The calibration problem is that of estimating x at which (x) = 0 . For the linear logit model, argue that a confidence interval

The calibration problem is that of estimating x at which π(x) = π₀. For the linear logit model, argue that a confidence interval is the set of x values for which |α̂ + β̂x – logit(π₀)|/[var(α̂) + x² var(β̂) + 2x cov(α̂, β̂)]^1/2 < z_α/2.