Re: [R] p-value from GLM

2012-03-05 Thread Dunia Scheid
Dear Peter, Many thanks for the reply. Sure, this is just an example and it makes zero sense! Regards, Dunia 2012/3/4 peter dalgaard > > On Mar 4, 2012, at 12:21 , Dunia Scheid wrote: > > > Dear all, > > > > I am fitting a GLM similar to > > > > library(MASS) > > anorex.1 <- glm(Treat~Postwt+

Re: [R] p-value from GLM

2012-03-04 Thread peter dalgaard
On Mar 4, 2012, at 12:21 , Dunia Scheid wrote: > Dear all, > > I am fitting a GLM similar to > > library(MASS) > anorex.1 <- glm(Treat~Postwt+Prewt,family = binomial, data = anorexia) I hope that's just for an example. The actual analysis makes zero sense to me... > > I have found two ways o

[R] p-value from GLM

2012-03-04 Thread Dunia Scheid
Dear all, I am fitting a GLM similar to library(MASS) anorex.1 <- glm(Treat~Postwt+Prewt,family = binomial, data = anorexia) I have found two ways of computing the p-value of the fitted model: pval1 <- 1-pchisq(anorex.1$deviance,anorex.1$df.residual) pval2 <- 1-pchisq(anorex.1$null.deviance - an