical variable (stage 1 or stage 2) to identify moderating effects. At stage 2 all the latent variable (LVs) scores are standardized but the categorical is not. Can I use it without standardising in stage 2?
I am using SmartPLS 3.2.4 version, this has MGA function. It provides results for 2 groups only. Can I use it comparing the means for more than two groups by running it simultaneously? Is it appropriate?
Thanks
Mina
My model includes both second-order formative and first order reflective constructs so I used two-stage approaches to get the main model. But I am wondering at which stage I should use the categorical variable (stage 1 or stage 2) to identify moderating effects. At stage 2 all the latent variable (LVs) scores are standardized but the categorical is not. Can I use it without standardising in stage 2?
I am using SmartPLS 3.2.4 version, this has MGA function. It provides results for 2 groups only. Can I use it comparing the means for more than two groups by running it simultaneously? Is it appropriate?
Thanks
Mina