Examining student, parent, and school factors predicting science achievement using a multilevel approach: the case of Hong Kong from the Program for International Student Assessment 2015

This study utilizes the 2015 data from the Hong Kong sample of the Program for International Student Assessment to examine the relative importance of various factors on the science learning performance of Hong Kong students. Using hierarchical linear modeling (HLM), we examined the effect of students' affective characteristics, involvement, and perceptions of science by parents and school environment on science achievement in Hong Kong. The results of HLM demonstrated that school characteristics explain 33% variance in science achievement, which indicates that school is a critical predicting factor in identifying science achievement. The findings highlighted that identifying as male, having high socioeconomic status (at the student and school levels), high self-efficacy, low test anxiety, less parental support, positive view of science by parents, more school resources, and less teacher autonomy positively influence the science achievement of students. In the final model, school-level predictors, such as disciplinary climate in science classes and teacher-directed science instruction, explained 58% of variance at the school level in positive directions. Lastly, the study discussed the implications of achievement gaps between gender and socioeconomic status and between parents and school influence.