0

I have a variable called experience which was coded as numeric and contains 3 values (1,5,10). I changed the class to factor using df$experience<-factor(df$experience) and it changed to factor. Next I run a GLM model as 

reg<-glm(cbind(win,loss)~experience, data=df, family=binomial)

but when I get summary(reg), only one level of the experience variable shows in the table (i.e. experience10). Shouldn't there also be another categorical variable, experience5?

rcs
  • 67,191
  • 22
  • 172
  • 153
finstats
  • 1,349
  • 4
  • 19
  • 31
  • 2
    http://stats.stackexchange.com/q/26539/11849 – Roland Feb 17 '14 at 12:37
  • The categorical variable experience has 3 levels, shouldn't 2 of these levels be shown in the summary output? I understand that I will have a base case (for example experience1) which will not be shown in the output. But I am only getting one level which is the experience10. Thanks. – finstats Feb 19 '14 at 12:24
  • We can't help you without a [reproducible](http://stackoverflow.com/a/5963610/1412059) example. – Roland Feb 19 '14 at 12:45

1 Answers1

0

After using sub-samples of my data, the other levels were showing up. I guess it was just a matter of Multicollinearity where the other level was not being calculated somehow! Thanks!

finstats
  • 1,349
  • 4
  • 19
  • 31