A vast majority of the studies exploring the associations between student and school related factors and standardized test scores were conducted in developed countries. On the other hand, research suggests that the generalization of the findings of those studies to developing countries often leads to incorrect conclusions. The purpose of this study was to explore the effects of selected student- and school-level factors on 15-year old Turkish students’ scientific literacy achievement. The Program for International Student Assessment (PISA) 2006 database was utilized to explore a) if there were differences among schools in terms of their mean scientific literacy scores, b) which student-level factors can explain the differences in students’ scores within a particular school, and c) which school-level factors can explain the scientific literacy differences between schools.
Hierarchical linear modeling (HLM) method was selected as the analytic method due to its usefulness in exploring relationships between a dependent variable and sets of layered independent variables. While the dependent variable of the study was represented by five plausible scores, independent variables consisted of 25 variables. Among those predictors, 15 were measured at student-level (level 1) and grouped in 4 clusters (background characteristics, teaching and learning factors, affective factors, and out-of school science related activities), and the remaining variables were measured at level 2 and grouped under two clusters (school resources and school context).
The results of the study indicated that more than half of the variation in students’ scientific literacy scores occurred among schools. While eight student-level variables—grade, economic, social and cultural status, general and personal value of science, responsibility for sustainable development, science self-efficacy, in-school time spent on science learning, and hands-on activities— explained about one-third of the variation at the student-level, three school-level predictors—school sector, mean economic, social and cultural status, and mean in-school time spent on learning— accounted for more than 70% of the variation at that level.
Several findings of the present study were different than what mainstream research suggested. Instrumental motivation and interest in science were found to be insignificant in predicting scientific literacy of Turkish students; while the students’ perceptions about the general value of science had a positive association with scientific literacy, their perceptions on the personal value of science had a negative association. Many factors included in the teaching and learning cluster had either no relationship (investigations and interactions) or negative relationship (hands-on experiences) with the scientific literacy of Turkish students. Teacher-student ratio, lack of instructional resources, and science teacher shortage were among the non-significant predictors of Turkish students’ PISA scores.