Reader response to the Taiwanese Joint College Entrance Examination: English reading section.

摘要

This study complements conventional psychometric methods used in the process of test construction and evaluation. Such methods provide useful information with respect to item difficulty and item discrimination, but do not give information that explains why a particular item may be viewed as too difficult or not properly discriminative. Therefore, we need to ask test takers to explain their responses to test items, thereby gaining more direct insights as to why certain items are found to be problematic.;Three areas were investigated in this study: test makers, test material, and reader responses. First, two test makers were interviewed in order to understand the construction processes of the Taiwanese Joint College Entrance Examination (JCEE) English subtest. Second, comprehensive textual analyses were conducted with a corpus of 10 English reading passages and 19 tasks taken from the 1987-1993 JCEE, which were judged to be problematic by psychometric standards, serving as the basis for individual interviews and group probes in the final phase of the study. A total of 763 Taiwanese university freshmen, senior high graduates, and last-year senior high students responded to the test material in the corpus and/or participated in the interviews and probes.;The findings indicate that much of the test takers' confusion in response to the given test material resulted from ambiguities or errors found in the test material itself, which in turn were largely caused by the disparity in perspectives between test makers and test takers. The rich information obtained through interviewing test takers could enable test makers to better understand how test takers in general perceive and respond to test items.;The researcher proposes that for test improvement, reader response methods, such as interviews, and performance and knowledge probes, be used as a supplement to the psychometric approach. With the use of such methods, test makers could improve their test construction and evaluation, especially if they were to use these methods in the early stages of test development. Such methods would also enable test takers to play a more active role in developing the test material that can have such a great impact on their future lives.