American Review of Mathematics and Statistics

Hypothesis Testing and Statistical Confidence: An Overdue Observation on the Efficacy of a Hypothesis Test
David D. Marshall, Brandi N. Falley, Mark S. Hamner

Abstract
We introduce the novel argument that the general concept of statistical confidence applies both to an interval estimate of a parameter and to a hypothesis test. Measured degrees of statistical confidence are mathematical probabilities of accurate parameter identification established prior to drawing samples. Such probabilities serve as the foundation for a statistician’s expectation and conviction that a hypothesis test will correctly identify a true hypothesis, and more familiarly, that an interval estimate will properly identify a population parameter. The incidental and potentially misleading role of the P-value is discussed in the context of statistical confidence.

Full Text: PDF     DOI: 10.15640/arms.v3n1a3