Article appearing in Shiken 29.2 (Dec. 2025) pp. 27-40;
Article Doi: https://doi.org/10.37546/JALTSIG.TEVAL29.2-2
By Trevor A. Holster1 and Masaki Akase2
1Fukuoka Jogakuin University
2National Institute of Technology, Nagano College
Abstract
This study investigated the validity of TOEIC Bridge® as a placement test for use in a Japanese college of technology. Practice tests were administered to 185 students and analyzed using the dichotomous Rasch model. The data was found to sufficiently meet Rasch model requirements of unidimensionality, local item independence, and data-model fit to allow valid inferences about student English proficiency. The reliability coefficient of .90 was sufficient to separate students into low, mid, and high-proficiency groups and the range of item difficulty was acceptable for this purpose. However, there were very few items difficult enough to test the highest ability students in this group, so a test intended for intermediate level English speakers would be more suitable for making placement decisions within the highest group of students.
Keywords: TOEIC Bridge®, Rasch analysis, test validity, placement test