confidence-interval

A confidence interval is a specific kind of statistical inference procedure — a competent speaker would call it a form of inference. It maps sample data to a population-level claim with quantified uncertainty, which is exactly what statistical inference does. Nearest-kind.INSTANCE_OF, not a leap (Law 11e).

Confidence intervals are estimation procedures within statistical inference. Remove statistical inference and confidence intervals cease to operate — they are not standalone constructs but inference tools.

which-came-first: confidence intervals (1930s, Neyman) predate prediction intervals (1930s, also Neyman but later in his work on tolerance regions). Prediction intervals were built on the confidence interval framework, extending it from parameter estimation to future observation. Confidence intervals existed first and fed into prediction intervals.