probit
when the DV is categorical (binary), you need a probit model
myprobit<- glm(Alpha ~ DominanceZ + Sex, family=binomial(link="probit"), data=fWHR, na.action=na.pass)
mylogit<- glm(Alpha ~ DominanceZ + Sex, family=binomial(link="logit"), data=fWHR, na.action=na.pass)
summary(myprobit)
anova(mylogit , test="Chisq")
anova(myprobit, test="Chisq")
confint(myprobit)
aod::wald.test(b=coef(myprobit), Sigma=vcov(myprobit), Terms=2:2)
excellent example here
http://nlp.stanford.edu/~manning/courses/ling289/logistic.pdf
page revision: 2, last edited: 10 Jun 2012 13:44