Search

This dissertation considers the importance of prosecutors as legal actors in shaping the welfare of defendants in the US criminal justice system. Broadly, it shows that prosecutor choices have significant impacts on the charges and sentence lengths defendants receive and that racial minorities disproportionately receive harsher punishment from these decisions. In the first chapter, I consider whether changing the costs or benefits to prosecutor charging decisions impacts racial disparities in charges. Prosecutors face a tradeoff between expending resources to gather evidence and building a stronger case. I exploit a circuit split in charging rules to observe variation in the cost of evidence requirements. I compare how prosecutors differentially increase charges for racial minorities across each evidence cost regime by utilizing a bunching point in mandatory minimum charging of federal drug crimes. Using a difference-in-differences design, I find that when evidence costs are lowered, prosecutors disproportionately increase charges for Black defendants compared to White counterparts. I then consider how decreases to the expected sentence length for increased charges affects prosecutor choice and racial disparities using the 2018 First Step Act (FSA). I find that in circuits with high evidence costs, the FSA decreased the race gap in increased charges. But in circuits with low required evidence, the FSA increased this disparity. The results suggest that increasing evidence costs is an effective policy tool for reducing racial disparities in charges. In the second chapter I consider the impact of mandatory sentencing. I estimate the actual increase to sentence lengths created by mandatory minimum eligibility for federal drug crimes. I decompose eligibility effects into effects from the law and effects driven by manipulation of cases near the eligibility cutoff. This is done by fitting distributions through non-manipulated cases and extrapolating these distributions int

Background In Australia, the National Safety and Quality Health Service Standards (2012) stipulates that partnering with health consumers to improve health‐care experiences is one of the criteria health‐care organizations are assessed and accredited against. This standard has given rise to a role: health engagement professionals (HEPs). While there are no standard requirements for recruitment into this role, this study contributes to much needed research into understanding their responsibilities and capabilities, and their contributions to engagement outcomes. Methods Using a qualitative, interpretive approach, 16 HEPs and 15 health consumer representatives (who have experiences of interacting with HEPs) participated in an in‐depth phone interview in December 2019. We explored (a) the purposes of the role, (b) the responsibilities and work activities and (c) the capabilities required to carry out the responsibilities. Results Health engagement professionals are specialists in designing engagement mechanisms for health‐care organizations to co‐design health services with health consumers. They facilitate partnerships between health‐care organizations and health consumers. They play significant roles in listening to, facilitating understanding amongst different stakeholder groups (eg hospital management, health‐care workers and health consumers) and navigating the bureaucratic structures to influence outcomes. Four major responsibilities (advocacy, education, facilitation and administration) and four categories of capabilities (relational, communication, professional and personal) were identified. Conclusion A list of job responsibilities and desired capabilities of HEPs is provided to help health‐care organizations better understand the requirements for the role. This would help them decide how applicants to these roles would meet the requirements (eg experience of navigating bureaucratic systems).

Chapter 1 examines the effects of state-level workers’ compensation reforms on workers’ access to compensation benefits. In nearly all liability settings the probability that an accident victim will receive payments and the amount of those payments are strongly influenced by laws or rules related to burdens of proof, legal procedure, access to medical treatment, and other nonquantitative factors. Yet, their impact is often difficult to measure empirically because it is difficult to measure the true liability and size of damages because those are also at issue in the proceedings. We study the impact of such nonquantitative laws on average accident payments per covered worker under workers’ compensation, where fault is not at issue and damage payments are set by statute. The results show that states experienced sizeable drops in cash and medical payments per covered worker of 20 and 16 percent, respectively, after they adopted burden of proof laws. A decomposition of national means between 1997 and 2016 shows that about 11 percent of the decline in the national average of benefits per covered worker might be attributed to the patterns of adoption of these laws. The analysis also shows that average statutory expected benefits rose during this period and thus served to counteract the decline in benefits per covered worker. Chapter 2 investigates the causal effects of restricting prosecutorial discretion on racial and gender disparities. In North Carolina, sentence enhancements may be imposed if prosecutors meet the legally required burden of showing the existence of aggravating factors. Blakely v.Washington 542 U.S. 296 (2004) imposed a significant constraint on state prosecutors’ discretion by raising their burdens of proof in successfully seeking sentence enhancements from “preponderance of evidence” to “beyond a reasonable doubt.” First, we use a regression discontinuity design to exploit the timing of the Blakely decision and find that Blakely decreases the defendan

