Using Principled Assessment Design and Item Difficulty Modeling to Connect Hybrid Adaptive Instructional and Assessment Systems: Proof of Concept
| Autor: | M. Christina Schneider, Jing Chen, Paul D. Nichols |
|---|---|
| Rok vydání: | 2021 |
| Předmět: | |
| Zdroj: | Adaptive Instructional Systems. Adaptation Strategies and Methods ISBN: 9783030778729 HCI (32) |
| DOI: | 10.1007/978-3-030-77873-6_11 |
| Popis: | Coherence between adaptive instructional and summative assessment systems should provide teachers stronger support in challenging each student. Coherent systems should lead to accelerated student learning, and students being ready for the next grade. Range achievement level descriptors (RALDs) describe a state’s theory of what increasing knowledge, skills, and abilities look like in their standards as students become more sophisticated thinkers on their journey to proficiency and beyond. Systems can be connected to better align interpretations of student performance using task features that align to evidence statements in RALDs. Our proposition is that through coding tasks in both systems using common schema, including RALD-to-task match, assessment and instructional system inferences about student progress can be bridged. We combined two approaches for linking inferences across systems measuring mathematics that do not rely on common students or common tasks as a proof of concept. Using RALDs and other task features, we predicted 45% of the variance in task difficulties for a secondary data source. Holding all else constant, RALDs were the strongest feature for modeling increases to task difficulty. This suggests that RALDs could be leveraged in instructional systems to support interpretations of student growth, increasing their value for teachers. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|