earticle

영어

This study examined the validity of subscores for a reading comprehension section in a TOEIC practice test for which subscores are routinely provided for test takers. The study employed a bifactor approach, a constrained version of full-information item factor analysis, in its investigation of the subscore characteristics. The results showed a) all the subscores have lower reliability indices compared to the reliability of the total score and b) that the information of the subscores is trivial and redundant after accounting for their contribution to the general reading ability. These results agree with most of the previous studies on subscore quality and again demonstrate that language testing programs should be cautious about reporting subscores despite the pervasive demand for subscores in the testing industry.