An Empirical Study Of Authenticity Of CET Listening Comprehension Tests

Based on the testing theory, this paper undertakes an empirical study mainly from the facets of input in CET listening comprehension tests. It sets its main aim to promote the authenticity of CET listening comprehension tests by using improved input texts, and then promote its validity. It mainly probes two questions: (1) college students have different reactions to the scripted texts and semi-scripted texts. Which one will be easier for them? (2) Listening to scripted texts, and listening to semi-scripted texts will yield different backwash effects. Which one will bring more effective backwash? Both qualitative and quantitative methods are used to collect and analyze the empirical data. Subjects in this study are randomly selected from Guangxi University. There are 72 junior non-English majors, and 30 sophomore English majors. The total number of subjects is 102. According to their listening ability, they are randomly divided into two parallel groups with similar listening proficiency. This experiment used modified listening tests designed by the author. All tapes used in the listening tests were recorded by two native speakers. The same material was recorded twice by the scripted way and semi-scripted way respectively. Control group students listened to the scripted version, while experimental students listened to the semi-scripted version.