PENERAPAN HIERARCHICAL LINEAR MODELING UNTUK MENGANALISIS DATA MULTILEVEL

  • Dewi Wulandari Universitas PGRI Semarang
  • Ali Shodiqin Universitas PGRI Semarang
  • Aurora Nur Aini Universitas PGRI Semarang
Keywords: HLM, regression, multilevel data, OLS, GLS

Abstract

Multilevel data are data that are nested within the other data which are in the higher level. As an example is students are nested in the classes. The student is the level-1 variable and the class is the level-2 variable. Multilevel data are not restricted only level-2 but also more than it. As an example we have taken, school is the level-3 variable, region is the level-4 variable etc. Students in one class will be different from another class, classes in one school will be dierent from another school, etc. Because of this variation then we need Hierarchical Linear Modeling (HLM) to
analyze it. This method is a complex form of OLS (Ordinary Least Square) regression. In estimating the parameters we use GLS (Generalized Least Square). In this research, we use mathematics score of Nasima junior high school student Semarang. From the analysis result which are got by using software HLM student version we can conclude that there's no signicant variation within groups or classes, then it's enough using OLS regression to analyze the factors affecting mathematics score. Hypothesis of the reason of this is the amount of unit in level-2 variables are not enough, they are only 4 units. To prove this hypothesis, we need another research.

Published
2016-06-01
How to Cite
Wulandari, D., Shodiqin, A., & Aini, A. (2016). PENERAPAN HIERARCHICAL LINEAR MODELING UNTUK MENGANALISIS DATA MULTILEVEL. UJMC (Unisda Journal of Mathematics and Computer Science), 2(1), 16-21. https://doi.org/https://doi.org/10.52166/ujmc.v2i1.444