using a co-factor when comparing two groups?

[colombeli]

Hi,

I would like to compare two groups, one treated, one control.
Usually, I would simply use a t-test, assuming normal distribution
and equal variance.
However, I would now like to use another parameter as a co-factor:
I am comparing two groups of animals and the effect I observe
could be due to a difference in weight.

Which test can account for this? I know that if I had three groups
or more I could use one-way ANOVA, but since I only have two groups...

Probably, this is a very basic question, I'm just not very firm with stats.

Thanks for your help

 

[Dason]

Is the covariate continuous? You're probably looking at a case where you want to look into ANCOVA.

 

[colombeli]

Wow, a super fast reply.

Yes it is. What exactly is ANCOVA?

 

[Dason]

ANCOVA - it's essentially ANOVA but you're also modeling the effect of a continuous covariate.

 

[colombeli]

Alright, I managed to run an ANCOVA using an online form (because stupid Excel 2008 for Mac does not support ANOVA and Prism won't let me do it either).

However, I have some trouble interpreting the values.

Is the P-value given in the ANCOVA summary table for adjusted means
the value I should be interested in?

And what does the test for homogeneity of regressions tell me?
Is the P-value for between regressions at all relevant?

Sorry to be shooting all these questions. I tried to find answers for lay people like me, but didn't...

 

[colombeli]

Alright, I think I figured it out by myself. Thanks again for the help!