2017 CSS Fellows
Project: Improving Refugee Integration Through Data-Driven Algorithmic Assignment
Department: Political Science
Abstract: The world is currently experiencing the worst refugee crisis since World War II. The U.S. has long run the world’s largest refugee resettlement program, yet little empirical evidence exists on the efficacy of different approaches to integrating refugees within American society. Currently, the assignment of refugees across domestic resettlement locations is based primarily on local office capacity, and does not leverage the rich insights that historical data can provide regarding which refugees do better in which locations. To address this issue, we will develop a data-driven assignment algorithm that combines supervised machine learning and optimal matching to assign cohorts of refugee arrivals across resettlement locations in such a way that maximizes refugees’ employment outcomes. The algorithm will achieve this by identifying and leveraging synergies between refugee characteristics (such as age, gender, language skills, etc.) and resettlement locations, and will be tested via randomized controlled trials with partnering resettlement agencies.
Nano Barahona, Josh Kim, and Sebastian Otero
Project: Food Labeling and Information: Effects on Supply and Demand of Nutritional Content
Abstract: This project studies how the introduction of a nation-wide regulatory label based on the nutritional content of food products sold in supermarkets affects both the bundle of products demanded by consumers as well as the nutritional composition of food offered by companies. By looking at both demand and supply responses, this project studies the general equilibrium effects of information-based regulations intended to reduce obesity, contributing to our understanding of both public health and economic behavior. To do so, we take advantage of a unique, large and highly-detailed dataset containing all transactions and purchases for any food item sold in each retail store belonging to Walmart in Chile during the last 4 years. We are also collecting an additional dataset detailing the nutritional composition of each listed food item. By combining reduced form analysis techniques and a general equilibrium model of demand and supply for nutritional content in food we quantify the causal impact of the regulation and understand its overall welfare consequences.
Project: Scientific Models as Guides for Disciplined Perception
Department: Graduate School of Education
Abstract: Scientific data on relevant issues for our societies – global warming, vaccination, evolution theory, gene manipulation – is too complex and messy to provide clear and crisp stories that everyone agrees on. Such data is prone to cherry picking, where people can easily see what they expect to see and ignore the rest. Thus, an essential goal of science education is to teach students to be “critical observers,” i.e. participants of scientific conversations that use evidence from real-world data instead of ideologies or subjective beliefs to critically review and evaluate scientific claims. Little research exists, however, on how to best structure student experiences to teach the necessary skills and mindsets of critical observation. To study this, I will conduct two studies that address the following research questions: 1. Does the explicit comparison of scientific models with data improve participants’ ability to perceive meaningful, yet subtle patterns in rich, unstructured real-world data? 2. Compared to situations where only the target model is given, does the presence of competing models influence what evidence participants seek in the real-world data, and how they evaluate the target model?
Project: Lexicon-Based Factuality Classification of Media Consumption Data
Abstract: Factuality, defined as the degree of certainty about the factual status of a situation or an event, is one of the fundamental factors in several applied research fields such as machine translation and claim detection. This project will develop a lexicon-based method for factuality classification based on the theories and literature in computational linguistics. Previous work on factuality and other related concepts, such as modality, evidentiality, and uncertainty, have focused mainly on the sentence level and well-formed language data. Text analysis at the sentence level cannot cope with non-standard language data containing tokens that do not form sentences. In focusing on the word-level lexical analysis, this project represents one of the first attempts to deal with non-standard language use, such as the language used in texting and social media platforms, in factuality research.
Caue Dobbin and Tom Zohar
Project: Determinants of Inequality and Economic Opportunity Using Administrative Databases from Israel
Abstract: Equality of opportunities is a topic of widespread interest, central to both academic and political discussions. Economic studies have traditionally estimated the intergenerational elasticity (IGE) in income. However, income based measures can greatly underestimate the intergenerational persistence of opportunities. Individuals from wealthy families often pursue careers which do not necessarily maximize their earnings (e.g. teachers, artists and academics), while those from less advantageous backgrounds might not have this choice. Therefore, ignoring the non-monetary component of jobs is misleading. This projects aims to bridge this gap. For this purpose, we build on a recent innovation proposed in Sorkin (2017), in which a revealed preference method is used to estimate the total value of each job, including both monetary (wage) and non-monetary (amenities) components. Combined with linked parent-child data on labor market outcomes, it allows us to estimate the intergenerational persistence of utility, i.e. the degree to which family background determines one’s wellbeing.
