A Study Of The Relationship Between The Stem Types And Students' Performance In TEM-8 Reading Comprehension

This study is intended to investigate the relationship between stem types and the students'reading performance. It has two objectives: one is to find out the distribution of stem types in RCT in TEM-8 from 2000 to 2005 and the other is to find out the relationship between stem types in RC and the students'reading performance.163 English majors from Yangzhou University participated in this study. They were selected as subjects because they were linguistically competent and had not taken TEM-8 yet. The instrument employed in this study was a reading test including 6 test papers. Data analysis involved four steps: scoring, classification of stem types, calculations and comparison. Three major findings were yielded from this study:One is that among the three stem types, fact/detail questions have the highest correlation (0.753). This indicates that fact/detail questions determine to a great extent whether the score for reading is high or low. The higher the score of this stem type is the higher the test performance is. This contradicts the findings of Li (1997) and Lu (2002).Another is that though main idea/topic question is least occurred, it still meets the requirement of the National Syllabus for English Majors (1999) and objectively reflects the students'reading ability in an all-round way for in 2000-2004 TEM-8 there is skimming and scanning, through which the test-takers'global understanding of main idea is tested. But the total amount of the reading has been reduced since the skimming and scanning was deleted by the 2004 Test Syllabus for RC of TEM-8. The design of stem types is still to some extent in accordance with the syllabus.Lastly, the accuracy rate of main idea/topic questions is the highest, which may indicate that the students'ability to summarize has been well developed after three years study and main idea/topic questions of TEM-8 are somewhat easy.Although great efforts have been made, the present study has its own limitations and calls for improvement in future research.Firstly, this study only covered original reading passages and questions of TEM-8 from 2000 to 2005, the cases before 2000 and after 2005 haven't been studied.Secondly, the test would be more reliable if all the six test papers were done by the same subjects.Thirdly, since the researcher herself chose the content of the tests and categorized the three types of questions, the results of the study could be subjective. These may to some extent affect the results of the study. If the stem type is categorized in another way, the results may be a bit different.