Title Information
Type of Resource (primo): dissertations
Name: Personal
Name: Personal
Name: Personal
Name: Personal
Name: Personal
Name: Personal
Name: Corporate
Origin Information
Physical Description
Note: thesis: Thesis (Ph. D.)--Brown University, 2026
Genre (aat): theses
Abstract: This dissertation examines the contours of algorithmic eligibility determination for Medicaid Long-Term Services and Supports (LTSS). In doing so, I provide the first population-level evidence on how algorithmic decision-making changed over the past decade and how different types of algorithmic decision-making are associated with inequality in access to LTSS. Prior research on algorithmic decision-making and inequality has been idiosyncratic, examining the social dimensions of a single algorithm in a single context. Leveraging variation in Medicaid LTSS eligibility determination, I show how algorithms’ characteristics interact with the political and organizational contexts to shape inequality. I propose a novel typology for characterizing algorithmic decision-making and have developed an original dataset of the algorithms used in Medicaid LTSS eligibility. Chapter 1 examines how these algorithms changed from 2012 to 2020 and finds that the use of fully automated algorithms decreased while automated algorithms that incorporate a “human-in-the-loop” increased. States where most beneficiaries are evaluated by this second type of algorithm tend to be more politically conservative. Chapter 2 shifts to analyzing the consequences of different types of algorithms for racial inequality in nursing home care from 2012 to 2020. Using first differences regression analysis, I find that shifts from fully automated algorithms that employ medical professionals to algorithms with different combinations of automation and user expertise are associated with decreasing percentages of non-white residents in nursing homes. However, I find that as state-level Medicaid generosity increases, these differences become less pronounced. While the first two chapters provide a national perspective on algorithmic decision-making, Chapter 3 uses in-depth interviews and analysis of eligibility appeal records in Wisconsin to understand how a Medicaid eligibility algorithm is used in practice. I show that even within the same state, the same algorithm can be used differently across organizational contexts, namely eligibility screeners are bound to the algorithm’s standardized structure while appeal judges have more flexibility. This dissertation highlights the growing importance of human actors in algorithmic decision-making and shows how organizational factors, state political contexts, and the characteristics of algorithms work in tandem to produce inequality in resource allocation.
Subject (fast) (authorityURI="http://id.worldcat.org/fast", valueURI="http://id.worldcat.org/fast/01041902")
Subject (fast) (authorityURI="http://id.worldcat.org/fast", valueURI="http://id.worldcat.org/fast/01013657")
Subject
Subject
Language
Record Information