RESEARCH SETTING AND PARTICIPANTS

66 research question dealing with learning achievement, a set of measurement data in the form of test scores were obtained through a series of pre-test and progress- tests. These test scores are quantitative in nature. Meanwhile, to answer the research question that dealt with students learning process, pre- and post- program questionnaires were used to obtain quantitative data. In addition, qualitative data were also collected using in- depth interview, students’ reflective journals, and researcher’s observation field notes.

a. Tests

There were two types of tests administered throughout the research: pre- test and progress tests. The pre-test was conducted before the course program began, aiming to assess students’ initial speaking performance, as well as to confirm that students ’ proficiency ranged. Meanwhile, progress tests were administered at the end of each action research cycle, aiming to assess students’ speaking skill development. Thus, the tests were to answer the second research question regarding students’ achievement subsequent to their involvement in the course program. The overall design of the test scaling was adopted from the CEFR Common European Framework of Reference as Luoma 2004, p. 71 recommends for several reasons. The first reason is that it contains a range of ‘illustrative descriptors’ of language ability, which are useful for the examiners to analyze learner at different levels. Secondly, the descriptors of each level provided in CEFR are written for general purposes, so the functions and language use contexts are modifiable to suit other test purposes. There are five scoring criteria, PLAGIAT MERUPAKAN TINDAKAN TIDAK TERPUJI 67 namely: range, accuracy, fluency, and interaction or coherence. It is suggested that the descriptors of interaction and coherence are used either one depending on the task given. As Luoma 2004 explains, the interaction scale is useful for rating interactive skills, which are performed through interactions between two or more speakers. On the other hands, the coherence scale rates the performance of a single speaker doing long turns. As the students were measured in both types of performances in each test, both interaction and coherence scoring criteria were used in this research instrument. In CEFR scales, there are six levels at which the test participants are to be placed. Those levels are Basic A1 and A2, Independent B1 and B2, and Proficient C1 and C2. However, for the purpose of speaking assessments in this research, the levels were converted into numbers that function as scores ranging at 50 for level A1, 60 for level A2, 70 for level B1, 80 for level B2, 90 for level C1, and 100 for level C2. The conversion into scores enabled the researcher to see the average score of the learners in performing the overall speaking skills presented in the assessment units. In order to maintain the consistency and uniformity of the test scores between two examiners, an analytic descriptors of spoken language use Council of Europe, 2001, p.28-29 was adopted to guide the process of giving marks. For the pre-test and progress tests, the interaction pattern was adapted from Cambridge Key English Speaking Test CEFR Level A2 Council of Europe, 2015. In this test format, there are two students and two examiners: one examiner acts as the interlocutor who manages the interaction by asking questions and set