Displaying 1 result from an estimated 1 matches for "quasi_glm".
2002 Jan 10
0
quasibinomial glm
...have a glm with family=binomial, link=logit but there is
over-dispersion. So, in order to take into account for this problem i
choose to do a glm with family=quasibinomial(). I'm not an expert on
this subject and i ask if someone could validate my approach (i'm not
sure for the tests) :
quasi_glm(myformula,quasibinomial(),start=mystart)
summary(quasi) # test t for coefficients
anova(quasi,test="F") # F test (quasi glm) instead of Chisq test
(classical glm)
Thanks in advance.
--
St?phane DRAY
---------------------------------------------------------------
Biom?trie et Biologie ?...