In glm, is there a way to specify the family to be exponential, or Gamma but with alpha=1? This is required by Dobson Exercise 4.3d, page 48. But without specifying alpha=1, the asnwer is the same as in the book. If reply to the list, please cc me a copy. Thanks. -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
glm: family=exponential() or family=Gamma(alpha=1)?
2 messages · Michael Wang, Thomas Lumley
On Mon, 26 Feb 2001, Michael Wang wrote:
In glm, is there a way to specify the family to be exponential, or Gamma but with alpha=1?
Yes. You do it in summary(), not in glm(). summary(fit) estimates the dispersion summary(fit, dispersion=1) sets it to 1 (exponential).
This is required by Dobson Exercise 4.3d, page 48. But without specifying alpha=1, the asnwer is the same as in the book.
The estimated coefficients do not depend on alpha. That's why the family object doesn't have to specify a value. The standard errors do, but if the data are really exponential the estimated dispersion will be close to 1 with reasonable sample sizes (estimating the dispersion and fixing it at the true value are asymptotically equivalent). -thomas Thomas Lumley Asst. Professor, Biostatistics tlumley at u.washington.edu University of Washington, Seattle -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._