count <- c(53,414,11,37,0,16,4,139)
V <- factor(rep(c("w","b"),c(4,4)))
D <- factor(rep(c("w","w","b","b"),2))
P <- factor(rep(c("y","n"),4))
P <- relevel(P,"n")
death.DV.DP.PV <- glm(count ~ D*V + D*P + P*V, family=poisson())
summary(death.DV.DP.PV)
fits <- fitted(death.DV.DP.PV)
round(cbind(count,fits),2)
# Fitting the simpler model for part (c)
death.DV.PV <- glm(count ~ D*V + P*V, family=poisson())
summary(death.DV.PV)
resids <- residuals(death.DV.PV,type="pearson")
h <- lm.influence(death.DV.PV)$hat
adjresids <- resids/sqrt(1-h)
round(cbind(count,fits,adjresids),2)