An embedded and charset-unspecified text was scrubbed... Name: not available Url: https://stat.ethz.ch/pipermail/r-help/attachments/20060222/3fefc3c6/attachment.pl
Error in RBloomberg
4 messages · Sumanta Basak, I.Szentirmai, Thomas Lumley +1 more
Dear All, I'm using a glm to analyse the effect of year on a variable with poisson distribution. Using the summary() I can test whether years are different from each other, but how can I test whether year has an overall effect on my respons variable? My guess is that by a Wald-test, but I don't know how exactly. Could someone give me advice? Many thanks, Istvan
On Wed, 22 Feb 2006, I.Szentirmai wrote:
Dear All, I'm using a glm to analyse the effect of year on a variable with poisson distribution. Using the summary() I can test whether years are different from each other, but how can I test whether year has an overall effect on my respons variable? My guess is that by a Wald-test, but I don't know how exactly.
You can get a likelihood ratio test by also fitting a model without year and using anova() to compare the two models. I think multiple people have written wald test functions. Mine is regTermTest in the "survey" package. -thomas
Dear Istvan, You might want to take a look at the Anova() and linear.hypothesis() functions in the car package. Regards, John
On Wed, 22 Feb 2006, I.Szentirmai wrote:
Dear All, I'm using a glm to analyse the effect of year on a variable with poisson distribution. Using the summary() I can test whether years are different from each other, but how can I test whether year has an overall effect on my respons variable? My guess is that by a Wald-test, but I don't know how exactly.