Government Analytics

Permanent URI for this collection

https://jscholarship.library.jhu.edu/handle/1774.2/38219

Browse

Now showing 1 - 20 of 57

Algorithmic Discrimination in the U.S. Justice System: A Quantitative Assessment of Racial and Gender Bias Encoded in the Data Analytics Model of the Correctional Offender Management Profiling for Alternative Sanctions (COMPAS)
(2017-04) Li, Yubin
The fourth-generation risk-need assessment instruments such as Correctional Offender Management Profiling for Alternative Sanctions (COMPAS) have opened the opportunities for the use of big data analytics to assist judicial decision-making across the criminal justice system in U.S. While the COMPAS system becomes increasingly popular in supporting correctional professionals’ judgement on an offender’s risk of committing future crime, little research has been published to investigate the potential systematic bias encoded in the algorithms behind these assessment tools that could possibly work against certain ethnic or gender groups. This paper uses two-sample t-test and ordinary least-square regression model to demonstrate that COMPAS algorithms systemically generates a higher risk score for African-American and male offenders in terms of the risk of failure to appear, risk of recidivism, and risk of violence. Although race was explicitly excluded when the COMPAS algorithms were developed, the results showed that such an analytic model still systematically discriminates against African- American offenders. This paper introduced the importance of examining algorithmic fairness in big data analytic applications and offers the methodology as well as tools to investigate systematic bias encoded in machine leaning algorithms. Additionally, the implications of this paper also suggest that simply removing the protected variable in a big data algorithm could not be sufficient to eliminate the systematic bias that can still affect the protected groups, and that further research is needed for solutions to thoroughly address the algorithmic bias in big data analytics.
American Trust: Confidence in the Ability of the United States Government to Successfully Respond to Acts of Cyberterrorism
(2018-04) Sanders, Michael
Determining the impact of cyberterrorism – a relatively recent construct – on individuals’ trust in the United States government is an area of study lacking longitudinal research. Existing studies provide only topical analysis, leaving room for significant academic research into an emerging topic of substantial concern. This study aims to determine whether American citizens trust the United States government to respond to cyber attacks targeting government entities and public infrastructure, areas likely to be determined to be acts of cyberterrorism. The results from this study indicate that queried individuals’ age, sex, political ideology, education, and employment statuses have a statistically significant impact on the respondents’ overall trust in their government’s ability to adequately respond to cyber attacks targeting both government and public entities and infrastructure.
The Cause of Economic Freedom - A Model of Prosperity
(2020-08) Akers, Philip R.
Economic freedom in the form of open markets and minimal government intervention has been demonstrated in numerous studies to be the most effective approach for achieving wealth and prosperity around the world. This paper examines various sociopolitical and geographic variables as well as changes in their values over time to determine their potential causal impact on the level of economic freedom across a large array of countries. It hypothesizes that democratic governance is of prime influence on free market growth. Having been established as a beneficiary of free markets, Gross Domestic Product (GDP) levels are analyzed for reciprocal causality regarding economic freedom levels, while corruption, civil rights, religiosity, and geography are also explored as potential economic freedom determinants. Of the variables explored, only democratic governance, corruption, and changes to GDP were found to be statistically significant, along with certain regional categories primarily located in Africa. The other variables of interest, initial GDP levels, fundamental rights, and religiosity were not found to be significantly related to economic freedom. This study also found that changes in economic freedom ratings over time were significantly related only to changes in GDP levels, of all the variables studied. The lack of statistical significance found in the initial GDP variable was an unexpected result that potentially calls into question prior research indicating a stronger relationship to economic freedom.
Child Pornography Crimes as Reported to Law Enforcement Agencies
(2019-12) Idle, Megan