Project: Using Virtual Reality to Understand Prejudice and to Reduce Implicit Biases
Abstract: Understanding prejudice and the full extent of its consequences is difficult. Unless it is personally experienced, some people do not understand what it is like to be discriminated against based on their race or socioeconomic status. Others don’t believe it occurs simply because it doesn’t happen to them or because they’re not prejudiced themselves. To address this issue, researchers have begun to integrate virtual reality (VR) and perspective-taking (i.e. imagining what it is like to be someone else under specific circumstances), allowing users to viscerally experience what it is actually like to be someone else. In the past, VR experiences where users were able to take on the perspective of others (e.g. the elderly, schizophrenics, or the colorblind) reduced prejudice and increased empathy, understanding, and helpful behaviors. The proposed study will examine the effectiveness of VR perspective-taking at reducing implicit biases and prejudice toward black people and members of stigmatized groups as well as the effect that different immersive features have on attitudinal and behavioral change inside and outside of VR.
Project: Understanding the Global Gender Gap in STEM Engagement with Agent-Based Simulation and Online Social-Psychological Interventions
Department: Graduate School of Education
Abstract: Despite educators and researchers’ concerted efforts, a gender gap in science, technology, engineering, and mathematics (STEM) fields continues to persist across the globe. Overall, girls take fewer STEM courses relative to boys, and women acquire fewer Ph.D. degrees and occupy fewer tenure-track positions in STEM higher education institutions than men. Previous studies have extensively examined how micro-level factors (such as personal traits and classroom cues) affect girls’ and boys’, women’s and men’s STEM preference, yet little is known about how macro-level, sociocultural contexts (such as individualistic or collectivistic cultures) shape preference development, especially for large populations over time. Using big data, agent-based simulation and mass online experimenting platforms, this research brings a sociocultural perspective in understanding causes of the gender gap in STEM engagement and challenges previous biology- and trait-based explanations.
Chloe Lim and Dan Thompson
Project: Covering Congress: The Impact of a Congressperson's Roles on Her Coverage
Department: Political Science
Abstract: Countless social scientists have spent their careers theorizing about and empirically teasing out the returns members of Congress receive for their efforts at policy making. These returns come in the form of pet projects, campaign donations, and future income. But, these payments are partly a means to an end—they allow members to raise their profile at home. Does this work? Do members who serve in more important roles receive more media attention? Our project marries techniques from computational social science with traditional research designs for drawing causal inferences to answer this question directly.
Project: Estimating the Determinants of Trust Using Online Field Experiments
Abstract: The question of what makes people trust each other has renewed its importance in the era of the sharing economy. A number of recent research studies have shown that despite claims that the sharing economy can promote trust between individuals in society, the fact is anti-social behaviors such as discrimination are still widely experienced by certain groups of individuals on platforms like Airbnb and Uber. These revelations have revitalized long-standing research questions on trust. Namely, what causes feelings of trust and trustworthiness between individuals –particularly strangers –and what sort of structures can be implemented to engender a greater sense of generalized trust? The goal of this project is to use computationally-oriented experimental methods that can leverage new sources of data available online by conducting a set of online field experiments using subjects recruited from a large population of users both from within the sharing economy as well as the general population.
Project: Attributions of Responsibility in Non-Democratic Political Systems
Department: Political Science
Abstract: Conducting public opinion research in non-democratic political systems can be difficult, because surveys often need to be censored and respondents may fear the consequences of providing honest answers. Recruiting respondents into online surveys with Facebook ads offers one potential strategy for addressing these issues. This study will use Facebook ads to implement a survey with embedded experiments exploring how individuals in non-democracies attribute responsibility for governance outcomes across different political actors.
2016 CSS Fellows
Project: Understanding movement patterns of internally displaced persons in the Democratic Republic of the Congo
Advisers: Ran Abramitzky and Shripad Tuljapurkar
More than 65 million people worldwide are currently displaced from their homes by conflict, the highest recorded number since World War II. Humanitarian organizations focused on these emergency contexts rarely prioritize data collection, leading to a significant gap in the academic and policy understanding of forced migration. This project, based in the Democratic Republic of the Congo, will investigate the socioeconomic impacts of displacement and whether a machine-learning algorithm linking records across multiple databases could identify duplicate entries to more efficiently target assistance. The project is co-authored by Toly Rinberg and Andrew Bergman at Harvard.
Project: Planting passwords through online games
Department: Computer Science
Advisers: Bahman Bahmani, Russell Poldrack, and David Mazieres
Authentication is one of the major problems faced by society in interacting with cyber technology. Passwords, challenge questions, out-of-band text messages, and physiological biometrics create friction with user experience, and yet are increasingly bypassed by hackers.
