Ability Estimation with Polytomous Items in Computerized Multistage Tests

Authors : Hasibe YAHSİ SARI, Hülya KELECİOĞLU

Pages : 171-184

Doi:10.21031/epod.1056079

View : 107 | Download : 65

Publication Date : 2023-09-30

Article Type : Research Paper

Abstract :The aim of the study is to examine how the ability estimations of individuals change under different conditions in tests consisting of polytomous items in an computerized multistage test environment. The research is a simulation study. In the study, 108 insert ignore into journalissuearticles values(3x3x6x2=108); conditions were examined consisting of three categories insert ignore into journalissuearticles values(3, 4 and 5);, three test lengths insert ignore into journalissuearticles values(10, 20 and 30);, six panel designs insert ignore into journalissuearticles values(1-2, 1-2-2, 1-3, 1-3-3, 1-4 and 1-4-4); and two routing methods insert ignore into journalissuearticles values(Maximum Fisher Information insert ignore into journalissuearticles values(MFI); and Random);. Simulations and analyses were carried out in the mstR package in R program, with a pool of 200 items, 1000 people and 100 replications insert ignore into journalissuearticles values(e.g., iterations);. As the outcomes of the research, mean absolute bias, RMSE and correlation values were calculated. It was found that as the number of categories and test length increase, the mean absolute bias and RMSE values decrease, while the correlation values increase. In terms of routing methods, although MFI and random methods have similar tendencies, MFI gives better results. There is a similarity between the panel designs in terms of results.
Keywords : Computerized multistage tests, polytomous items, routing methods

ORIGINAL ARTICLE URL