[R] glm(family=binomial(link=logit))

Robin Hankin r.hankin at noc.soton.ac.uk
Fri Jul 15 17:00:44 CEST 2005


I am trying to make glm() work to analyze a toy logit system.

I have a dataframe with x and y independent variables.  I have

L=1+x-y          (ie coefficients 1,1,-1)

then if I have a logit relation with L=log(p/(1-p)),

If I interpret "p" as the probability of  success in a Bernouilli
trial, and I can observe the result (0 for "no", 1 for "yes")
how do I retrieve the coefficients c(1,1,-1)
from the data?

n <- 300
des <- data.frame(x=(1:n)/n,y=sample(n)/n)   # experimental design
des <- cbind(des,L=1+des$x-des$y)            # L=1+x-y
des <- cbind(des,p=1/(1+exp(des$L)))         # p=1/(1+e^L)
des <- cbind(des,obs=rbinom(n,1,des$p))      # observation: prob of  
success = p.

My attempt is:


But it does not retrieve the correct coefficients of c(1,1,-1) ;
I would expect a reasonably close answer with so much data.

What is the correct glm() call to perform my logit analysis?

Robin Hankin
Uncertainty Analyst
National Oceanography Centre, Southampton
European Way, Southampton SO14 3ZH, UK
  tel  023-8059-7743

More information about the R-help mailing list