Re: range issues in ANCOVA

Isabelle wrote:
A research scientist told me that he wants me to study some of the ranges issues that may exist with the predictor variables in his ANCOVA or Linear Model Analysis.

His understanding of statistics is intuitive, rather than rigorous.

The only range issue I can think of are that which arise when a categorical var has the values 1 and 2 (say man=1, woman=2) instead of the values 0 and 1. If one wants the gender coefficient to be meaningful, one would rather use a 0/1 variable than the 1/2 variable.

Can anybody think of other such range issues that may arise?

Your help is much appreciated.


The issue that comes to mind for me is the distribution of the covariate for your two (or however many) groups. Suppose there is a gap between the maximum value of the covariate for one group and the minimum for the other. This means that when you compare the adjusted means, you are extrapolating into no man's land (i.e., beyond the range of the observed data) for both groups.

--
Bruce Weaver
bweaver@xxxxxxxxxxxx
www.angelfire.com/wv/bwhomedir
.