Frequentist statistical inference without repeated sampling

Abstract Frequentist inference typically is described in terms of hypothetical repeated sampling but there are advantages to an interpretation that uses a single random sample. Contemporary examples given indicate probabilities for phenomena interpreted as classical probabilities, and this equally likely chance outcomes applied statistical using urn models. These used address Bayesian criticisms frequentist methods. Recent descriptions p -values, confidence intervals, power viewed through the lens probability based on sample from population.