Prev: legend in sgplot
Next: symput parameters error
From: sousimou on 16 Mar 2010 05:44
Hi all,

My question is statistical task related.
the objective is to know what the factor (between x and z : both are
categorical variables)) that have more impact on Y (explicative
variable).
I had performed ANOVA on this way :

model : Y= x z x*z

the interaction is significative.
P (x) <0.0001
P(z) < 0.0001
P(x*z)=0.005

What is the next step to do in order to detect the factor that have
more impact on Y btween (x and z) ?

If this is not clear please send me an email? I am waiting for help.

thanks,
From: Paige Miller on 16 Mar 2010 16:31
On Mar 16, 5:44 am, sousimou <sousi...(a)gmail.com> wrote:
> Hi all,
>
> My question is statistical task related.
> the objective is to know what the factor (between x and z : both are
> categorical variables)) that have more impact on Y (explicative
> variable).
> I had performed ANOVA on this way :
>
> model :   Y= x z x*z
>
> the interaction is significative.
> P (x) <0.0001
> P(z) < 0.0001
> P(x*z)=0.005
>
> What is the next step to do in order to detect the factor that have
> more impact on Y btween (x and z) ?

You need to change your thought patterns to: X is an important
variable, Z is an important variable, and X*Z is an important
variable. Among these three, the most important is the one with the
largest sum of squares (and that assumes an orthogonal experimental
design). There is no way to narrow this down to answer your question
about only X and Z.

--
Paige Miller
paige\dot\miller \at\ kodak\dot\com
 | 
Pages: 1
Prev: legend in sgplot
Next: symput parameters error