Alternatywna metoda szacowania zgodności wewnętrznej testu

Abstract

Commonly used methods Tor assessing the discriminatory capacity of test items and test intemal consistency refer mainly to the analysis of variance or the corrclation ratio which, in fact, is its particular example. In order to make them work properly, we need to challenge the fairly strong assumptions conceming, for example, the myth of normal distribution. Failurc to do so would lead to paradoxcs. Just following the author’s suggestion to use the analysis of set isomorphism and taxonomic distance between them would not rcąuire any challenging assumptions. The compulational proccdures applied are logically simple and easy to use.