Past studies have shown that child pornography is associated with multiple, violent offenses. This article presents a current picture of child pornography incidents as reported to law enforcement agencies. The analysis presents a multinomial logit regression on data from the 1991 to 2016 National Incident-Based Reporting Software (NIBRS) confirming findings that child pornography is associated with violent offenses. However, it also identifies clear differences between child and adult pornography crimes, such as ages for offenders and victims that are younger than expected, an association with gang activity and violent offenses, and suspected computer use by offenders. Many child pornography offenders are minors, rather than older adults; most victims are pre-teens (10 to 14), not young adults; and child pornography offenses are more likely to include victims than adult pornography. The updated information provided in this analysis can assist in the prevention of child exploitation nationwide.
Climate Change Policy Adoption in US States: A Cluster Analysis
(2020-05) Woods, Melissa
While the United States steps back from taking action on climate change at the federal level, many are looking to the states for leadership. Existing literature examining enactment of climate legislation in the United States often aims to identify causal mechanisms or quantify relationships driving the enactment of specific policies like renewable portfolio standards. This study takes a different approach by using clustering algorithms to identify commonalities among states in terms of: their legislative behavior; impacts of climate change; political, economic, and demographic characteristics; and energy generation and emissions. The clusters identify nuanced groups such as politically conservative coal states encouraging carbon sequestration, politically liberal states with low emissions and low policy adoption rates, and states generating large portions of their energy from natural gas also proposing renewable energy legislation. These insights highlight opportunities for future research and potential strategies for political advocacy organizations working to enact climate legislation.
Compact Growth and Smart City Development: The Unsustainability of Urban Sprawl
(2019-12) Pacilli, Dean
As cities become more populated, are municipal governments struggling to manage the influx of demand in the realm of power usage and city resources regarding critical infrastructure and energy management? Contemporary research centers on a prevailing view of the benefits of the compact growth smart city model over the outdated urban sprawl model regarding energy and environmental sustainability. In this study, data was collected and analyzed to provide meaningful insights for city planners and government leaders to weigh the costs and benefits of adopting the compact growth model of urban development in lieu of the urban sprawl model of expansion. Through collecting municipal data of high population cities in America, regressions were run to examine how density has an effect upon various factors supporting a city’s operational efficiency. The results of this study indicate that the urban sprawl model, popularized during the 20th century, is less sustainable than the compact growth model of smart city development, especially with increasing populations. Furthermore, future urban development plans can employ the compact growth model of smart city development to maximize infrastructure and improve energy efficiency. This paper will delve into the contention that urban areas will need to adopt the compact growth model for sustainable operations.
Count MENA: Clarifying the Impact of U.S. Hate Crimes on the Middle Eastern & North African Community
(2021-12) Zakhari, Nivine K.
Frequent racial, ethnic, and/or religious miscategorization of Arabs, Middle Easterners and North Africans (MENA), Muslims, and Sikhs has downstream impacts in general census reporting and hate crime statistics, among other government programs relying on accurate racial/ethnic or religious population data. This research analyzes U.S. hate crime statistics since 1991, along with the latest U.S. Census and religious population estimates, to determine the impact of inconsistent victim identification and their approximate per capita hate crime victimization rates. Analysis shows it is still highly likely that the MENA population in the U.S. is significantly undercounted, as well as the incidents of hate crimes targeting persons of MENA descent and/or Muslims or Sikhs, which are often conflated. Unless government race and ethnicity reporting standards are updated, along with enforcement of Uniform Crime Reporting standards to improve hate crime motive accuracy, efforts to address challenges impacting the MENA community will continue to fall short.
Democracy and Economic Growth: A Social Level Measurement of Growth
(2019-08) Garcia, Lina V.
There are contradictory findings in the literature about the effect democracy has on economic growth. In the current literature, the research indicates that democracy positively influences economic growth directly as it ensures property rights and improves business environment through advanced technological innovation. Yet, there is substantial research that states that there is no evidence that democracy has a considerable effect on economic growth. This article examines individual/ country specific indicators such as Human Rights Protection Score, Political Regime Score, Child Mortality Rates, and Years of Schooling and the effects this has on economic growth, measured by GDP per capita. This article utilizes a pooled time series cross sectional analysis to analyze the said variables from 1985-2014 in 145 countries. The existing literature analyzes hard indicators such as property rights and technological innovation to measure economic growth. This article takes a different approach as it aims to research a more social aspect of countries and how this effects GDP per capita. This study finds that years of schooling is the most significant variable affecting GDP per capita.
