What method is used to calculate confidence intervals in R's MASS package function confint.glm?

The helpfile states that:

These confint methods call the appropriate profile method, then find the confidence intervals by interpolation in the profile traces.

Can anyone point me in the direction of some literature that could help me to understand what this means?

  • 3
    $\begingroup$ See this previous answer. $\endgroup$ – Cyan Mar 14 '12 at 5:47
  • $\begingroup$ @Cyan Thanks. Is it adequate to say in a manuscript that "confidence intervals were estimated using the profile likelihood method (Venables & Ripley, 2002)" ? If so, add this as an answer and I'll accept it. $\endgroup$ – fmark Mar 14 '12 at 7:33
  • 2
    $\begingroup$ I also provided some references in this response. I think you can find additional references by looking at the ProfileLikelihood package, or this on-line course. $\endgroup$ – chl Mar 14 '12 at 10:14

To summarise the above comments for posterity, the confidence interval is known as the "profile likelihood confidence interval".

An explanation of the method is given by Stryhn and Christensen, and in Venables and Ripley's MASS book, §8.4, pp. 220-221.

It has weaker assumptions that the better known Wald method, but requires more computation.

| cite | improve this answer | |

See this document for a pretty good explanation of how profile confidence intervals are determined: http://www.math.umt.edu/patterson/ProfileLikelihoodCI.pdf

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.