[R-SIG-Finance] dummy variables in regression

[markleeds at verizon.net]

I don't remember what I said in my previous message but I was playing 
with something
else and realize that, if I said contr.sum, that was wrong with respect 
to their being a zeroed
out effect for the baseline. the contrasts that do that are 
contr.treatment and they are
the default in R. I always get this stuff backwards.