Nowe ujęcie problematyki mocy
testów statystycznych
Elżbieta Aranowska, Jolanta Rytel*
Wydział Psychologii Uniwersytetu Warszawskiego
NEW APPROACH TO THE PROBLEMS OF STATISTICAL TESTS POWER
Summary. This article presents a discussion of the statistical tests power concept based
the Neyman-Pearson approach to the null hypothesis testing. The notion of power is
presented from the classical point of view. Necessity to include two inference errors in
planning research is postulated and the examples of calculating the test power as well as
the sample size needed to secure the desired power are described. Moreover a revised
concept of power based on Kaiser's (1960) proposal of 3-alternative hypothesis testing is
discussed and suggestions are offered for assuming of Type I and Type II errors values
Autor: Ewelina Soszyńska
Ostatnia aktualizacja: 20.06.2007, godz. 15:11 - Ewelina Soszyńska