Message-ID: <BEE4E369-0552-4646-8A7E-4B58CB4EBCF0@comcast.net>
Date: 2012-05-12T02:58:35Z
From: David Winsemius
Subject: Multiple regression Categorical data
In-Reply-To: <A29987079A478A4A91B7210BC46CE0D05B576A312A@exmbx3.ad.slu.se>
On May 11, 2012, at 9:13 AM, Rosario Garcia Gil wrote:
> Hello
>
> It is possible to set up an lm() model where none of the categories
> of the categorical independent variable need to be used as
> references, I mean use the total mean instead.
Yes.
?contrasts
--
David Winsemius, MD
Heritage Laboratories
West Hartford, CT