MANCOVA and MANOVA Discussion > MANCOVA?
Michael,
With the limited information we are dealing with, I'd say that MANCOVA is a reasonable approach. I'd also say that the multiple covariates and DV's are a reasonable rationale (if we had only one DV and multiple covariates, you could use ANCOVA; if you had no covariates, but multiple DV's, you could use MANOVA).
Depending on how you conceptualize the relationships between variables (and your outcomes, relative to each other), I might also consider a structural equation model (SEM) and/or a hierarchical linear model (HLM; also known as MLM).
January 26, 2012 |
Jeremy Taylor
Jeremy Taylor
This is a very good discussion, I feel that the multivariate analysis of variance is a very good idea
June 21, 2012 |
Darcy
Darcy
So for a qualifying exam, I have to just have a basic idea of what analysis to perform. My proposed study is to investigate whether exposure to an intervention will lead to growth in 3 outcome measures (state test scores, high school exit exam scores, and GPA). I know the intervention is my IV and the measures are my DV, and I have several covariates (age, grade, identification for special education, type of school setting, etc). I think MANCOVA is my best choice because of the multiple covariates and dependent variables, but I'm not sure and not 100% confident with the rationale.
Help!!