Determinants of Forced Migration: The Varying Effects of Violence and Economic Conditions on Syrian Refugee Flight
(2016-12) Byrne, Maureen
The majority of existing research on the impact of civil war on forced migration flows uses observations at the annual-global level, limiting the applicability of results to explaining aggregate trends. This article studies refugee flows from Syria to Jordan from 2012-2015 at the weekly level to test how violence and economic conditions affect the fluctuation of migration processes. The results of regression analysis offer support for the argument that all violence does not affect migration decisions uniformly; rather some types of violence produce higher migration flows, while others, such as chemical warfare, render conditions too unsafe to flee. Furthermore, while economic conditions in the origin country affect migration flows, conditions in the destination country do not, suggesting that economic opportunities outside the country are less consequential as a determinant of forced migration during civil war. The research demonstrates the importance of using data at low levels of temporal aggregation to uncover causal mechanisms underlying refugee flight.
Determinants of Military Expenditure for Deterrence: Interaction with Nuclear Weapons, Democracies, and Alliances
(2016-12) Kim, Kenneth
During his campaign trail, the President-elect announced increasing military expenditure to enable deterrence to maintain the security of the United States. However, limited quantitative research exists to determine if military expenditures effectively support deterrence. Moreover, current research does not estimate if possession of nuclear weapons, a democratic form of government, and alliances contribute to prohibiting militarized conflicts through deterrence. This article examines the determinants of military expenditures and its effects on increased deterrence by analyzing ten countries engaged in deterrence. A regression of military expenditures as the dependent variable establishes a baseline to identify predicted values based on the theories of nuclear deterrence, democracy peace, and free riding of defense burden within alliances. The regression further explores interactions between possession of nuclear weapons and democracies, between democracies and alliances, and between possession of weapons and alliances. Limited results convey that deterrence can be increased with increased military expenditure but at a cost.
Do Media Companies Drive Bias? Using Sentiment Analysis to Measure Media Bias in Newspaper Tweets
(2018-04) Thomsen, Taylor
As natural language processing tools are advancing in their use to study corpuses, it is important to harness these tools to study media content for potential media bias. Existing studies have attempted to measure political sentiment by using lexicons to determine whether political texts have a positive or negative connotation and other studies have looked at media bias by manually classifying media content. Although machine learning models are widely used to simulate human decisions about sentiment, few studies have used automated sentiment analysis on newspaper content to measure media bias. This paper uses an out-of-the-box sentiment analysis model on several newspapers’ tweets from four major media companies during the month leading up to the 2016 presidential election. The sentiment analysis results in a sentiment score for each tweet mentioning Republican or Democratic keywords. Overall, this paper finds that only some newspapers had significant differences in the sentiment scores for Republican and Democratic tweets. Additionally, Republican and Democratic sentiment scores were not significantly different between media companies, showing that the companies that own the newspapers may not be driving biased content of the individual newspapers.
Does Healthcare Availability Impact the Veteran Suicide Rate?
(2020-05) Diaz, Fredy
Past research on a recent veteran suicide spike phenomenon has focused on Post-9/11 combat veterans suffering Post-Traumatic Stress Disorder (PTSD) and Traumatic Brain Injury (TBI) from combat in Iraq and Afghanistan. This article presents a different approach by looking at the importance of healthcare accessibility – measured through state uninsured rates – as a possible way to reduce the veteran suicide rate. The Veterans Affairs (VA) only provides healthcare services to honorably discharged veterans, leaving those veterans who received an administrative or punitive military discharge – commonly referred to as “bad paper veterans” – to turn elsewhere for mental health services. This analysis uses VA suicide data covering the 50 states and the District of Columbia from 2011 to 2017 to determine whether post-Affordable Care Act state actions to lower the uninsured rate have impacted veteran suicides. The analysis offers strong support that the uninsured rate explains and influences the veteran suicide rate.
