Using Automated Detectors to Examine the Relationships Between Learner Attributes and Behaviors During Inquiry in Science

Principal Investigator: 
Co-Investigator: 
Project Overview
Background & Purpose: 

We will triangulate survey measures of student attributes, automated detectors of student behavior in log files, and tests of student learning in order to develop highly precise models of how disengaged student behaviors mediate the relationships between student attributes and learning outcomes.

Setting: 

This project takes place in public middle schools in diverse urban and suburban districts in central Massachusetts.

Research Design: 

This project is designed to generate evidence that is descriptive [observational]. Original data are collected through assessments of learning, observation [personal observation and weblogs], and survey research [self-completion questionnaires].

Data will be stored in the Pittsburgh Science of Learning Center DataShop to enable future secondary data analysis. Researchers will build behavior detectors using machine learning and will use classification algorithms to study the behaviors that emerge from specific individual differences. Researchers will use path models to study how student behaviors mediate the relationships between student attributes and learning.

Publications & Presentations: 

No project documents yet to report.

Other Products: 

This project is anticipated to develop new detectors of student disengaged behaviors for Science Assessments.