A Content-related Validation Study Of TEM-4Listening Comprehension

Since1990, the validity research on TEM-4has been a continuous heated topic, however,most validity studies of TEM are about reading part, many of which are based on validity ofclassification view. What’s more, the numbers of items, time allotment and answering time etc.have been changed since the testing syllabus reformed in2005, so it is necessary to collectand analyze the content-related evidences of listening part of TEM-4to validate the listeningtasks from2005to2013.In order to achieve the research goal, we’ve read lots of data about communicativelanguage test including introductions to related theories and empirical research achievements.Based on Bachman&Palmer’s framework of task characteristics with great flexibility andadaptability, and referring to Weir’s interlocutor theory, the author puts forward a revisedframework including the input, the expected response, and the relationship between input andexpected response for validating listening comprehension. On the ground of this revisedframework with the help of the teaching and testing syllabuses and combing with thecharacteristics of the TEM listening comprehension part, the study has showed us to whatextent the TEM-4listening comprehension tasks have being matched with these twosyllabuses, based on what it has demonstrated the content-related validity of TEM-4listeningpart.In terms of the three characteristics in the revised framework, we examined9years’TEM-4papers, including102topics, and270question items. Results show that TEM-4listening comprehension (2005-2013) has a high content-related validity. It can be manifestedin the following three aspects:1) About the input. Most characteristics of listeningcomprehension parts of TEM-4after the reform meet the requirements in the relevantteaching and testing syllabuses in the matter of the length, speed and accent of the audiomaterials, the proportion of the new words, and the range of topics.2) About the response.Question types mainly are items answered from literal comprehension or reorganizedinformation, taking into account the reference and judgment questions, examining the full range of listening expected skills.3) About the relationship between input and response. Thereare absolute proportions of narrow scope and direct relationship tasks, which is consistentwith the distribution of the test taker’s ability levels and has few content-irrelated answers.Meanwhile there are some other discoveries: A few of audio materials’ length are longer orshorter than the requirements; Only when test takers take the advantage of the questionsprovided before listening in TEM-4, are they able to answer the questions successfully; Thetest takers are required to improve the listening ability as for the acceleration of speech rate;And taking authenticity into account, it is supposed that the accent of the audio material bemore classical like in the target language domain as long as the pronunciation and intonationis clearly expressed.This study also has some limitations and the future research space is very large. Becausevalidation is an endless process, only when more researchers do a lot of research fromdifferent ways, can validation work be improved and upgraded theoretically and practically.