Re: [R] CI with confint

2012-03-28 Thread Benjamin Caldwell
Oh, interesting - here I thought I was getting a confidence region for the model (which of course you're right has two parameters) and in fact confint is giving me the confidence interval for each of the two parameters in the model. The ellipse package does a nice job of helping me visualize that.

Re: [R] CI with confint

2012-03-28 Thread peter dalgaard
On Mar 28, 2012, at 21:33 , Benjamin Caldwell wrote: > Hello all, > > I'm trying to use confint from the MASS package to compute confidence > intervals for an nls object. When I plot the results, however, they don't > make sense - lines cross over the fitted model or just don't match the > data.

[R] CI with confint

2012-03-28 Thread Benjamin Caldwell
Hello all, I'm trying to use confint from the MASS package to compute confidence intervals for an nls object. When I plot the results, however, they don't make sense - lines cross over the fitted model or just don't match the data. Code is : Thanks for help dat<-data.frame(a,b) with(dat, plot(a,