Publication Date

2019

Abstract

The 2015 Programme for International Student Assessment (PISA) has drawn a substantial amount of attention from science educators and educational policymakers because it marked the first time that PISA assessed students' ability to evaluate and design scientific inquiry using computer‐based simulations. We undertook a secondary analysis of the PISA 2015 Taiwan dataset of 7,973 students from 214 schools to identify critical issues of student learning and potentially reshape our educational system and policies. Thus, this study sought to identify potential latent clusters of students' scientific literacy performance according to a set of focus variables selected from the PISA student questionnaires. In addition, significant determinants of students' scientific literacy and resiliency were analyzed. Cluster analysis results demonstrated the presence of four clusters of high, medium, low, and inferior scientific literacy/epistemology/affective dispositions. Specifically, students in cluster 1 compared with other clusters showed that the higher the scientific literacy scores are, the more positive epistemic beliefs about science, achievement motivation, enjoyment of science, interests in broad science, science self‐efficacy, information and communications technology (ICT) interest, ICT autonomy, more learning time, more teacher supports and teacher‐directed instructions are. Regression results indicated that the most robust predictor of students' scientific literacy performance is epistemic beliefs about science, followed by learning time, interest in broad science topics, achievement motivation, inquiry‐based science teaching and learning practice, and science self‐efficacy. Decision tree model results showed that the descending order of the variables in terms of their importance in differentiating students as high‐ versus low‐performing were epistemic beliefs about science, learning time, self‐efficacy, interest in broad science, and scientific inquiry, respectively. A similar decision tree model to determine students as resilient versus non‐resilient also was found. Various interpretations of these results are discussed, as are their implications for science education research, science teaching, and science education policy.

School/Institute

School of Education

Document Type

Journal Article

Access Rights

ERA Access

Access may be restricted.

Share

COinS