Summary: A Critique of The Markup’s Investigation into Predictive Models of Student Success
Their conclusions include “DEWS is wrong nearly three quarters of the time when it predicts a student won’t graduate on time” and “it’s wrong at significantly greater rates for Black and Hispanic students than it is for White students.”
The three points I want to make are:
It’s worthwhile to build predictive models of student success as part of advising and support systems.
They include a quote from a state employee involved in the system: “the model… over-identifies Black, Hispanic and other students of color among the non-on-time graduates.” That is, the model is mis-calibrated.
The Markup repeatedly makes statements such as “[DEWS] was wrong nearly three quarters of the time it predicted a student wouldn’t graduate on time,” implying that the model is fundamentally flawed.
Recently, tech-journalism site The Markup ran a long, detailed, critical investigation of a predictive machine learning model used by the State of Wisconsin to identify public school students at risk of not graduating.
Learning from historical data, it’s easy for the model to improperly conclude that being non-white reduces your likelihood of graduation, yielding scores that are too low.
Similar Articles
A Critique of The Markup’s Investigation into Predictive Models of Student Success
The Markup’s investigation into student success models in Wisconsin
missed some important context around the value of these models, how
to incorporate race and …
Read the complete article at: www.harlan.harris.name
About The Author
