摘要: | This study aimed to evaluate the reading comprehension items of the Scholastic Ability English Test (SAET) and the Department Required English Test (DRET) from 2004 to 2008. Specifically, the study intended to answer the following three research questions: (1) What reading skills were measured on the SAET and the DRET reading comprehension sections and what was the percentage of the items for each of these skills identified? (2) How did the examinees in general perform on reading comprehension items measuring each of the reading skills on the SAET and the DRET? (3) For both tests across the five years, which reading skill identified could consistently best discriminate between the high scorers and the low scorers? For the purpose of answering the research questions, Nuttall’s (2000) categorizations of reading skills and question types were mainly used as the coding scheme. Two experts in the field of English were invited as raters to classify each of the 134 reading comprehension items into one of the 11 reading skills. The results showed that six reading skills were identified on the SAET from 2004 to 2008, including (1) Interpreting (39.24%), (2) Comprehending literal meaning (25.32%), (3) Reorganizing (18.99%), (4) Recognizing implications and inferences (7.59%), (5) Recognizing functional value (6.33%), and (6) Recognizing and interpreting cohesive devices (2.53%). As for the DRET, the same six reading skills were also identified along with one more sub-skill, Recognizing style and tone. The respective percentages of the seven reading sub-skills identified on the DRET were: Interpreting (40%), Recognizing implications and inferences (18.18%), Reorganizing (16.36%), Comprehending literal meaning (12.73%), Recognizing and interpreting cohesive devices (5.45%), Recognizing functional value (3.64%), and Recognizing style and tone (3.64%). The SAET takers performed best on the Comprehending literal meaning items, but worst on the Recognizing functional value items, whereas the DRET takers performed best on the Re ognizing functional value items, but worst on the Recognizing style and tone items. Furthermore, the examinees generally performed better on the SAET than those on the DRET, in terms of the mean passing rate for each of the reading skills identified. Finally, none of the reading skills could consistently best discriminate the high scorers from the low scorers for both tests across the five years. |