I'm currently at Princeton pursuing a PhD in Sociology and Social Policy, with a Demography specialization through the Office of Population Research.  In Fall of 2020, I will be joining Dartmouth's Quantitative Social Science Department as an Assistant Professor. Currently, I am visiting as a Data Scientist with The Lab at DC and working as an Associate Fellow with GSA's Office of Evaluation Sciences on projects at the intersection of data science, education, health, and housing.

My broad interests are in how law and ethics shape the allocation of scarce resources. 

My dissertation focuses on one way that bureaucracies allocate scarce resources: human committees decide which categories count as needy (e.g., having a disability; being a child) and prioritize individuals who display a higher count of needy characteristics. I study these dynamics in school districts. State legislatures allocate more resources to districts that have a higher presence of students with tags for characteristics like poverty, language status, and disability. Past research largely focuses on how tags shape allocations to districts. Less is known about what happens to resources once they enter the district. I study how the due process rights that accompany different tags shape whether and how funds reach tagged recipients. Here are a couple visualizations related to the project:

  • In which school districts do parents file complaints over resources? These are some results from geo-coding complaints that parents of children with disabilities file against school districts arguing that the district failed to provide an appropriate amount of services, obtained via an open records request in New Jersey. The figure highlights that while higher socioeconomic status (SES) districts face more complaints, there's considerability variability within SES levels:  https://rjohnson.shinyapps.io/nj_dp_visualize/.
  • What do supporters of parents versus supporters of school districts emphasize when arguing over resources? a recent Supreme Court case (Endrew v. Douglas County) covered what level of benefit/resources districts owe students with disabilities. This visualization of the text of the amicus briefs filed in favor of the parents versus the district shows that briefs for the district emphasize terms related to budgets, deference to districts, and (wariness) of judicial intervention, while those for the parents emphasize terms like college, development, and early intervention: https://rjohnson.shinyapps.io/endrew_amicus_visualize/

map loadingRecently, human services agencies have moved from using discrete tags to define who counts as needy to using predictive modeling. For instance, rather than allocating resources based on an individual's additive count of 'needy' characteristics, bureacracies might use machine learning to predict an individual's risk of some bad outcome (e.g., dropping out of school) and then allocate resources to individuals with risk above a threshold. In Summer of 2018, I participated in a Summer Fellowship with the Data Science for Social Good program, where we worked on one such algorithm for helping a NYC agency quantify which tenants face the highest risk of landlord harassment to then help the agency prioritize those individuals for resource allocations.


Before coming to Princeton, I received a B.A. from Stanford with honors in Psychology, minors in Economics and Religious Studies, and an M.A. focused on political theory. After graduation, I was a summer research associate at the University of Pennsylvania’s Scattergood Program for the Applied Ethics of Behavioral Healthcare, then spent two years as a pre-doctoral research fellow at the National Institutes of Health's Department of Bioethics, with research focused on dual meanings of diagnostic labels as potential sources of stigma or tags for accessing resources.


Google scholar: google scholar page

Github: github page