Using Rasch Model to Evaluate Physics Competence Tests and its Implications for Teachers’ Development of Competency-based Items

Abstract

Core competencies are the focus of the 12-year basic education curriculum reform. The purposes of the study were to introduce the development and validation of two physics competence tests through the cooperation of the seed teachers in the physics curriculum center, K-12 Education Administration, Ministry of Education, with a science education researcher, and to investigate strategies for developing competency items based on the analysis of multidimensional partial credit model, one of the Rasch models, and the interpretation of Wright maps. Forty-two physics competence items were selected and divided into test A and test B with a similar difficulty level. Stratified and convenience sampling was adopted. Two tests were administered to 1119 Grade 12 students from 19 public senior high schools in Taiwan. The results showed that most of the items in the two tests fit the multidimensional partial credit model, and the average of item difficulty for test A and test B were approximately 0. However, based on the Wright maps, some items were easier, and there was a lack of items with higher difficulty to assess students with higher ability. The study intends to inspire practical educators to build a national assessment framework for scientific literacy and to develop constructed-response items, items assessing the competence of evaluating and designing scientific inquiry, and items with high difficulty. The use of Wright maps may enhance teachers’ sensitivity to item difficulty and their understanding of students’ ability.

Keywords

• assessment
• item response theory
• physics competence
• rasch model