Cecil Edward (Tom) Barnes took part in the famous sabotage operation of the Gorgopotamos Bridge in November 1942 as leader of the demolition party. Remaining in Greece after the attack, he rose to become one of two Area Commanders for the Allied Military Mission in Greece. The archive he left behind at his death in 1952 includes over 1000 photos, war diaries, letters and reports, mainly relating to his experiences in Greece. This study demonstrates how these eye-witness records illuminate important aspects of these critical years in modern Greek history including the Gorgopotamos operation itself and the trek across Greece which followed, the "Animals" operation which was conducted to convince Hitler that the Allied landings would take place in Greece and distract his attention from Sicily, surrender overtures from the German commander in Epirus (General Hubert von Lanz), and the so-called first two rounds of Civil War in 1943 and 1944

This purpose of this project was to understand how Arizona tenants experience the judicial process of summary eviction and whether that experience might be meaningfully improved. While the psychological, social, and economic effects of eviction are currently in the national spotlight, few scholars have investigated the legal documents that quietly shunt tenants out of their houses and the justice innovations designed to help them resist. In addition, there has been surprisingly little qualitative research of tenants during the eviction process that investigates why they choose to attend their hearing, how they perceive housing court, and whether they are able to accomplish their stated goals through their attendance and participation. By comprehensively analyzing the tenant experience, I provide court practitioners with recommendations for redesigning housing court more equitably. This project is comprised of four stand-alone studies. Chapter 1 describes a linguistic and rhetorical analysis of the documents that evict. As a case study, I analyze the notice and pleading documents most filed in an Arizona housing court, measure their real-world impact, and present model eviction notice and pleading forms. Chapter 2, written with Margaret Hagan, integrates existing expert-oriented and user-centered approaches to designing justice innovation and presents a first attempt at establishing a standard methodology for creating and vetting new justice interventions. We present the methodology used to create the eviction self-help materials used in this project as an example and argue that a human-centered, participatory approach should become the standard in justice innovation design. Chapter 3, written with Andy Yuan, describes the results of a randomized study to provide self-help information to tenants facing eviction in one Arizona court. We find no evidence that provision of self-help improved court attendance or case outcomes; in contrast, treated tenants were significan

I attempt to address an important issue of the portfolio allocation literature – none of the allocation rules from prior studies consistently delivers good performance. I develop an approach that aggregates information from a wide range of sources to make allocation decisions. Specifically, this approach models the optimal portfolio weights as a function of a broad set of portfolio weights implied by prior allocation rules, and determines the relative contribution from each allocation rule through Elastic Net, a machine-learning technique. Out-of-sample tests suggest that my approach consistently achieves good performance, whereas none of the alternative rules can match the consistency.

A test question is said to contain Differential Item Functioning (DIF) if people who have the same ability level but are from different subpopulations have unequal probabilities of answering correctly. Detection of DIF is crucial to maintaining test fairness. Numerous DIF detection methods exist; none are ideal across a wide range of settings. I introduce a randomization test for DIF detection, and compare its performance to that of the Mantel-Haenszel, Logistic Regression and Lasso-based methods under a Rasch model using Monte Carlo simulations.

The incidence of complications secondary to fracture stabilization, particularly osteolytic lesions and bony tumor formation, has long been difficult to evaluate. The objective of this study was to describe the long-term incidence of aggressive bony changes developing in dogs with long bone diaphyseal fractures stabilized by metallic bone plates compared to a breed-, sex-, and age-matched control group. The medical records of a tertiary referral center were retrospectively reviewed for dogs that matched each respective criterion. Signalment, history, cause of death (if applicable), and aggressive bony changes at previous fracture sites were recorded. Ninety dogs met the criteria for inclusion in the fracture group and were matched with appropriate control dogs. Four of the dogs in the fracture group developed aggressive bony changes at the site of previous fracture repairs most consistent with osseous neoplasia. One lesion was confirmed with cytology as neoplastic. The population of dogs was mixed with regard to breed and body weight, but all dogs with aggressive bony lesions were male. Incidence of aggressive bony lesion formation in the fracture group was 4 (4.4%) and was 0 (0%) in the control group; three (75%) of the affected dogs in the fracture group included cerclage as a component of their primary fracture stabilizations. Incidence of aggressive bony lesions in the fracture group compared to the control group was determined to be statistically significant (p = 0.0455), as was the incidence of cerclage among dogs affected by aggressive bony lesions compared to the rest of the fracture group (p = 0.0499). Development of aggressive bony lesions is an uncommon complication of fracture fixation. Additional research is needed to further identify and elucidate the long-term effects of metallic implants in dogs.

