Citation

Brand MR. Reliab. Eng. 1980; 1(2): 115-125.

Copyright

(Copyright © 1980, Elsevier Publishing)

DOI

10.1016/0143-8174(80)90003-7

PMID

unavailable

Abstract

In attempting to establish reliability when only a few test items are available, three Bayesian non-parametric formulae have often been applied. After deriving all three, the 'equivalent sample size' formula, which relates to extended testing, is shown to be based on a false premise and to be incorrect. The other two formulae are contrasted to their classical counterparts and it is argued, mainly for reasons of consistency and common sense, that the classical formulae should be employed.