Hi all,
I'm struggling to find the appropriate analysis for my experimental study. I've got the following variables:
Independent:
Info available (dichotomous: experimental vs. control group)
Dependent:
Consumer behaviour (Continuous)
Willingness to pay (Continuous)
Moderator:
Environmental awareness (Continuous Moderator,)
Do I first conduct t-tests between environmental awareness and each DV for each individual IV group? But then what is the next statistical analysis to account for the moderator?
Hello Hannah. With two DVs, I think the first thing you need to figure out is whether you want to do two univariate analyses or one multivariate analysis. Notice that I am using univariate and multivariate to describe the number of DVs, not the number of explanatory variables. The classic article by Huberty & Morris (1989) is a good resource on this issue. HTH.
https://pdfs.semanticscholar.org/ee6c/77c99c8e4530d0cccaedf85ed525fb22a02d.pdf
Hello Hannah,
After you've answered the multivariate / univariate question posed by Bruce Weaver , a classical approach to analysis would be a two-stage regression:
Stage 1: IVs = Group membership and moderator.
Stage 2: Add an interaction IV (group x moderator).
DV = one or both measures, based on your choice of univariate/multivariate.
If stage 2 adds a significant increase in explanatory power (R-squared increase), then this is evidence of a moderation effect.
Good luck with your work.
Bruce, thanks for the plug. The full-text for that PB article is also here on RG, and, of course, I agree :-)
David's advice is also dead-on.
No worries, John. I recommend that article frequently because so many people still believe that a multivariate test prior to 2 or more univariate tests provides some kind of protection (a la Fisher's protected t-test). I also ask the students in one of my classes to read it and write a brief summary as one of their assignments.
Cheers,
Bruce
Cool! I am told by my colleagues in Psy that it has had an effect, but I suppose that the effect size is questionable:-) Tradition is difficult to overcome.
How to do the multi-group CFA for moderator of latent construct? ( Moderator is in between Mediator and DV, Moderating variable is Work-Life Balance). Looking for any guidance on how to do it in AMOS. Thanks
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis