Since the 1920s, plea bargaining in the United States criminal justice system has become status quo. Today, over 95% of federal cases have resulted in plea bargains and never been tried in court. The power of the prosecutor within the legal system has also increased substantially over the last century leading many scholars to believe that the proliferation of plea bargaining is a direct result of heightened prosecutorial power. Prior literature has addressed the increased power of the prosecutor and the need for reform if plea bargaining rates are to change, but has not yet provided an empirical observation as to whether prosecutorial reform makes a significant impact in plea bargaining. In this paper I theorize that the prosecutor’s discretion over exculpatory evidence plays an integral part in states’ reliance on plea bargaining and that states that attempt to reign in prosecutorial power experience reduced rates of pleas. To test this theory systematically, I have run a multivariate analysis at the county-level to compare rates of plea bargaining across states that have reformed prosecutorial discretion and states that have not. If states with a model rule suppressing prosecutorial power show a statistically significant difference in plea rates, criminal justice reform advocates should turn their attention to prosecutorial reform.

One of the most common questions that scholars of democratic theory address is how to best allocate and balance power across the different structures within a government. Historically, many theorists and philosophers have postulated that structures with separation of power are more effective at resisting tyrannical rule. Though researchers have established the effects of an imbalance of power between two branches of government, no study has attempted to systematically account for the relative balance of power among all three branches working in tandem or develop empirical metrics to this end. In this study, I theorize that designing separate branches of government that are equally strong strengthens conflicts across institutions, which ultimately leads to more robust protections against tyranny. I evaluate this model quantitatively by developing and introducing a new measure, the Separate Powers Index (SPI). My SPI assesses the balance of power between the three branches of government as postulated in a sample of 130 of the world’s constitutions. Using multivariate regression methods, I compare the SPI with a cross-national index of free expression in a cross-sectional analysis during the year 2008, I can systematically examine whether there is a relationship between structural provisions of institutions that distribute power and their outcomes to protect their citizens. In addition to providing a novel measure of tripartite power balance in national constitutions, the result of this study has a large impact on all scholars of constitutionalism and civil liberties.

Scholars recognize that politicians’ perceptions of their electability influence their home style, or the way in which they present themselves to their constituents. Marginality, or the proportion of a politician’s co-partisans in the electorate, is common indicator of electability. However, marginality fails to capture how polarization of the electorate augments politicians’ vulnerability. This study introduces a new indicator of electability that captures statewide polarization in the electorate. Building off the finding that more marginal Senators emphasize support for appropriations to build non-partisan support, I expect that Senators in more polarized states will emphasize their support for appropriations after controlling for marginality. Appropriations credit-claiming builds non-partisan support without alienating more partisan voters. I employ multilevel linear regression analysis to examine the relationship between state-level partisan polarization and topic expression in Senate press releases systematically. A positive relationship between partisan polarization and appropriations credit-claiming may reveal a mechanism by which polarization paradoxically minimizes the partisan content of Senators’ home styles.

For this project, I examine why the majority of white women voted for Trump in the 2016 election, a fact that came as a shock to many observers. However, with few exceptions, white women have been shown to support conservative candidates in presidential elections since the 1950s. Yet, Donald Trump’s election can be considered unique as he was still able to gain most white women’s support despite his displays of various behaviors that could be regarded as overtly sexist. In this research design, I analyze white female voters using data from a 2016 Pilot Study by the American National Election Survey and hypothesize that racial resentment and internalized sexism are the primary factors that drove white women to support Donald Trump, controlling for partisanship, economic anxiety and other factors that might influence vote choice. I employ linear regression models using R programming software to examine a relationship between 2016 vote choice, racial resentment and internalized sexism systematically. In doing so, my analysis takes an intersectional approach, where both race and gender dynamics are useful in providing an explanation for white women’s support for Trump.

Although legislative partisanship has traditionally been studied through measures like floor votes or debates, scholars have found that social media also provides a conducive environment for negative and positive displays of party politics. This project codes Tweets from both Washington and Texas State legislators in 2017 as either “partisan” or “neutral” in order to create a proportion for each category and correspondingly identify the extent to which these legislators participate in partisan behavior online. Then, these proportions are compared with an individual legislator’s roll call votes, markers of their general level of polarization, in order to view whether or not they are behaving in a more partisan manner online than their votes would indicate. My paper posits that the unique conditions of social media cause legislators to behave differently, resulting in a comprehensive increase in legislator partisanship. This research holds importance in future studies by shedding light on how social media is used by those in our state governments, especially as it pertains to their party posturing online.

Methylmercury (MeHg) is a bioaccumulative neurotoxin, dangerous to human health even at trace levels. In inundated soils, MeHg is formed from inorganic mercury by mercury-methylating microorganisms; a process termed methylation. Demethylation, by contrast, converts MeHg into less-dangerous inorganic mercury, and also occurs via microbial activity throughout the aquatic soil profile. Rice grains can be contaminated with MeHg when grown in soils where methylation rates are high; human exposure to MeHg is thus a serious public health concern in places where rice cultivation, high rates of consumption, and soil mercury (Hg) contamination overlap. Our research aims to better understand the soil conditions that favor demethylation over methylation – this information can then be used to reduce rice grain contamination through agricultural practices or rice breeding programs. Specifically, our research focuses on the role of oxygenation and carbon root exudates on the net MeHg accumulation throughout the soil profile. Rice plants grow in flooded, oxygen-free (anoxic) soils, but their roots can leak oxygen (making the rice rhizosphere oxygenated in varying degrees), as well as carbon root exudates. Our project simulated both fully oxic and transiently-oxic (transition) zones, with two different levels of root exudates; we use isotopic tracers to assess respective methylation and demethylation rates in all four treatments in both the vegetated (rhizosphere) and non-vegetated (bulk) soil. Carbon root exudates have been collected from hydroponically-grown rice variety M-206, and can be applied to different soil zones via tubules. Oxygenation of the soil can be measured with mm-scale optode imagery, which allows delicate testing of various oxygen-introduction designs. My role in this interdisciplinary project has been to develop, scale-up, automate, and verify the accuracy and dependability of root-oxygenation and root-exudate introduction systems to be used in upcoming experiments.

LGBTQ+ students face overwhelmingly negative statistics in and out of school, here and across the world. Supportive teachers and the educational efforts they are involved in like GSAs (Gay Straight Alliance), inclusive curriculum, and specifically inclusive policies make significant positive effects on LGBTQ+ students. The teachers that can understand these students the most though are LGBTQ+ themselves. This project explores the following questions. 1. How do teachers who identify as LGBTQ+ interact with their public education system? 2. Do teachers who identify as LGBTQ+ feel supported in the daily work life and in their professional goals as teachers? 3. What insights do teachers who identify as LGTBQ+ have for their LGBTQ+ students? Through collaboration with various equity and educational leaders, primarily from Washington State and the UW College of Education, a Google Forms survey was created and shared through various digital platforms. Qualitative data was taken from the survey responses of 20 LGBTQ+ identifying teachers in the United States and 2 in China. Responses were analyzed using coding techniques where I deep read responses and categorized them by relationships and on a scale of trust. A majority of teachers were open about their identity and stated feeling supported by their schools. However, analysis revealed that most long answer text responses showed varying levels of mistrust: between teacher and faculty; teacher, student and the school community; as a factor of race and ethnicity; and learned from their life history from student to teacher. The suggested solutions are to recreate and implement discrimination and protective policies, curriculum, and space and face reforms that are specifically inclusive to the entire range of identities that LGBTQ+ individuals inhabit. Additionally, more research needs to be done in how policies and LGBTQ+ issues affect the symbiotic relationships between students and educators.

Since the 1990s, legislation in many US states has increasingly criminalized and restricted undocumented immigration. However, despite these restrictive policies, the population of undocumented immigrants has increased and overall immigration flows from Mexico to the US only began to decline following the 2009 recession. This research responds to decades of increasingly punitive policy and intends to test the efficacy of these policies. Many studies have focused on the efficacy and implications of federal policies, but far less attention has been given to state-level policies in the US. This paper analyzes the efficacy of punitive, state-level immigration laws from 2010 to 2016 in the United States. I hypothesize that there is no significant relationship between the passage of punitive state-level immigration policies and the year-to-year change in state populations of undocumented immigrants. However, I also hypothesize that the introduction of punitive state-level immigration policies affects yearly immigration flows—or the number of people immigrating from Mexico to a specific US State. To conduct this study, I created an index that aggregates a broad spectrum of laws, including policing, licensing, education, public benefits and labor. I then selected a few high-profile state laws intended to deter or curb undocumented immigration. I used multivariate regression analyses to test whether the introduction of punitive policy immigration policy systematically influences migration flows and changes in migrant populations while controlling for relevant economic and demographic factors.

This study evaluates whether more punitive state welfare policies depress rates of voter turnout in communities with high concentrations of welfare recipients. While scholars have studied the influence of demographic group belonging and the effects of state institutional contact (prison, welfare, etc.) on political participation and voting, researchers have not studied how voting behavior shifts in response to policy-level variations in welfare states. To address this gap in the literature, I use data from the Urban Institute's Welfare Rule’s database to generate a novel index of state welfare punitiveness that will be broadly useful for scholars interested in state welfare policies and political socialization. I use multivariate regression analysis to test whether high concentrations of welfare recipients in more punitive states systematically affects voting behavior. Results of this study will be relevant for scholars concerned with voting behavior, political learning, institutional contact, and democratic participation.

