Hi Dustin,
It seems that the estimates() function returns the following error when the Gamma family is specified in glm()
g <- glm(minutes.fighting ~ willpower, data = avengers, family = Gamma(link="log"))
estimates(g)
Error: object 'OR' not found
It does appear to work fine for Gamma mixed models when gmler() is used.
It seems to also work fine for both glm() and glmer() when family=binomial.
However, the same error appears when the poisson family is specified. I have not checked other amilies.
g <- glm(injuries ~ willpower * minutes.fighting, data = avengers, family = poisson)
estimates(g)
Error: object 'OR' not found
I suspect the estimates function is expecting an odds ratio when glm is detected, but does not get one when the family is binomial / poission, etc.
Cheers,
Tony
Hi Dustin,
It seems that the estimates() function returns the following error when the Gamma family is specified in glm()
g <- glm(minutes.fighting ~ willpower, data = avengers, family = Gamma(link="log"))
estimates(g)
Error: object 'OR' not found
It does appear to work fine for Gamma mixed models when gmler() is used.
It seems to also work fine for both glm() and glmer() when family=binomial.
However, the same error appears when the poisson family is specified. I have not checked other amilies.
g <- glm(injuries ~ willpower * minutes.fighting, data = avengers, family = poisson)
estimates(g)
Error: object 'OR' not found
I suspect the estimates function is expecting an odds ratio when glm is detected, but does not get one when the family is binomial / poission, etc.
Cheers,
Tony