[R] GLMMPQL and negbinomial: trouble with the X-axis in PREDICT

Erin Latham els.latham at gmail.com
Wed Oct 28 18:12:15 CET 2009

Ahh, yes. The level made all the difference. I originally thought the
level only applied to binomial response variables, so I disregarded
And I'll just add that >plot(e) below still produced an indexed x
axis, so I just combined the vectors and then was able to plot the
scaled variable properly.

> MyData<-data.frame(BERRIES = seq(from =20, to=3600, by=100))
> e<-predict(m1,MyData,type="response", level=0)
> q2<-cbind(MyData,e)
> plot(q2$BERRIES,q2$e)

Thanks for the fast responses!


Erin Latham, MGIS Candidate
Geography Dept, University of Calgary, AB, Canada

More information about the R-help mailing list