Does Immigration Cause Crime? Evidence from the United States
(2018-08) Zhang, Mengyi
Donald Trump announced that immigration should be responsible for higher crime incidents in the United States, and he in turn aimed to strengthen his anti-immigrant policies. Recently, his entire anti-immigrant agenda has triggered debates all over the United States. There are not too many previous studies focusing on empirical evidence, and they have never reached a consensus. This paper investigates the relationship between three kinds of immigration and crime in different regional groups to provide an updated assessment, including unauthorized immigrants, foreign population and Mexican unauthorized immigrants. State level cross-sectional data in 2014 is analyzed using multivariate regression. The results of the regression analysis reveal that immigration has significantly positive effects on violent crime. Compared with foreign population, the influence of unauthorized immigration appears to be stronger. Compared with foreign population and Mexican unauthorized immigrants, the influence of unauthorized immigration appears to be stronger. Contrary to inland state group, the study reveals that immigration accounts for crime. The evidence in fact shows that poverty rate increases the amount of violent crime and crime rate significantly. In the end, the findings provide important implications for the concerned authorities and policymakers.
Does Teacher Multicultural Training Matter? Evidence from the U.S. Scores on Pisa 2015
(2018-10) Audisio, Ariana
The racial and ethnic achievement gap in the USA ranges from about 3 to 5 years of schooling. Existing research about diversity in education has focused on studying the effect of matching students with teachers of the same culture or gender, and the different mechanisms that mediate those effects, including bias, stereotypes and prejudices. Multicultural competence training has been an essential part of work training and professional development in the military and business industries. In the field of teacher education, it has also been discussed and incorporated to some extent but there is no research about the effectiveness of the training and its impact on student performance. This study takes opportunity of the new teacher questionnaire from PISA 2015, a large-scale international assessment study administered every 3 years to 15-year-old children around the world to measure their performance in reading, math and science. Using multilevel analysis, the effect of multicultural training on the student’s performance in science was measured for immigrant and native students in the USA. The results show that multicultural training has no significant effect on student performance in science, even after accounting for the economic, social and cultural status of students and schools, and common indicators of teacher quality.
Does US Foreign Aid Foster Innovation in Recipient Countries?
(2018-12) Boampong, Angelina H.
Previous works studied the impact of foreign aid on economic growth with mixed results. US foreign aid promotes peace, security, and economic development, and provides humanitarian relief. These factors also spur innovation so it is assumed that foreign aid also fuels innovation. This study examines if US foreign aid impacts innovation in 146 recipient countries from 2011 to 2017. The study leverages fixed effects and ordinary least square regressions to assess the relationship between US foreign aid and innovation, while controlling for GDP, population size, and geographical region. The results of the analysis show that US foreign aid has a negative impact on innovation input factors for countries in Europe and Eurasia. GDP and population size were found to have measurable impacts on innovation under various scenarios. The results will provide useful insight in future considerations about how and when the US provides aid to other countries.
The Economic Integration of Immigrants: A Comparative Analysis of U.S. States
(2018-05) Parlej, Patrick K.
This paper measures the economic integration of immigrant populations in the United States by comparing incomes of native-born and foreign-born individuals across states in years 2005-2016. Whereas previous research on this topic primarily focused on immigrant economic and demographic traits in national or regional studies, this paper compares aggregate data at the state level that incorporates the economic and demographic characteristics of native-born Americans. To that end, this paper proposes the ‘integration ratio’, an economic metric used to compare earnings between immigrant and native median household incomes. Furthermore, this paper highlights the relative success of southern states, which overperform on measures of economic integration relative to other regions in the United States. Results from OLS model give insights into the conditions under which immigrants most effectively economically integrate, and thereby are of interest to policymakers and newly arrived immigrants seeking circumstances most conducive to immigrant integration.
