Analisis Skor Literasi Membaca Siswa Indonesia Menggunakan Linier Mixed Models

Abstract

The linear mixed models is a development of the linear model which includes both fixed and random effects in the model. Random effect in the model is used to model complex data that has a grouping structure. The grouping structure can occur because the same observations are measured repeatedly or each observation is measured only once but these observations have some form of group structure. Students who participate in the Program for International Student Assessment (PISA) are nested in several schools, so the PISA data structure is quite complex and requires a more in-depth analysis. Quantitative studies on PISA, especially in reading literacy, are still rarely done. The purpose of this study is to determine what factors effect the Indonesian student’s PISA reading literacy scores using a linear mixed model approach with school being used as a random effect in the model. The findings of the study are that the factors that affects Indonesian student’s PISA reading literacy scores are the class being taken, gender, mother's highest education, facilities at home, school entry age, student discipline and failed a grade. The result of the estimation of random effect variance which is not equal to zero indicates that there is a random effect from the student’s school on PISA reading literacy scores. Based on model diagnostics and parameter testing, it was concluded that the model obtained is fitted in modeling Indonesian student’s PISA reading literacy scores.