0

I estimated the following Poisson regression in R:

poissmod2 <- glm(aerobics$y ~ factor(aerobics$x1)  + aerobics$x3 + aerobics$x4,
                 family = poisson))

No problem so far. The next step, however, is to determine a 95% confidence interval for aerobics$x1.

I have tried following two codes:

confint.default(poissmod2,"aerobics$x1")

and

confint.default(poissmod2,"factor(aerobics$x1)")

For both codes I get a confidence interval 'Na-Na'... This code, however, seems to work for non-factor variables.

Does anyone have an idea what I am doing wrong?

Zheyuan Li
  • 71,365
  • 17
  • 180
  • 248
user3387899
  • 601
  • 5
  • 18
  • 3
    Possible duplicate of [calculating confidence interval of coefficient in poisson regression](http://stackoverflow.com/questions/22223301/calculating-confidence-interval-of-coefficient-in-poisson-regression) If you don't know what it is a reproducible example search in google. But usually implies coping some data so others can try the same code as you to get at the same point. After having the same problem they can start thinking how to solve it easier than from just a couple of lines without any data – llrs Mar 10 '14 at 12:31
  • Can you specify how this is different from your previous question? – Ben Bolker Mar 10 '14 at 12:43

0 Answers0