Equivalency Analysis of State-Level Air Pollution Emissions Estimates from an Integrated Assessment Model: GCAM-USA
(2019-08) Besch, Brianna
The use of Integrated Assessment Models (IAMs) has expanded from their primary application, global climate assessment, to examine broader environmental trends such as air pollution emissions, including those at sub-national levels. While most IAMs are evaluated for internal validity, little literature evaluates IAM performance against real-world data. In this paper, 2015 estimates from the Global Climate Assessment Model (GCAM), an IAM, of state-level air pollution from five macro-economic sectors and six air pollutants, are tested against EPA data using two equivalency analysis tests. One-sided t-tests and regression analysis determine if modeled emissions fall within 20% of U.S. Environmental Protection Agency data. The results are mixed, showing that GCAM emissions estimates from only one sector and one pollutant pass both tests, and two sectors and two pollutants pass neither test. This indicates that GCAM-USA is more appropriate for examining national trends than specific sectors or pollutants at the sub-national level.
Evaluating Primary Care Same Day Access as a Method to Reduce Avoidable Hospitalizations Related to Ambulatory Care Sensitive Conditions
(2018-04) Fay, Jason
Current evidence in support of measures to reduce hospital admissions associated with Ambulatory Care Sensitive Conditions (ACSC) is extensive and represents decades of global research. My study improves this by examining primary care access from the perspective of same day care. This aggregate analysis emphasizes the importance of primary care access as the primary intervention to reduce avoidable ACSC admissions. This study expands on existing research by being the first to examine the relation-ship between same day accesses with their provider and ACSC admissions through a cross-sectional analysis of 140 Veteran A airs hospitals during fiscal year 2017. The secondary purpose is to evaluate contemporary recommendations on maximizing PC access to develop a comprehensive tool-kit for VHA administrators to reduce ACSC admissions, decrease costs and improve quality of care. The results and evidence presented in this article indicate a significant inverse relationship exists between same day access and ACSC admissions.
A Flaw in the Federal Adult Education Funding Formula: Academic Gains and English Language Learners
(2017-08) Judd, Amy Elizabeth
Existing research in the field of adult basic education diverges between transformative learning theory, which focuses on how student growth is accomplished in the classroom, and human capital theory, where adult education is viewed as an investment in a strong economy. US adult education programs are funded by federal legislation intended to create a strong workforce. Grant-funded programs are charged with serving adults deficient in the basic skills of literacy and numeracy or lacking a high school diploma, as well as adults who are limited in English language proficiency. Yet, the federal funding formula for state-level allocations has been historically based on Census calculations of adults lacking a high school diploma, and ignores those who do not speak English well, despite the expectation for grant-funded programs to serve both populations. In states serving large numbers of adults with limited English proficiency, programs struggle to meet performance requirements for student educational gain. The findings in this paper indicate that a better-aligned funding formula could more equitably distribute adult academic gains in each state, a matter of importance as US policymakers struggle to find common ground on issues of immigration, diversity, and upskilling the American workforce.
Fraccidents: The Ostensible Link Between Oil and Gas Development and Car Accidents in Post-Shale Boom Texas Counties
(2017-12) Benitez, Richard
New developments in the exploration of unconventional Oil and Gas (O&G) resources have become an energy industry trend. Nevertheless, there are underlying environmental and socio-economic impacts attributed to “fracking”, which could strain public resources and local transportation networks. From 2012 to 2016, a quarter-fold increase in the number of producing wells occurred simultaneously with a 29 percent increase in total car crashes as well. This study explores the purported link between O&G development and car crashes using Texas county-level panel data. Despite this anticipated relationship, the results presented a tenuous relationship between the two. However, the overall magnitude of the relationship between O&G development and car crashes improves upon the existing literature. The research also maintains that there is still an emerging opportunity for future scholarship and policy formulation regarding O&G development and its residual impacts on public safety and quality of life among other communities of interest.

Browse

Browsing Government Analytics by Title

Results Per Page

Sort Options