An assessment database for supporting educational research

Abstract

One of the challenges of research in science education is storing, managing and querying the large amounts of diverse student assessment data that are typically collected in many Science, Technology, Engineering and Mathematics (STEM) courses. Furthermore, longitudinal studies across courses and ABET accreditation necessitate tracking students throughout their academic programs in which each course will have different types of data. Researchers need to manage, assign metadata to, merge, sort, and query all of these data to support instructional decisions, research and accreditation. To address these needs we have constructed a database to support both data-driven instructional decision making and research in STEM education. We have built upon existing metadata standards to define an extensible Educational Metadata Language (EdML) that enables assessments to be tagged based on taxonomies, standard psychometrics such as difficulty and discrimination, and other data to facilitate cross-study analyses. Once a collection of assessment data are available, faculty can examine their assessment data to evaluate historical trends, analyze the effectiveness of pedagogical techniques and strategies, or compare the performance of different teaching and assessment techniques within their course or across institutions.