Remedial Instructions

Purpose of Study: This study is an application of differential person functioning (DPF) on the data of Taiwan PISA 2015
science literacy. We use this method to understand the strength and weaknesses of Taiwan students in science literacy. It is also
hoped that the results could help improve the content of the curriculum and remedial instruction.
Methodology: We transpose the person-item data matrix into an item-person matrix and use the Mantel-Haenszel method to
analyze this matrix. The focal and reference groups change from persons to item clusters, that is, cognitive domains in the
PISA science assessment.
Principal
Findings: Results show the proportions of DPF are very few. This implies most Taiwan students have a comprehensive
science literacy. Up to 5.9% of students perform differentially in three science competencies. Most DPF students perform better in
the explain phenomena scientifically (EP) competency than the others no matter what proficiency level they belong to. This
may reflect the current situation of science education in Taiwan. The experts and teachers should develop more curriculums to
reinforce the students’ competencies about evaluate and design scientific enquiry (ED) and interpret data and evidence
scientifically (ID).
Limitations: The effect sizes of DPF are not provided, the degree of DPF is unidentified. DPF analyses only test the relative
difference in two cognitive scales. Future studies can incorporate another cognitive diagnostic model to obtain more information
in detail.
Importance of Study: The PISA assessments provide a favourable framework to examine students’ literacy. The results of
international reports show that Taiwan students performed well in science. The application of differential person functioning
can find out the weaknesses of Taiwan students in science literacy. Furthermore, we applied the DPF procedure to explore the
characteristics of students in different levels of performance.