Academic streaming for secondary schools in Malaysia using a twostep clustering technique / Ely Salwana Mat Surin

• Main Author: Ely Salwana, Mat Surin
• Format: Thesis
• Published: 2015
• Subjects: L Education (General); QA75 Electronic computers. Computer science
• Online Access: http://studentsrepo.um.edu.my/6303/1/ely.pdf; http://studentsrepo.um.edu.my/6303/

Academic streaming, for students, is used to identify a learning pathway based on academic performance. This is important in assisting them to reach decisions related to their future careers; by distinguishing their academic strengths, weaknesses, interests, and abilities. There are two types of main academic streaming for secondary schools in Malaysia; the science stream and the arts stream. It is crucial to assign students to the correct stream, to avoid a distribution imbalance of students in both streams; as this will affect the needs potential for human capital in Malaysia. To achieve this, in 2011, the Ministry of Education (MOE) set human capital needs ratios of 40% for the art stream and 60% for the science stream. Based on the literature review, problems in determining streaming include: the academic performance of students in schools does not link substantially to the academic streaming process and the academic streaming cannot be determined systematically, due to a lack of understanding of the important factors that influence the streaming process; hence it is therefore difficult to identify the real potential of individual students. Despite several clustering techniques currently used in the educational environment, there is still a lack of appropriate techniques to be used with educational data which relate to academic streaming and academic performance. Based on these problems, this study aims to i) investigate current practices applied in academic streaming for secondary school in Malaysia and the student’s academic performance system; ii) identify factors contributing to the academic streaming process and propose an academic streaming framework based on identified factors; and iii) identify and apply a suitable clustering model to distribute students based on their academic performance. This study uses a mixed method research design; whereby the activities of data collection are performed using quantitative and qualitative methods for two different levels of respondents, namely teachers and secondary school students. iv Data was obtained from interview sessions with 17 teachers, surveys, and documents analysis for 465 secondary school students. The findings of this research have highlighted issues in the academic streaming process; mainly on the poor linkage between academic performance and the school’s streaming process since both of these processes are still connected manually. This research has also managed to identify important factors that relate to the academic streaming and clustering techniques that consists of nine categories which are education institution, peers, family, historical trends, workplace, globalization, employment market, community groups and socio economic status. Based on these findings, in order to stream students systematically, three frequently used technique for clustering which are TwoStep, K-means and Kohonen are analyzed and compared with each other using IBM SPSS Modeler. Based on the analysis, a TwoStep clustering technique is proposed to allocate students to appropriate streams based on their academic performance data. Through this study, current practices applied in academic streaming can be understood more clearly and the linkage between academic streaming and academic performance is identified. Further, a conceptual academic streaming framework for secondary schools is proposed to help teachers and students to understand all relevant streaming determination factors, and provide balance streaming for secondary schools in Malaysia based on academic performance using a TwoStep clustering technique.