A Comparative Study Of Inference Tasks In Reading Comprehension Between CET-4and CET-6

Since its first administration in1987, College English Test (CET) has beenexerting tremendous impact on the English teaching and testing development in China.As a criterion-related and norm-referenced achievement test, CET battery consists oftwo bands (Band4and Band6), aiming to assess the English proficiency ofnon-English major undergraduates in China. According to the CET test syllabus, bothCET-4and CET-6attach great importance to assess students’ inference ability inreading comprehension so as to reflect the requirements of the teaching syllabus.However, the question of how the inference questions in reading comprehension inCET-4differ from those in CET-6has rarely been investigated. This present studyaims to address this question by comparing the inference questions in these two testsfrom the perspective of task characteristics and also the result of an empiricalexperiment conducted on106freshmen from East China University of Science andTechnology.This current study, by conducting both qualitative and quantitative analysis,intends to address the following three questions:(1) How do the task characteristics of the inference reading comprehension tasksin CET-4differ from those in CET-6?(2) Is there any difficulty level difference between inference readingcomprehension tasks in CET-4and CET-6? If there is, to what extent does thedifficulty level of such tasks in CET-4differ from that in CET-6?(3) How do the test takers think about the implicitness of the inference readingcomprehension tasks respectively in CET-4and in CET-6?The Test Method Facets (TMF) Framework proposed by Bachman (1990) is usedas a guideline and checklist for the establishment of a more specific revisedframework particularly used for analyzing the inference reading comprehension tasksfor this research. Altogether twenty official test papers, ten for CET-4and ten forCET-6, administrated from the year2007to2011are regarded as the resource ofqualitative analysis on test tasks. Inference reading comprehension tasks in these test papers are first picked out by an expert group and then analyzed from the perspectiveof vocabulary frequency, lexical density, syntactic complexity and sentence type, etc.It is found that key sentences, stems and options of inference reading comprehensiontasks in CET-4differ in all the above-mentioned aspects from those in CET-6.A test paper containing six reading passages and corresponding twenty inferencequestions is designed by the author and administrated on106freshmen from EastChina University of Science and Technology for the purpose of revealing differencesin difficulty level of inference reading comprehension tasks in CET-4and CET-6.Besides, an additional feedback question is designed so as to, with the collectedfeedback data, compare the test takers’ viewpoints on how implicit the inferencereading comprehension tasks are in CET-4and in CET-6. SPSS is used as the tool forthis quantitative analysis and the result shows that the inference readingcomprehension tasks in CET-6are more difficult than those in CET-4and also areregarded as having a higher degree of implicitness by the test takers.This research is meaningful in the sense that it is a comparative study conductedon a specific kind of CET test task that has been rarely explored. The study clearlypresents the task characteristics of inference tasks in reading comprehension sub-testin CET-4and CET-6as well as the difference of difficulty level and inference degree,hoping to provide the test designers with more information for the optimization of thetask.