The Influence of Different Measurement Approaches on Student Affect Transitions Using Ordered Networks
摘要
Affect detection is integral to creating affect-sensitive learning systems, but the impact of measurement methods needs further research. This paper uses ordered network analysis (ONA) to compare the affect dynamics of two suites of affect detectors trained on complementary data (i.e., labels from an in-the-moment self-reporting (SR) tool vs labels from field observations) in game-based learning. We then use ONA difference models to assess how divergence in learning and motivational measures impact affect dynamics.