Re-envisioning a K-12 Early Warning System with School Climate Factors

Mengchen Su; Lukas A. Olson; Daniel C. Jarratt; Sashank Varma; Joseph A. Konstan; Rebecca J. L. Keller; Bodong Chen

doi:10.1145/3491140.3528670

What is it about?

The Every Student Succeeds Act (ESSA) prescribes holistic measures of schools for student success and well-being. However, many early warning systems rely exclusively on the "Attendance, Behavior, Course" (ABC) taxonomy, which misses potentially crucial determinants such as school climate and students' socioemotional learning. We report early findings from a larger project that aims to apply machine learning methods to improve an early warning system by incorporating factors related to school climate.

Photo by Andy Kelly on Unsplash

Why is it important?

Above and beyond the traditional "Attendance, Behavior, and course" (ABC) taxonomy, the projects explore the potential of social and emotional related school factors in improving the machine-learned early warning system (EWS) in K-12 education. In addition, the study combines school administrative data and survey data about school climate, which is traditionally separated, to seek means to support students’ holistic development in schools. The study further spotlights and explores the way how to improve the explainability, fairness, and robustness of early warning systems, which could facilitate districts' and schools' use of EWS.

Perspectives

This project is funded by Schmidt Futures. We offer a special thanks to the Nevada Department of Education and American Institutes for Research (AIR) for data preparation, early partnership, and feedback on this project. I hope this work-in-progress paper may help data scientists and researchers to rethink why the early warning model, and how it plays the role to help educators/administrators to better serve K-12 students. After we get the risk scores and information, how to transform it into useful messages to each community, district, school, and parents should be considered too.
Mengchen Su
University of Minnesota Twin Cities

This page is a summary of: Re-envisioning a K-12 Early Warning System with School Climate Factors, June 2022, ACM (Association for Computing Machinery),
DOI: 10.1145/3491140.3528670.
You can read the full text:

Read

Resources

URL
Re-envisioning a K-12 Early Warning System with School Climate Factors
The Every Student Succeeds Act (ESSA) prescribes holistic measures of schools for student success and well-being. However, many early warning systems rely exclusively on the "Attendance, Behavior, Course" (ABC) taxonomy, which misses potentially crucial determinants such as school climate and students' socioemotional learning. We report early findings from a larger project that aims to apply machine learning methods to improve an early warning system by incorporating factors related to school climate and socioemotional learning. These preliminary analyses suggest that a culturally inclusive, socially supportive, and emotionally and physically safe school climate is related to academic success and fewer engagement/behavior problems at the school level. They suggest the promise of integrating these features into early warning systems to help schools change their practices to better support student well-being.

Contributors

The following have contributed to this page

School climate factors in machine-learned early warning model in K-12 education

What is it about?

Why is it important?

Perspectives

Resources