A Complex Systems Approach to the School to Prison Pipeline

David McMillon

My research employs an under-utilized set of tools to rigorously study how behavioral mechanisms studied in various fields can interactively lead to various (sometimes unintended) consequences for the School-to-Prison Pipeline (STPP) and the achievement gap. Using NLSY data with an approach that models the criminal career as a 5-state Markov chain, I show that prevention is the best cure for long-run crime and incarceration. This illuminates the systemic importance of reducing behavioral infractions and academic failure in schools. Hence I focus chiefly on the school context. I use discrete choice models that are designed to consider how developmental trajectories evolve dynamically through a complex of interacting factors that have largely been studied separately in the literature; but which nonetheless produce important, synergistic effects, including the possibility of tipping points and phase transitions. Focusing on behavioral infractions, I show conditions under which tipping is more likely for a group that is more homogeneous, likely to be harshly punished for the same offense, and for a group whose behavior is likely to be interpreted as deviant, concluding that African-American schools are more likely to exhibit tipping. Using a similar model estimated with the LSAY and focusing on achievement, I show whether tipping is more likely for students with fixed vs growth mindsets and for student groups who are more vulnerable to stereotype threat such as women and minorities. Tipping points allow small changes to make dramatic differences, and leveraging them is crucial in the fight to dismantle the STPP and reduce the achievement gap.

About David McMillon

David B. McMillon is a doctoral candidate at the University of Chicago. He was a Marjorie Lee Brown Fellow at the University of Michigan where he earned two Master's degrees, in Applied Mathematics and Operations Research. He was awarded the National Science Foundation's Graduate Research Fellowship Honorable Mention in 2014 and the Institute of Education Sciences (IES) fellowship in 2016. His research interests lie in the application of cutting-edge quantitative techniques and complex systems theory to contemporary issues in social policy that affect low-income groups. He constructs and estimates mathematical models of crime, mass incarceration, the achievement gap, and the School-to-Prison Pipeline, and has published in the Public Library of Science Journal (PLOS One). These interests emerge from the combination of his mathematical training, his upbringing in inner city Saginaw, Michigan, and his conviction that systemic disadvantage is best-solved using systems thinking. They are ultimately motivated by his faith in GOD, which was sparked by his mother, Dr. Gwendolyn McMillon (a past Spencer fellow), and his father, Rev. Dr. Vincent McMillon. It has been sustained in part through the recognition that average people, for all of their faults and harmful tendencies, are capable of divine love. He is fortunate to recognize that he has never been without love, and this has made him whole in every sense. Finally, this wholeness is a rare privilege in today's world, which is why he considers helping those who were denied that privilege a personal responsibility, as his late grandfather would demand.