Two Versions of the TOEIC(R) Test: A Comparative Study Based on the Classical Test Theory

ITO Akihiro, SHIMATANI Hiroshi, NORIZUKI Ken, KINOSHITA Masayoshi

doi:10.20622/jltaj.12.0_26

This paper presents a preliminary portion of the reports on our research project. In May 2006, the TOEIC(R) was revised in order to make it a more suitable test battery to measure English communicative language ability. The purpose of the project was to compare the older version of the TOEIC(R) test and the new one from the perspective of established testing theories. University students learning English as a foreign language (N=136) in Japanese universities made up the participants in this study. In November and December 2008, the two versions of the TOEIC(R) were administered to the participants of the study with an interval of one to two weeks. Based on the reliability estimation method, the two-way ANOVA (TOEIC(R) Version (2) × Subtest (2)), and item analyses of the participants' test scores, the following research findings were obtained. (1) Both versions of the TOEIC(R) test indicated high reliability coefficients (old: α=0.928; new: α=0.933); (2) the participants performed better in the new TOEIC(R) than the old one; (3) The participants performed identically in both TOEIC(R) Reading Comprehension sections, but performed better in the new Listening Comprehension section than the old one; (4) the new TOEIC(R) Listening Comprehension section contained more satisfactory test items than the old one in terms of item difficulty; and (5) the old TOEIC(R) Listening Comprehension section contained more satisfactory items in terms of actual equivalent number of options. The directions for further research on TOEIC(R) as a placement test in the language classroom setting are also discussed.

Two Versions of the TOEIC(R) Test: A Comparative Study Based on the Classical Test Theory

Bibliographic Information

Search this article

Abstract

Journal

Details 詳細情報について

Export

Report a problem