摘要

Adaptively administered tests with dichotomously scored items are already well described in the relevant literature and used in practice. The presented study analyses the possibilities of adaptive administration of test with polytomous items, which are commonly used in personality testing. Based on analysis of simulated adaptive administration of NEO PI-R, the limits and benefits of this approach are discussed. It was found that adaptive administration successfully and more effectively reconstructs the level of measured traits in comparison with full scale administration. On the other hand, significant problem consists in overexposure of several items with highest item discrimination power. Representative sample built for the purposes of Czech standardization of NEO PI-R was used (N = 2084).

  • 出版日期2011