#791: Semi-Parametric Non-Linear Modeling of Student-Centered Learning Challenges

Student-centered learning (SCL) emphasizes learners’ autonomy, engagement, and responsibility for constructing knowledge. Yet when external supports such as scaffolding, guidance, and authentic audiences are limited, learners may encounter heightened uncertainty and strain. This study uses Electrodermal Activity (EDA) as a real-time indicator of physiological arousal associated with task engagement in unsupported SCL-like tasks and applies semi-parametric spline modeling to characterize temporal dynamics in the signal. Results indicate recurring arousal patterns that, in relation to task context and theory, are interpreted as possible episodes of heightened challenge, brief recovery, and later stabilization. By relating these patterns to the OLSit (Own it, Learn it, Share it) design framework, the paper highlights the importance of balancing autonomy with timely support. The study contributes both a methodological approach for modeling EDA time series and design hypotheses for supporting productive affective engagement.