Re: [R] Standard Error for difference in predicted probabilities

I think you can use the bootstrap to obtain the std error. My attempt for your problem and data is below. I would be interested if anyone can point out a problem with this approach. Darin y=rbinom(100,1,.4) x1=rnorm(100, 3, 2) x2=rbinom(100, 1, .7) diff <- vector(mode="numeric", length=200) for

[R] Standard Error for difference in predicted probabilities

Is there a way to estimate the standard error for the difference in predicted probabilities obtained from a logistic regression model? For example, this code gives the difference for the predicted probability of when x2==1 vs. when x2==0, holding x1 constant at its mean: y=rbinom(100,1,.4)