The study developed from realisation that there is no information available about strategies or processes in the iBT speaking scoring rubrics, although ETS (Educational Testing Service) claims that the iBT speaking test is designed to measure strategic processes, which is one constructs of academic proficiency. Therefore, the study investigates which strategic processes are used to complete given speaking tasks. This would provide evidence to help in the evaluation of the validity claims proposed by the test designers. Six Korean participants, studying at English-medium universities, completed 2 independent tasks and 2 integrated tasks both in a test and in their class. Participants’ speech samples were collected during the performances and stimulated recall verbalisation was conducted after they had completed the tasks. Speech samples were coded into five categories: approach, compensation, cognitive, metacognitive strategies and feelings. Consequently, the study examined how strategies reported through stimulated recalls were present in actual speech. The findings showed that metacognitive strategies were used most frequently under both conditions. Fair-level speakers employed more strategies in the test, while good-level speakers used more strategies in the class. Moreover, integrated task types elicited more strategy use for both conditions. Speakers reported that they felt significantly more negative under test conditions than in the class. More importantly, two conditions shared 67.74% of the strategy types, and 84% of the strategy types used in the test were also used in the classroom, which may strengthen the validity of the iBT speaking test in terms of strategy use. Finally, evidences of strategy use were identified in actual speech, which can open the way to operationalised strategy use assessment in speaking test. However, the figure of evidenced strategy use was very low: 5.28% and 2.66% respectively in the test and in the class. It is recommended that future research be carried out with a large number of participants in order to generalise strategy use in speaking performance. Moreover, further studies might be conducted to examine the significance of observable strategic evidence in speech, to inform decisions to include strategies in the scoring rubrics.

abstract: English proficiency is one of the major factors affecting international students’ academic adjustment to American universities. Many international college students select to improve their academic English proficiency through an English intensive program. Collaborative learning is an educational approach to teaching and learning that allows students and teachers to engage in a common task and work together to search for knowledge and skills. This thesis study aims to develop, design, and iteratively refine strategies to help English intensive program teachers build collaborative learning and promote international students’ effective collaboration, so as to improve students' academic English proficiency. In this study, two different collaborative learning strategies were designed, implemented and iterated. Data was collected using qualitative methods and follow the principle of design-based research (DBR; Barab, 2014) The results of this study suggest that successful instructional strategies for collaborative learning should be designed in the following ways. First, gathering participants’ opinions and feedback at all phases of design and iteration; Second, linking the new strategies or activity to students’ grade should be the center of the design. Third, in DBR，researchers need to be patient to build good relationships with practitioners, which can provide a basis for continuing research. / Dissertation/Thesis / Masters Thesis Art 2020