In classical mixture modeling, each data point is modeled as arising i.i.d. (typically) from a weighted sum of probability distributions. When data arises from different sources that may not give rise to the same mixture distribution, a hierarchical model can allow the source contexts (e.g., documents, sub-populations) to share components while assigning different weights across them (while perhaps coupling the weights to "borrow strength" across contexts). The Dirichlet Process (DP) Mixture Model (e.g., Rasmussen (2000)) is a Bayesian approach to mixture modeling which models the data as arising from a countably infinite number of components: the Dirichlet Process provides a prior on the mixture weights that guards against overfitting. The Hierarchical Dirichlet Process (HDP) Mixture Model (Teh et al., 2006) employs a separate DP Mixture Model for each context, but couples the weights across contexts. This coupling is critical to ensure that mixture components are reused across contexts. An important application of HDPs is to time series models, in particular Hidden Markov Models (HMMs), where the HDP can be used as a prior on a doubly infinite transition matrix for the latent Markov chain, giving rise to the HDP-HMM (first developed, as the "Infinite HMM", by Beal et al. (2001), and subsequently shown to be a case of an HDP by Teh et al. (2006)). There, the hierarchy is over rows of the transition matrix, and the distributions across rows are coupled through a top-level Dirichlet Process. In the first part of the dissertation, I present a formal overview of Mixture Models and Hidden Markov Models. I then turn to a discussion of Dirichlet Processes and their various representations, as well as associated schemes for tackling the problem of doing approximate inference over an infinitely flexible model with finite computa- tional resources. I will then turn to the Hierarchical Dirichlet Process (HDP) and its application to an infinite state Hidden Markov Model, the H

An important objective in statistical risk assessment is estimation of minimum exposure levels, called Benchmark Doses (BMDs) that induce a pre-specified Benchmark Response (BMR) in a target population. Established inferential approaches for BMD analysis typically involve one-sided, frequentist confidence limits, leading in practice to what are called Benchmark Dose Lower Limits (BMDLs). Appeal to hierarchical Bayesian modeling and credible limits for building BMDLs is far less developed, however. Indeed, for the few existing forms of Bayesian BMDs, informative prior information is seldom incorporated. Here, a new method is developed by using reparameterized quantal-response models that explicitly describe the BMD as a target parameter. This potentially improves the BMD/BMDL estimation by combining elicited prior belief with the observed data in the Bayesian hierarchy. Besides this, the large variety of candidate quantal-response models available for applying these methods, however, lead to questions of model adequacy and uncertainty. Facing this issue, the Bayesian estimation technique here is further enhanced by applying Bayesian model averaging to produce point estimates and (lower) credible bounds. Implementation is facilitated via a Monte Carlo-based adaptive Metropolis (AM) algorithm to approximate the posterior distribution. Performance of the method is evaluated via a simulation study. An example from carcinogenicity testing illustrates the calculations.

This dissertation consists of three essays evaluating topics in industrial organization. The first essay investigates a market structure or property regime in which a final good exists only by assembling multiple, monopoly-supplied components. In such dynamic settings, any sunk cost results in an outcome of hold-up, also known as tragedy of the anticommons. I design a model showing conditions for which two factors that reduce sunk cost, refunds and complementarities, mitigate hold-up. If the first component purchased has positive stand alone value or the first seller offers a full refund, hold-up is mitigated. My results suggest several policies that can mitigate inefficient outcomes in assembly problems, including legal requirements on full refunds, regulation on the purchasing order of components, and prohibition of price discrimination. The second essay applies Bayesian statistics to single-firm event studies used in securities litigation and antitrust investigations. Inference based on Bayesian analysis does not require an assumption of normality that potentially invalidates standard inference of classical single-firm event studies. I investigate ten events, five from actual securities litigation cases. Various Bayesian models, including replication of the frequentist approach, are examined. A flexible Bayesian model, replacing parametric likelihood functions with the empirical distribution function, also is explored. Our approach suggests an alternative, valid method for inference with easy implementation and interpretation. The third essay, motivated in the context of pharmaceutical advertising, analyzes demand rotations caused by an exogenously determined advertising parameter under Cournot oligopoly competition. We find that firms and consumers prefer extreme levels of advertising, but preferences for which extreme do not necessarily align. However, these differences can be alleviated with few or many firms in the market or cheap or expensive technologies. T