Autore
De Martini, Daniele

Titolo
The emotionality of the p-Value, the rationality of the observed power: how to measure the riability of statistical tests results.
Periodico
Università degli Studi del Piemonte Orientale 'A. Avogadro' : Facoltà di Economia - Dipartimento di Scienze Economiche e Metodi Quantitativi "SEMEQ" - Quaderni
Anno: 2005 - Volume: 05 - Fascicolo: 82 - Pagina iniziale: 1 - Pagina finale: 19

The reliability of statistical tests results is usually measured by the p-value. But, what is reliability? Recently many authors have discussed the opportunity of using the information provided by post-hoc power evaluation to reinforce the result that the null hypothesis is true. Some others suggest using the observed power independently of statistical significance. Both have in mind adding power evaluation to the results of a statistical test. The aim of this paper is to supply an adequate explanation about the Observed Power as an indicator of the reliability of a statistical test result, and its relationship with the p-value. We argue that the concept of reliability follows simply by observing that a statistical test is a Bernoulli random variable, and that the test is constructed with the logic of being an "empirical proof" of the fact that the alternative hypothesis HI is true. Then, the reliability is to be considered in terms of reproducibility of the desired outcome HI, which coincides with the power of the test, and which can be estimated. The estimate of the power, that is the Observed Power, is thus a rational indicator-of the reliability of tests results.




Esportazione dati in Refworks (solo per utenti abilitati)