← All reports

Overrepresentation of low-income parents in datasets used for child abuse risk prediction can lead to harmful outcomes.

Data EthicsPredictive ModelingLegal IssuesSocial IssuesApr 20, 2026score 0.172 posts · 0 replies across 1 instances
The thread discusses the potential harm of overrepresentation in datasets used for predictive modeling, specifically highlighting a case in Allegheny County, Pennsylvania, where low-income parents were overrepresented in a child abuse risk prediction dataset.

Claims

Overrepresentation of low-income parents in datasets used for child abuse risk prediction can lead to harmful outcomes.
Parent: Predictive ModelingEntity: Child Abuse Risk PredictionImpact: negativeDate: Apr 20, 2026Target: Overrepresentation of low-income parents in datasets used for child abuse risk prediction

Source posts

@[email protected]
do-no-harm-guide.pdf hypothes.is/a/vW-uzDxzEfG8cxvd3gHSrQ In other cases, however, such overrepresentation can be harmful, such as when parents with low incomes were overrepresented in a dataset that was used as an input to predict the risk of child abuse in Allegheny County, Pennsylvanthis can also be harmful in…
0 boosts · 0 favs · 0 replies · Apr 20, 2026
@[email protected]
do-no-harm-guide.pdf hypothes.is/a/vW-uzDxzEfG8cxvd3gHSrQ In other cases, however, such overrepresentation can be harmful, such as when parents with low incomes were overrepresented in a dataset that was used as an input to predict the risk of child abuse in Allegheny County, Pennsylvanthis can also be harmful in…
0 boosts · 0 favs · 0 replies · Apr 20, 2026