A Comparison Of The Test-taking Processes: Long Conversation Versus Short Conversation

Listening comprehension lies at the heart of language learning, but it is the least understood and least researched skill. With the deepening of the CET reform, Long Conversation (LC) item-format in listening comprehension component is adopted to supplement the previous test type --- Short Conversation (SC). The present study is conducted to compare the test-taking processes of LC and SC through analyzing testees'retrospective verbal reports.The study adopts Buck's (2001) framework of listening constructs, which classifies the componential competence of listening ability into language competence and strategic competence. Language competence consists of grammatical knowledge, discourse knowledge, pragmatical knowledge and sociolinguistic knowledge; strategic competence consists of cognitive strategy and metacognitive strategy. By investigation of the testees'verbal reports, the study focuses on the following three research questions:1) What kinds of knowledge of language competence are applied in test-taking processes of SC and LC?2) What individual strategies of strategic competence are applied in test-taking processes of SC and LC?3) What construct-irrelevant strategies are employed in test-taking processes of SC and LC?The instrument of the study comprises LC and SC parts excerpted from listening comprehension in CET-4 administered in June, 2008. Four second-year non-English majors from Shanghai Dianji University took part in the empirical study. Immediate retrospection task was conducted after the subjects had completed the test, and then a well-organized interview followed. The verbal reports were transcribed, coded and analyzed. The following three major findings are presented based on the interpretation of the protocols.Firstly, grammatical knowledge and discourse knowledge of language competence defined in the framework of listening constructs (Buck, 2001) are applied in both LC and SC, but pragmatic knowledge and sociolinguistic knowledge are not employed in SC or LC. The performance of SC relies more on grammatical knowledge while SC more on discourse knowledge.Secondly, test-takers use inferencing strategy, elaboration strategy and self-monitoring strategy significantly in the test-taking process in SC. In the process of LC, they employed the strategies such as note-taking, advanced organization, and selective attention, as well as inferencing, elaboration and self-monitoring.Thirdly, some strategies which are irrelevant to listening constructs are identified. Superficial matching is identified only in LC. However, it is not identified in SC.The comparison between the competences involved in the test-taking processes of SC and LC leads to the conclusion that test-takers apply wider varieties of listening comprehension subskills in the test-taking process of LC than that of SC. Thus, it can be argued that LC has better construct validity and will result in positive washback to the teaching of listening comprehension.