Abstract: PISA2015 results attracted a wide attention from the whole society. PISA2015 data for B-S-J-G(China)and Singapore were accessed from OECD online database to compare 15-year-old students' science literacy performance. Hierarchical Linear Modeling was employed to explore the key factors that influence students' science literacy performance in two respective economies. Results demonstrate that B-S-J-G(China)students' test scores on the overall scale of science literacy test and on all the subscales of science literacy test are significantly lower than that of Singapore students, with a score gap ranging from 32.82 to 43.29 and a moderate effect size. Students from B-S-J-G(China) also significantly show a stronger instrumental motivation, but less enjoyment of science and less science self-efficacy than Singapore students. The factors such as students' ESCS, enjoyment of science, disciplinary climate in science class, and the ratio of science teachers to all school teachers significantly predicted students' science literacy performance positively, while students' test anxiety, instrumental motivation, and student-perceived enquiry-based instruction had significant negative impacts on students' science literacy performance. In order to continuously reform science education and improve its quality, B-S-J-G(China) should conscientiously learn from the educational experience in Singapore. These measures include emphasizing inquiry teaching and the cultivation of critical thinking ability, strengthening the management of human resources in schools to improve teachers' professionalism, and paying attention to students' non-cognitive performance so as to stimulate their internal motivation.

Key words: PISA, science literacy, science education, key literacy, hierarchical linear regression