Determining the Difficulty and Discrimination Parameters of a Mathematics Performance-Based Assessment

Abstract

Performance-based assessment (PBA) is different from traditional testing methods in that PBA presents real-life problems for students to solve by integrating critical thinking with their content knowledge and skills. Implementing PBA regularly in mathematics classes is associated with improved student achievement and motivation to learn; however, there are concerns about the general lack of psychometric data to support the use of performance assessments. To address such concerns, this study applied item response theory to estimate the difficulty and discrimination indices of items that comprised a newly developed mathematics PBA. Data were collected by administering the PBA to 750 senior high school students in the Western Region of Ghana. The results indicated that the difficulty and discrimination levels of each item were satisfactory, which suggests that well-designed and properly vetted math PBA items would improve classroom assessments as well as high-stakes tests administered on a large scale. Additional recommendations are included at the end of this paper.