R help - Jun 2003 - Statistical Models: Ordered vs. Non-Ordered Factors

I'd like to understand better how ordered and non-ordered factors are
treated differently when included as explanatory variables in statistical
models.  Also, it appears as though glm [family=binomial] (base) treats
ordered factors differently than lrm (design).  For example, when fitting
the same model with each function the inclusion of an ordered factor results
in considerably different coefficients.  If I do not specify the factor as
ordered my resulting coefficients (glm and lrm) are essentially identical.

 

Any personal insight or good references would be much appreciated.

 

Thanks.

 

Marc

 

 


	[[alternative HTML version deleted]]