이요인 일반화부분점수 모형을 위한 공통-문항 기반 다차원 IRT 척도연계 방법

초록

The bi-factor generalized partial credit (BGPC) model, which is among the polytomous multidimensional item response theory (IRT) models, can be used to analyze the partially scored item response data of bi-factor latent structures. The purpose of this study is to present four scale-linking methods for the BGPC model under the common-item equating/linking design and investigate their characteristics and performances through computer simulations. The four linking methods presented are the least squares (LSQ), mean-least squares (MLS), item-category response function (IRF), and test response function (TRF) methods. The four linking methods estimate the dilation and translation coefficients in linear transformations, with the assumption of independence among all general and group-specific factors. Computer simulations that included three factors of (a) nonequivalence level between two examinee groups, (b) sample size, and (c) number of the common items to be used for linking were conducted to examinee the performances of the four methods. Overall, the four methods properly estimated the dilation coefficients. In estimation of the translation coefficients, however, the IRF, LSQ, and MLS methods functioned properly but the TRF method did not. In the recovery of both ability and item parameters, the IRF method performed best in mose cases, the LSQ or MLS method second best, and the TRF method worst, with few exceptional cases.

키워드