Yes, better powered, and it uses all the data. Otherwise you throw out group 2. You can do the same thing with the other covariates (ie, scale group1's age by 1, scale group2's age by 3, etc)

On 11/29/2012 02:58 PM, Mahinda Yogarajah wrote:

Hi - thanks - I was thinking about doing it the other way (3 separate groups) because as well as looking to see whether there is a linear change in dependent measures (eg area) across groups I also want to see if there is a linear change in the slope of dependent measure against age across the 3 groups. I can' t see how to do that with the single group method against 1,2,3 covariates.

Is there a way of doing it ?

Also why do you recommend the 1 group method ? Better powered

Thanks.

Mahinda