Validity of Multiple-Matching Tasks for reading comprehension tests

Validity of Multiple-Matching Tasks for reading comprehension tests

Yu,Chen, Hsuan

The purpose of this study aimed to investigate the validity of multiple-matching tasks by examining test-taking processes. Specifically, test-taking processes were analyzed by test-taking strategies. Hence, the researcher identified what test-taking strategies participants employed. Also, this study investigated whether high proficiency level learners employ more strategies than low proficiency learners do. Six Chinese-speaking English learners participated in the present study and were divided into high, intermediate and low proficiency groups. Participants were asked to think-aloud while performing one reading comprehension test. Findings indicated that participants employed five major test-taking strategies throughout the task. Moreover, locating the task in Messick’s (1989) theory, Kane’s (2006) and Bachman and Palmer’s (2010) validity frameworks, the researcher argued that the validity was violated in that (1) test takers could choose a correct answer without understanding the text; (2) test-taking processes did not correspond to the processes in natural settings and (3) test takers did not really show their comprehension because they were given main ideas to choose from. Therefore, cautions should be made for test users when they interpret test takers’ scores. In terms of the amount and types of strategy use cross different proficiency learners, results showed that except for one high-proficiency learner, other learners of different proficiency did not differ considerably from each other.

Conference or Workshop Item
PeerReviewed

2012-05

866