Autore
De Martini, Daniele

Titolo
Calibrating sample size estimation.
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: 81 - Pagina iniziale: 1 - Pagina finale: 12

The risk function is defined and it suggests using power greater than 90%. The methodology of sample size estimation based on a pilot sample is considered. Consequently, the power of the planned experiment results a random variable. Pointwise estimates of noncentrality parameters or estimates based on bound of confidence intervals with a classical threshold usually provide biased power. Unbiased bounds are defined, based on the knowledge of the investigated distribution F. Then, a bootstrap calibrated estimator of the bound is proposed. A simulation shows that the sample size computed through the bootstrap estimated bound, performs better than those based on pointwise estimation of noncentrality parameters and on a conservative bound of the noncentrality parameter with classical threshold, in the Z-test and in the i-test.




Esportazione dati in Refworks (solo per utenti abilitati)