Makers, Gamers, and Assessment: Machine Learning and Design Frameworks to Automate Assessment in Open Environments

• In Event: The Use of Process Data in Educational Assessment

Abstract

Assessment in activity-rich environments such as games, simulations and maker spaces introduces novel challenges to assessment design. Fortuitously, more than a decade ago a National Research Council (2001) advocated for use of rigorous assessment design frameworks that can be leveraged in these contexts. The recommended design frameworks are particularly valuable in contexts that emphasize complex constructs and complex activities in relatively open learning environments that require use of machine learning techniques. One such design framework - the Bear Assessment System - is described along with required modifications for its use in conjunction with machine learning techniques. A summative assessment for an e-textiles unit is presented as a concrete example.

Authors

• Seth Corrigan, GlassLab Games

• Sunrito Bhattacharya

• Kylie A. Peppler, Indiana University - Bloomington

• Constance Yowell