Re: [R] SE of difference in fitted probabilities from logistic model.

On Fri, 16 May 2008, Rolf Turner wrote: I am fitting a logistic binomial model of the form glm(y ~ a*x,family=binomial) where a is a factor (with 5 levels) and x is a continuous predictor. To assess how much ``impact'' x has, I want to compare the fitted success probability when x =

[R] SE of difference in fitted probabilities from logistic model.

I am fitting a logistic binomial model of the form glm(y ~ a*x,family=binomial) where a is a factor (with 5 levels) and x is a continuous predictor. To assess how much ``impact'' x has, I want to compare the fitted success probability when x = its maximum value with the fitted probab