267 / 2018-04-16 08:46:52

Alignment of the Standards-based Tests in Asia: TEPS and GEPT Comparison

CEFR,GEPT,TEPS,standards,test constructs,classification consistency analyses,content analysis,factor analyses,correlation analyses

全体主题 > Standards in validating language assessments

Language testing organizations in a number of Asian countries have developed English proficiency tests that are tailored to the needs of local learners. These tests are designed to be fit for purpose in their own particular local contexts. To provide further information for the interpretation of test scores they produce, some of these locally produced tests have been linked to the Common European Framework of Reference for Languages: Learning, Teaching, Assessment (CEFR) or other external benchmarks, such as the TOEFL or IELTS. In this study, the Test of English Proficiency developed by Seoul National University (TEPS), developed for EFL learners in South Korea, was related to the General English Proficiency Test (GEPT), developed based on the English language curriculum in Taiwan’s school system, as part of its validation effort for the revised version of TEPS. The GEPT is selected as an external referent because it is a highly researched test and is systematically linked to the CEFR levels.
The comparison study aimed to explore the relationships between similar skill sections of the TEPS and GEPT by answering the following research questions:
1. To what extent does the Revised TEPS meet the psychometric quality standards when it is evaluated with Taiwanese test takers?
2. To what extent do the section and total scores of the Revised TEPS correlate with those of the GEPT?
3. To what extent do the assessment constructs of the Revised TEPS relate to those of the GEPT at the section and total test levels?
4. To what extent do the scores and the proficiency levels of the Revised TEPS match those of the GEPT?
A total of 134 learners who had taken the GEPT within the previous six months took the Revised TEPS in November 2017. To investigate the relationship between the two tests in terms of standards, test constructs, and test results, classification consistency analyses, content analysis, factor analyses, and correlation analyses were performed. This paper will report the research framework and preliminary findings. It is hoped that the study will provide a small-scale demonstration of procedures to investigate test comparability both quantitatively and qualitatively to enhance test usefulness and to result in more beneficial effects on the educational and societal contexts within each region.