The Dropout Early Warning System (DEWS) in Wisconsin predicts students' risk of dropping out using factors such as race, test scores, attendance, disciplinary history, and socioeconomic status. This algorithm misjudges nearly three-quarters of the time, disproportionately affecting Black and Hispanic students. The Wisconsin Department of Public Instruction removed DEWS data from school dashboards after community feedback, prompting a potential reevaluation of the algorithm as part of a broader discussion on racial equity in education. The experiences of students like Mia Townsend and Maurice highlight the algorithm's negative impact on their school experiences.
The Dropout Early Warning System (DEWS) in Wisconsin predicts dropouts based on race, testing and attendance, often mislabeling Black and Hispanic students as high risk.
Predictions from DEWS were wrong almost three-quarters of the time, with a significant bias against Black and Hispanic students compared to their White peers.
After sharing their experiences in an investigation, the Wisconsin Department of Public Instruction removed DEWS data from school dashboards for potential reconsideration.
Mia Townsend expressed that learning about DEWS shocked her, revealing racial biases in the prediction system and the impact of automated labeling on students.
Collection
[
|
...
]