Singhing with confidence: visualising the performance of confidence procedures

Confidence intervals are an established means of portraying uncertainty about inferred parameter and can be generated through the use confidence distributions. For a distribution to valid, it must maintain frequentist coverage true parameter. This represented for precise by adherence cumulative unit uniform distribution, comparison that is referred here as Singh plot. article extends this imprecise procedures with bounds around describes how deviations convey information regarding characteristics designed inference prediction. An example plots highlight deficiencies in proposed methods given ProUCL Chebyshev upper limit estimator mean unknown distribution. quickly inform appropriate such these aid development novel avoid similar errors.