Reflection 14

Assessed Levels of Second Language

Speaking Proficiency: How Distinct?

By NORIKO IWASHITA, ANNIE BROWN, TIM MCNAMARA and SALLY O’HAGAN

This study investigated the nature of speaking proficiency at different level and found out that vocabulary and fluency were the factors that had the strongest impact on distinguishing between the proficiency levels. This consideration is particularly important in designing a high-stake test, such as TOEFL. First, it is important to understand what we mean by fluency and how a test can be useful in measuring this competency. Test usefulness includes 6 main qualities: reliability, construct validity, authenticity, interactiveness, impact, and practicality. Although there is a tension among these qualities, there is no total abandonment of any. Depending on the purpose of a test, some of these qualities may have higher degree than the others.

If we analyze the TOEFL speaking section, we can see that certain features of the test increase the degree of overall usefulness of the test. I completely agree that pronunciation, rhythm, stress, and intonation are important factors in speaking, which make our speech comprehensible and help us avoid miscommunications. So, the test focuses on it. The topics in the test require not only descriptions but also have test-takers justify an opinion or compare and contrast. If someone is good at describing, it does not mean that he or she is also good at comparing things or justifying an opinion. So, the test measures different types of talk to receive more accurate information about the candidates’ oral proficiency. Besides, the test limits the possibility of memorization of any topic for taking the test, thus increasing its construct validity. This yields the positive washback of the test on teachers and students in the sense that teachers should not have their students memorize the topics; if the student does not actually like coking, he does not have to say he likes it.

Although there is certain control over what happens in the interaction with the test tasks, there is little power over the examinee, while in other tests the interlocutor initiates interaction and asks questions, and the examinee is to comply and answer. The quality of authenticity is also enhanced as the test contains both independent and integrated tasks. This means that there is relatively high degree of authenticity and reliability in the speaking tasks, unless the topics are memorized. I think one concern relates to scorer reliability and inter-rater reliability, an issue which was discussed in our class.

This discussion leads to the pedagogic prescription which could provide the best investment for the development of speaking proficiency. It is not enough to expose our learners to the conventions of actual encodings, because speaking is not something that learners can be rehearsed to. The language that people actually produce as observable behavior presupposes a vast knowledge of language as unexploited potential. They draw on this knowledge pragmatically as a complement to a context. Thus, if we want our learners to become competent in speaking, we should develop both their memory-based knowledge and rule-based knowledge. Drawing on these two types knowledge, which interrelate and interact with each other, learners will be able to exploit the virtual language and adjust to the conventions of actual encodings as the context requires. This implies that we should focus our teaching on the language as a resource for making meaning, which includes both what is virtual in the language and what is actual in its encodings.