[R] About categorical explanatory variables

Rafael Costa rafaelcarneirocosta.rc at gmail.com
Thu May 7 20:43:22 CEST 2015


When you want to get the individual impact of each level of a single
categorical explanatory variable on a variable continuous quantitative
response, simply add the value of the estimated coefficient to the
intercept or, more directly, just consider all levels of the categorical
variable in the regression and omit the intercept, thereby obtaining in
each coefficient estimated the individual effect of that particular
category. But with 2 or more categorical variables, the intercept becomes
the value of a reference group, jointly considering several categories, one
for each qualitative variable. The coefficients in turn will be the
difference between a specific group and the reference group. But I like to
get the individual effects of each category and no effect of group
categories.

My question is: is there any way to get the individual effects in this
case, where there are two or more categorical explanatory variables (these
may be ordered, unordered, or both types)? If so, how to proceed in R?

I have been looked, but I found nothing about it.
Any help will be appreciated. Thanks for listening.

Yours sincerely,
Rafael Costa

	[[alternative HTML version deleted]]



More information about the R-help mailing list