Ussama، Ratchawan نويسنده Prince of Songkla University, Thailand , , Sinwongsuwat، Kemtong نويسنده Prince of Songkla University ,
This study investigated the differences between two types of oral English proficiency tests (i.e., interview interaction and two-party peer interaction) in assessing learners’ English speaking performance. It attempted to determine (a) whether the two test types produce different results when used in assessing students’ oral performance with a traditional rubric, (b) whether they differ in terms of the interactional features produced by the students, and (c) how the differences in interactional features possibly contribute to the different score results and the conversation-assessing ability of each test task. The participants of the study were ten students in the M.3 English Program at Pimanpittayasan School, Satun. Two test tasks were used: (1) a 5-minute interview interactions, and (2) a 5-minute two-party peer interactions. The conversations obtained from both test tasks were videotaped, transcribed, and analyzed according to the principles of Conversation Analysis (CA). The results revealed that the total scores obtained from both tests using the traditional rubric were not significantly different (at an alpha level of 0.05). It was suggested that both interview and two-party peer interactions produced similar results when traditionally assessing the overall oral English performance of the learners. However, close analysis of talks elicited from the two test tasks revealed different interactional features some of which were related to learners’ problems in conversation, indicating that as far as conversational competencies are concerned, interview interaction may be of more limited value than peer interaction.