Poster #3 - Modeling Student Interactions With an Online Map to Predict Differentiated Understanding of Scientific Observations in an Inquiry-Based Learning Environment

• In Event: 15.062 - MTCC Poster Session One
  In Poster Session: 15.062-1 - Computer-Based Experts and Models

Abstract

There has been an increasing interest in the benefit of automated, formative assessments while learning from visual representations. In this study, we evaluated data-mined models to assess moment-to-moment student learning using an interactive map. We collected data from 143 middle school students, who completed two modules of Research Quest, an inquiry-based learning environment designed to support students in scientific investigations using digitized fossils and paleontologist resources. Examination of the accuracy and error estimates for each model was found to outperform that of the majority class classifier (i.e., the baseline, equivalent to the best guess). We discuss the implications of a two-fold nested resampling method estimating the prediction error of models that drive scaffolding of interactive visual materials in real-time.

Authors

• Eric G. Poitras, University of Utah

• Kirsten R. Butcher, University of Utah

• Matthew Orr, University of Utah

• Michelle Hudson, Brigham Young University

• Madlyn Runburg, Natural History Museum of Utah