This study seeks to understand how to best support low-income, urban employment mobility through transportation. While scholars have studied the effects of urban planning, access to a private vehicle and public transportation for low-income residents, there is no consensus in the academic community for which method of transportation is most effective. Utilizing responses from the 2004 General Social Survey, the study measures perceived access to public transit, car ownership and employment mobility—the ability to switch to an equally desirable job— among low-income residents. Additionally, I employ census data measuring willing job-to-job transfers, car ownership and low-income public transit commuting time as a proxy for transportation access. I expect to find a positive relationship between car ownership and employment mobility, while I predict no relationship between public transportation and mobility. I utilize a multivariate regression analysis for the census data. This study has important implications for transit policy, as it can help inform how to better fund public transit to help low-income individuals and may lead to experimental policy like low-income car subsidies.

Treatments for cystic fibrosis (CF) have extended patients’ lifespan, resulting in CF-related diabetes (CFRD) as a major CF complication affecting 30-50% of adults. A key pathological feature of CFRD is the deposition of islet amyloid polypeptide (IAPP) as amyloid in pancreatic islets. In type 2 diabetes, Islet amyloid is associated with decreased beta-cell mass and function. Current mouse models of CFRD do not develop all the pathological features observed in human CFRD, including islet amyloid. To generate a mouse model of CFRD which exhibits islet amyloid deposition, we crossbred a CF mouse (CftrF508del) and a human IAPP (hIAPP) transgenic mouse and compared amyloid deposition in cultured islets from the resulting offspring. Islets from each of the mice genotypes (NT.Cftrdel/del (n=2), NT.Cftr+/del (n=4), NT.Cftr+/+ (n=5), hIAPP.Cftrdel/del (n=2), hIAPP.Cftr+/del (n=5), hIAPP.Cftr+/+ (n=3)) were isolated and cultured for 7 days in high (16.7 mM) glucose to induce amyloid formation. Islets were fixed in neutral-buffered formalin, paraffin-embedded and sectioned. Sections were stained with thioflavin S (for amyloid) after which amyloid prevalence (% islets with amyloid), amyloid severity (% islet area occupied by amyloid) were quantified. Islets from hIAPP transgenic mice developed islet amyloid when cultured in vitro, while islets from non-transgenic mice did not. Among hIAPP transgenic mice with different CF genotypes (hIAPP.Cftrdel/del, hIAPP.Cftr+/del, hIAPP.Cftr+/+, respectively), islet amyloid prevalence (74±13%, 73±11%, 81±5 %; p=0.89 by ANOVA)[NE2] or severity (3.4 ± 1.7%, 2.0 ± 1.1%, 3.57 ± 0.4%; p=0.64 by ANOVA) did not differ. The presence of amyloid in islets of mice expressing hIAPP genotypes supports the utility of this new mouse model to study this aspect of islet pathology as seen in human CFRD.

Endothelial cells within the pancreatic islet produce key factors for survival and functionality of insulin-secreting β cells. Laminin is one such factor, thought to be responsible for enhancing insulin release. Laminin is a heterotrimeric protein composed of three isoforms: α-chain, β-chain, and γ-chain. Islet endothelial cells from diabetic (db/db) mice show a significant decrease in the expression of laminin α4, β1, and γ1, and show a reduced ability to secrete insulin. To test the correlation between these in vivo observations, we used an immortalized islet endothelial cell-line (MS-1 cells), to decrease genes of interest; and generate conditioned media (CM) in which to culture isolated islets to test their insulin secretion. Islets cultured in CM from MS-1 cells with decreased laminin α4 expression did not have blunted insulin release. However, we found that that decreased laminin α4 increased expression of laminin α1, perhaps as a compensatory mechanism. We inhibited expression of laminin α4 and α1 in MS-1 cells, and cultured islets exposed to this CM also did not have decreased insulin secretion. Therefore, we hypothesize that reducing all three isoforms of laminin in MS-1 cells is necessary to cause a decrease in insulin secretion in islets exposed to this CM. To determine if the knockdown of laminin-411 is sufficient to impair islet insulin release, CM will be collected from MS-1 cells exposed to oligonucleotides designed to inhibit the expression of laminin-411 or a nonspecific control oligonucleotide for 24hrs. After the media is collected, glucose concentrations of the two types of CM are matched. Isolated islets are exposed to these CMs for 48hrs and insulin secretion will be determined for basal and glucose stimulated conditions. We anticipate that CM generated by cells with reduced laminin 411 to decrease insulin secretion in isolated islets, replicating the results observed from db/db mice.