Project: Bureaucratic mobility in China and India
Advisers: Xueguang Zhou and Matt Jackson
How does sponsorship or specialized experience facilitate or impede mobility in modern bureaucracies? How do the historical and political contexts of a bureaucracy affect its mechanisms of mobility? This research project will provide a systematic comparison between two of the largest bureaucracies from authoritarian and democratic regimes using field and archival research in China and India.
Project: Learning from startup failure
Advisers: Karen Cook and Dan McFarland
Learning from startup failure can help to avoid the same mistakes, to reduce risk, and to foster innovation, but the mechanisms for learning from failure are little understood. This research project takes advantage of a recent phenomenon – the online startup postmortem – to address the question of how people and organizations learn from startup failure.
Project: The nature of science misinformation in social media
Department: Graduate School of Education
Advisers: Jonathan Osborne and Sylvia Bereknyei Merrell
Contemporary web technology has allowed the public to do their own research easily and become “informed” about issues of concern. Some websites are professionally authored by experts in science, but a substantial portion of science-‐related web content is now published by Internet users who are non-‐experts and may contain invalid claims or inaccurate interpretations of science. This project will examine tweets about two recent controversial science issues with implications for personal and public health — the consumption of genetically modified organisms (GMO’s) as food, and vaccination safety -- to highlight the prevalence of scientific misinterpretations in this data.
Edgar Franco Vivanco
Project: The dynamics of crime, police violence, and corruption: A text analysis of citizens' anonymous crime reports in Rio de Janeiro
Department: Political Science
Advisers: Beatriz Magaloni and Justin Grimmer
Nine out of ten violent deaths in the world occur outside countries involved in direct conflict. 41 of the 50 most dangerous are in Latin America, the most violent region in the world. As a consequence of violence, around one million people have died in the past decade in this region. Despite its pervasiveness, our understanding of criminal violence is still very limited. This project will analyze a large dataset of anonymous calls reporting criminal activity in Rio de Janeiro, Brazil, combined with official governmental statistics on criminal activity. By implementing machine learning procedures and textual analysis models, it will offer insights into how the society reacts to crime, violence and corruption. The final objective of this project is to provide advice to researchers, policy makers and police departments on how to reduce violence and build better relations with the society.
2015 CSS Fellows
Project: Social Psychological Causes of Global Inequalities in MOOCs and Beyond
Advisers: Jeremey Bailenson, Geoffrey Cohen
It has been claimed that Massive Open Online Courses (MOOCs) have the potential to help bridge the education gap in developing countries. However, many students from developing nations are underachieving in MOOCS. This project proposes that this is due to identity threat (academic achievement gaps attributed to uncertainty about social belonging due to race, gender, or social status). It then addresses two questions: why do members of developing countries experience identity threat in international learning settings and are members of developing countries disadvantaged due to bias and identity threat in other high-stakes contexts?
Project: Continuous Multimodal Emotion Inference
Advisers: Noah Goodman, Jamil Zaki, Chris Potts
This project will build a model for performing continuous multimodal emotion inference from linguistic, audio, visual and psychophysiological (e.g., skin conductance, heart rate) cues. This model will help researchers understand how humans understand changes in the emotions of others, how emotion inference is affected in mental disorders, and will aid in building emotionally aware computational agents in the field of artificial intelligence.
Project: Understanding Social Dynamics in the ‘Gig’ Economy
Department: Management Science & Engineering
Advisers: Michael Bernstein, Melissa Valentine
The online labor market is a new economic labor system where people work as contractors using online platforms like ODesk, Amazon Mechanical Turk, and TaskRabbit to connect to potential employers. This project analyzes the problems that occur in a system where many of the employers have little to no experience in designing, defining, and managing work, and there is little oversight of the contractors in terms of qualifications and commitment they have to working on projects.
Project: Uncovering Political History using the Complete California Voter Records, 1900 - 1968
Department: Political Science
Advisers: Shanto Iyengar, Simon Jackman
Beginning in 1867, the state of California required all eligible California voters to register to vote. This created a database with the occupation, name, address, and party affiliation of every California voter. This project uses computational methods to explore a period of extraordinary enfranchisement and change in America by analyzing trends in voting records related to women gaining the right to vote, the Great Depression, and employment shifts.
Project: How Public Opinion Shapes Policy Outcomes in China
Department: Political Science
Advisers: Justin Grimmer, Jennifer Pan
To prevent rebellions and coups, autocrats wish to implement policies that the majority of people are in favor of. Because autocratic regimes have trouble understanding the public’s true opinion, this project uses computational methods to examine how governments strategically release feelers of policy ideas to the public, monitor discussion about the feelers on social media, and choose to confirm, deny, or adjust the policies based on the public discussion.
