Analisis Regresi Multilevel dalam Menentukan Variabel Determinan Nilai Ujian Akhir Nasional Siswa

Abstract

Hierarchical data are data where objects are clustered in their groups and for each of groups the variable(s) are set at different levels. It is common to analyze hierarchical data without examining individual’s data membership which affects the accuracy of analysis’ results.  Multilevel regression analysis is a method that can be chosen to overcomes issues regarding hierarchical data. This essay is aimed to apply multilevel regression analysis to evaluate score of national examination data from elementary school students at District of South Kuta, Bali.  These data were structured so that students are position as first level and are nested within their classes as the second level.  Furthermore, each of classes is nested within its respective elementary school as third level data. The application of three-level regression for these data set showed student’s score for national final test was significantly affected by teacher educational level of respective class.