Autore
De Martini, Daniele

Titolo
The emotionality of the p-Value, the rationality of the Post-Hoc Power: how to interpret the results of statistical tests.
Periodico
Università degli Studi del Piemonte Orientale 'A. Avogadro' : Facoltà di Economia - Dipartimento di Scienze Economiche e Metodi Quantitativi "SEMEQ" - Quaderni
Anno: 2004 - Volume: 04 - Fascicolo: 82 - Pagina iniziale: 1 - Pagina finale: 15

The stability of statistical test results is usually measured by the p-value. But, what is stability? Recently many authors have discussed the opportunity of using the information provided by the post-hoc power (PHP) to reinforce the result that the null hypothesis is true. Some others suggest using PHP independently from statistical significance. Both have in mind adding PHP to the results of a statistical test. But PHP has not yet received the right rationalization. The aim of this work is to supply an adequate and rational explanation about the PHP as an indicator of the "stability" of a statistical decision, i.e. a statistical test, and its relation with the other traditional indicator: the p-value. We argue that the concept of stability follows simply by observing that a statistical test is a Bernoulli random variable with mean equal to the power IT. Then, the stability is'to be considered in terms of non-variability, or, more precisely, of reproducibility of the desired outcome HI , which is measured by K, which can be estimated. The estimated power PHP is a rational indicator of the stability. Moreover, we focus on the distribution plug-in power estimator, which is a general and consistent statistical tool, and its relations with recent literature.




Esportazione dati in Refworks (solo per utenti abilitati)