\( \newcommand{\bm}[1]{\boldsymbol{\mathbf{#1}}} \DeclareMathOperator{\tr}{tr} \DeclareMathOperator{\var}{var} \DeclareMathOperator{\cov}{cov} \DeclareMathOperator{\corr}{corr} \newcommand{\indep}{\perp\!\!\!\perp} \newcommand{\nindep}{\perp\!\!\!\perp\!\!\!\!\!\!/\;\;} \)

3.4 shuttle data

shuttle contains the data in Table 1.3 of Davison (2003) on O-ring failures for the space shuttle.

To make the fitting code below work, we first remove the row names in shuttle

row.names(shuttle) <- NULL

To fit a binomial logistic regression model with covariate temperature:

fit <- glm(cbind(r, m-r) ~ temperature, data = shuttle, family = binomial)
anova(fit)
summary(fit)

Try fitting with and without both covariates. To assess model fit, try

plot.glm.diag(fit)

Do you find these diagnostics useful?

(Sections 10.1-10.4; Dalal et al., 1989)