## ----echo=FALSE,eval=FALSE----------------------------------------------------
# options(width=60)

## ----results='hide',eval=FALSE------------------------------------------------
# library(catdata)
# data(unemployment)
# attach(unemployment)

## ----eval=FALSE---------------------------------------------------------------
# durbin <- as.factor(durbin)
# table.durbin <- ftable(subset(unemployment, select=c("age", "durbin")),
# col.vars="durbin")
# rels<-table.durbin[,1]/rowSums(table.durbin)
# age.new <- min(age):max(age)
# 
# model1 <- glm(table.durbin ~ age.new, family=binomial)
# summary(model1)

## ----eval=FALSE---------------------------------------------------------------
# plot(age.new, model1$fitted.values, xlab="Age", ylab="Observed/Fitted values",
# type="l", ylim=c(0,1))
# points(age.new,table.durbin[,1]/rowSums(table.durbin))

## ----eval=FALSE---------------------------------------------------------------
# plot(model1$fitted.values,sqrt(rowSums(table.durbin))*rstandard(model1),
# xlab="Predicted values", ylab="Residuals")

## ----eval=FALSE---------------------------------------------------------------
# qqnorm(sqrt(rowSums(table.durbin))*rstandard(model1), main="",
#        ylab="Standardized deviance residuals")
# qqline(sqrt(rowSums(table.durbin))*rstandard(model1), lwd=2.5,
#        lty="dashed", col="red")