Türkiye’deki öğrencilerin fen okuryazarlığını etkileyen faktörler nelerdir? PISA 2015 verisine dayalı bir hiyerarşik doğrusal modelleme çalışması

The main purpose of this study is to investigate the student and school-level variables affecting Turkish students' science literacy using PISA 2015 data. In this way, we aim to build a hierarchical model of the variables predicting students' science literacy level. Particularly, when we consider the sharp decrease in Turkish students' success in PISA 2015, the implications of this study would be even stronger. Because of the nested nature of the data and a high intraclass correlation coefficient (ICC) value (0.52), we performed hierarchical linear modeling (HLM) analysis. As a result, we constructed a model including nine student-level and four school-level variables to predict students' science literacy scores. We classified the student-level variables into three categories as personal characteristics, variables associated with learning time, and variables associated with teaching-learning process. Similarly, we classified the school-level variables into two categories: school resources and learning environment in the school. While "weekly science learning time" is the most prominent variable at the student-level, "science specific resources", at the school-level, seems to be the most powerful predictor of students' success. One of the surprising findings in this study is that there is a significant negative correlation between "out-of-school study time" and science literacy scores.The main purpose of this study is to investigate the student and school-level variables affecting Turkish students’ science literacy using PISA 2015 data. In this way, we aim to build a hierarchical model of the variables predicting students’ science literacy level. Particularly, when we consider the sharp decrease in Turkish students’ success in PISA 2015, the implications of this study would be even stronger. Because of the nested nature of the data and a high intraclass correlation coefficient (ICC) value (0.52), we performed hierarchical linear modeling (HLM) analysis. As a result, we constructed a model including nine student-level and four school-level variables to predict students’ science literacy scores. We classified the student-level variables into three categories as personal characteristics, variables associated with learning time, and variables associated with teaching-learning process. Similarly, we classified the school-level variables into two categories: school resources and learning environment in the school. While “weekly science learning time” is the most prominent variable at the student-level, “science specific resources”, at the school-level, seems to be the most powerful predictor of students’ success. One of the surprising findings in this study is that there is a significant negative correlation between “out-of-school study time” and science literacy scores. Keywords: Science literacy, PISA 2015, hierarchical linear modelling (HLM), science educatio