Earticle

초록

영어

Extending the conventional item response theory(IRT) as a measurement model to explanatory IRT(EIRT) based on generalized (non)linear mixed modeling has opened a new possibility to incorporate many item and person properties in its modeling process so that their effects can be considered simultaneously. The purpose of the current study was to apply EIRT to a complex data set that consisted of a reading comprehension test, a student survey of English study, and information about the item categories of the reading test. Sequential EIRT modeling of the data set showed that some of the item and person properties consistently had significant effects on the item difficulty and on the probability of correct answers. The modeling process also revealed some statistical or computational challenges researchers might encounter when they try to apply EIRT to complex language test data.

목차

Abstract
 1. Introduction
 2. Previous Studies Using EIRT
 3. The Data Structure of the Study
 4. EIRT Models
  4.1. Model Presentation
  4.2. Model Estimation
  4.3. Model Fit Evaluation
 5. Results
 6. Discussion and Implications
 References

키워드

기술적 문항반응이론 설명적 문항반응이론 일반화 선형모형 수험자 특성 문항특성 고정효과 랜덤효과 시험 타당도 descriptive IRT explanatory IRT generalized linear mixed models person properties item properties fixed effects random effects test validation

저자

• Sae il Choi [ Chonnam National University ]

참고문헌

발행기관

간행물