MULTILEVEL REGRESSION ANALYSIS ON AVERAGE SCORES SMA/MA IMPLEMENTING UNBK

Abstract

At the end of the third year, all senior high school students are required to take the National
Examination, which affects their graduation. In the implementation of the UN in 2017, Kemdikbud
implemented the Computer-Based National Exam which aims to improve the efficiency, quality,
reliability, integrity, and economy of the UN implementation. This study aims to analyze the factors
that contribute to the dependent variable, namely the average score of SMA/MA implementing
UNBK in the Natural Sciences department in 2017. The analytical method used is multilevel
regression, which allows to model level-1 and level-2 independent variables. Level-1 independent
variables include school factors, namely school type, school status, and number of participants. In
addition, the level-2 independent variable of area (city) was also considered. The analysis shows
that the factors that significantly affect the average score of UNBK are the type of school (SMA or
MA) and the number of participants/students participating in UNBK.