Understanding Self-Directed Learning in an Online Laboratory

Abstract

Many studies on using online laboratories for learning focus on pedagogical contexts in K-12 education with well-defined problems as well as well-defined learning goals, assessments, and outcomes. We describe a study on the use of an online laboratory for self-directed learning through the construction and simulation of conceptual models of ecological systems. The learning goals and the demographics of the learners in this study are unknown; only the modeling behaviors and outcomes are observable. We analyzed the modeling behaviors of 315 learners and 822 instances of learner-generated models using machine learning techniques such as clustering and dimensionality reduction. We found three types of learner behaviors: observation (focused on simulation), construction (focused on conceptual model), and exploration (full cycle of model construction, simulation, and revision). We found that while the observation behavior was most common, the exploration behavior led to higher model quality.