Bayesian Versus Frequentist Estimation for Item Response Theory Models of Interdisciplinary Science Assessment

Abstract

Along with the trend emphasizing ID learning, ID assessments to measure students’ ID understanding have been developed by several scholars. The interdisciplinary science assessment for carbon cycling (ISACC) was developed to assess ID understanding among high school and college students in integrating knowledge from different science disciplines to explain a scientific phenomenon, global carbon cycling. The ISACC’s construct validity was checked using traditional item response theory (IRT) models in 2021. The current study was motivated by the desire to reveal the difference in IRT analysis results of the ISACC using a Bayesian approach in comparison with the results using the traditional approach. The Bayesian approach has several strengths over the traditional IRT. The results of the study imply the need for additional research for the development and validation of interdisciplinary science assessments through strong psychometric properties.