Dear R users: I am using the glm function in order to analysis data on fertility of rabbit females, the response variable are either 1, the female became pregnant, or 0, she does not. I am using the probit link function. Actually I have several measurement per female, thus it would be nice if I can include a random permanent effect of female in the model. I order to take into account for the variance covariance estructure of the data. I supply the calculation in this way: fert_probit<-glm(fer ~ gae + ctipo -1, family = binomial(link="probit"), data = FERTILIDAD) Thus my question is: Do exit same option in the glm() function to allow for random effects?, similar to the random option in lme() Thanks in advance Juan Pablo.
On Ti, 2005-12-20, 13:09, Juan Pablo S??nchez skrev:> [...] > Thus my question is: Do exit same option in the glm() function to allow > for random effects?, similar to the random option in lme()No, you can't fit GLMMs with the `glm' function. Instead, take a look at e.g. the `glmmML', `lmer' and `glmmPQL' functions in the glmmML, Matrix and MASS packages respectively. HTH, Henric