I would like to fit a generalized linear model in R, using glm()
. More precisely, it's a nb.glm()
. Apparently, coefficients fitted in R packages that deal with generalized linear models are all evaluated at mean values of the other variables. Is there a way to evaluate coefficients holding the values of other variables constant at e.g. median or zero?
For example, the example from UCLA's ATS site on NegBin regression:
require(foreign)
require(ggplot2)
require(MASS)
dat <- read.dta("http://www.ats.ucla.edu/stat/stata/dae/nb_data.dta")
dat <- within(dat, {
prog <- factor(prog, levels = 1:3, labels = c("General", "Academic", "Vocational"))
id <- factor(id)
})
summary(m1 <- glm.nb(daysabs ~ math + prog, data = dat))
The resulting coefficients are evaluated at the mean value of daysabs
, which is 5.96. It find it more intuitive to evaluate the coefficients at median values.