Integrating School Climate Factors in a K–12 Early Warning System: An Exploratory Study

• In Event: Marriott Roundtable Session Seven
  In Roundtable Session: Data Analysis and Data-Driven Decision Making (Table 1)

Abstract

Holistic approaches to education look beyond academic performance when considering student well-being. However, current machine-learned “early warning systems” exclusively rely on measures of attendance, behavior, and course performance, missing crucial factors like school climate and socioemotional learning. This study evaluates the integration of school climate survey data into an early warning system trained on electronic learning records. We found school climate factors enhanced the attendance-features-only model but not the overall model with all features. When further comparing model improvement across racial, gender, and socioeconomic status groups, we found that predictive quality did not differ across subgroups. These findings show both promise and limitation of using school climate/SEL data to predict educational outcomes.

Authors

• Mengchen Su, University of Minnesota
  Presenting Author

• Daniel Jarratt, Infinite Campus, Inc.
  Non-Presenting Author

• Sashank Varma, Georgia Institute of Technology
  Non-Presenting Author

• Joseph Konstan, University of Minnesota
  Non-Presenting Author

• Rebecca J.L. Keller
  Non-Presenting Author

• Bodong Chen, University of Pennsylvania
  Non-Presenting Author

• Lukas Andreas Olson, QuantumBlack
  Presenting Author