Re: Influence of combined variables on outcome?
From: Bruce Weaver (bweaver_at_lakeheadu.ca)
Date: 02/26/05
- Next message: clemenr_at_wmin.ac.uk: "Re: R versus Fortran?"
- Previous message: Marc Schwartz: "Re: R versus Fortran?"
- In reply to: Daniel Schott: "Influence of combined variables on outcome?"
- Next in thread: Richard Ulrich: "Re: Influence of combined variables on outcome?"
- Messages sorted by: [ date ] [ thread ]
Date: Sat, 26 Feb 2005 15:01:57 -0500
Daniel Schott wrote:
> Hi!
>
> Sorry, my stats are *totally* rusty. So, I need some help and advice
> for direction.
> I have the problem of a "treatment" that looks like this (to make it
> easier to understand -- and thinking about my English -- I broke it
> down to a color and sound problem):
>
> Red + Loud -> X1, X2, X3
> Red + Soft -> X1, X2, X3
> Blue + Loud -> X1, X2, X3
> Blue + Soft -> X1, X2, X3
>
> Red + Loud etc. are the "treatments", X1 to X3 are the dependent
> variables (interval scale). All treatment were conducted on the same
> "sample" (N=296).
Does that mean every subject received every combination of treatments?
I'll assume it does in my further comments.
>
> I want to know which statistical procedure I have to use if I want to
> know which variable (color or sound) has a higher influence on the
> observed differences in X1 to X3 between the four "treatments".
>
> Greez
> Daniel
Are X1, X2, and X3 three different variables, or 3 measurements of the
same variable?
If the latter, then as Art Kendall suggested, you could perform a 2x2x3
ANOVA with repeated measures on all factors. But if the former, then it
could be three separate 2x2 repeated measures ANOVAs (for X1 to X3), or
a MANOVA if you want to include all 3 outcomes in one analysis. Even if
you use MANOVA, you'll no doubt end up looking at the so-called
"univariate" tests anyway (i.e., the three 2x2 ANOVAs).
In any case, you'll have main effects of colour and sound plus their
interaction. The interaction tests the null hypothesis that the effect
of sound does not depend on colour (or vice versa).
-- Bruce Weaver bweaver@lakeheadu.ca www.angelfire.com/wv/bwhomedir
- Next message: clemenr_at_wmin.ac.uk: "Re: R versus Fortran?"
- Previous message: Marc Schwartz: "Re: R versus Fortran?"
- In reply to: Daniel Schott: "Influence of combined variables on outcome?"
- Next in thread: Richard Ulrich: "Re: Influence of combined variables on outcome?"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
|
