Topic: Prediction Intervals for New Outcomes

Point Predictions and the Unobserved Error

Pointwise Versus Joint Coverage for Several Predictions

Mean-Model Misspecification and Interval Coverage

Leverage and the Width of a Prediction Interval

How Predictor Location and Design Density Shape Width

Gaussian Errors and Exact Prediction-Interval Coverage

Effects of Sample Size and Residual Noise on Interval Width

Deriving the Prediction-Error Variance

Degrees of Freedom and t-Based Calibration

Defining the Future Outcome a Prediction Interval Covers