2014 CSS Fellows
Project: Using computational linguistics to explain organizational change
Advisors: Woody Powell (sociology), Mark Granovetter (sociology), Dan Jurafsky (linguistics)
These projects will use computational linguistics to study how ideas about organizing the nonprofit and public sector vary over time and across places. The first project is a mixed-methods investigation into the fact that nonprofit organizations are increasingly managed by professionals, who tend to make use of tools that are common in the for-profit sector. The first problem that will be addressed is the facts that the way nonprofits are categorized by the IRS is inconsistent. The second part of the project is to determine to what degree an organization uses for-profit practices without relying on expensive survey or interview data. It will compare web pages to the language of social movements, for-profit, and public organizations using natural language processing. The goal is to create a stand-alone computational model that classifies nonprofits based on characteristics that are not easily accessible otherwise.
The second project is an investigation into how city administrations describe the goals and activities of urban development. I will examine the diffusion of the idea that urban development has to be environmentally sustainable. The goal is to determine the degree to which global climate change and sustainability play a role on the web pages of urban planning departments. The computed probability that a city’s webpage reflects the sustainability paradigm serves as a unique dependent variable in explaining the diffusion of ideas. The idea is that cities with stronger ties to international organizations and INGOs are more likely to bring up sustainability in their urban development discourse. This work will help to explain how abstract ideas travel across the world and become implemented locally.
Project: The thin blue waveform: coding police officers' emotions from body camera audio
Advisors: Jennifer Eberhardt (psychology), Dan Jurafsky (linguistics)
Interactions between police officers and members of the public can be highly charged and immensely consequential for both officer and civilian: the officer must gauge the potential guilt of the suspect, while the suspect risks being arrested or jailed. Given the legal and physical authority officers wield in these interactions, it is critical to identify situations under which race may influence officers’ judgments and actions. Past psychological research on race and policing has focused on racial bias in laboratory settings. Researchers, for example, have examined which faces capture the attention of officers when they are prompted to think of violent crime or how stereotypes may influence officers’ decisions to shoot unarmed Black versus White suspects in a computer game. However, most civilian interactions with officers resemble interviews rather than arcade shooting ranges. Research on racial bias has revealed that bias can be conveyed in subtle ways, independent of (or even in spite of) egalitarian intentions. For instance, even when officers say the same things to Black and White suspects, their tone may send different messages. The proposed research utilizes footage from Oakland Police Department (OPD) body cameras to test if prosodic vocal cues reveal differential treatment of Black and White suspects in real-world settings.
Emanuele Colonnelli & Francisco Munoz
Project: Trust and firms' growth: Evidence from Brazilian corruption reports
Advisors: Tim Bresnahan (economics), Guido Imbens (Graduate School of Business), Dan Jurafsky (linguistics)
Corruption has recently been at the center of the policy attention, as is shown by the recent anti-corruption movements in India and the passing of a new Brazilian anti-corruption law at the beginning of 2014. This project aims to better understand the economic and financial effects of corruption in the context of one of the world’s most promising emerging economies: Brazil. It looks at a natural experiment that generates variation in the level of banks’ and firms’ information, and exploits potential heterogeneity in the presumed amount and quality of information these agents possess. The focus of this project is on firms’ innovation activities and creation of new businesses, and in understanding whether a potential decline in lending causes financial distress.
Project: Physiological response to robotic and animated agents in physical and virtual reality
Advisors: Wendy Ju (mechanical engineering), Jeremy Bailenson (communication)
This project investigates people’s emotional response to physical and digital agents in real-world and virtual-world environments. The findings from this study can help clarify how people experience physical interactions compared to virtual ones. We aim to use computational techniques in order to identify and classify features of a person’s physiological response (e.g., skin conductance and heart rate) to social robots and virtual agents that play the role of lecturers in both physical and virtual reality. These results have implications to theories of mixed reality, human-agent interaction design and the use of agents as opposed to instructor videos in the delivery of online educational content.
Erik Peterson & Jonathan Mummolo
Project: How local news shapes evaluations of politician performance
Department: Political Science
Democratic accountability requires voters to evaluate elected officials based on their performance in office, however, media reports may cause voters to obtain an unrepresentative view of what political figures have accomplished. If incumbents set the terms of coverage, the media may serve their interests rather than those of voters.
This project will use computational methods to examine the media's role in performance voting in gubernatorial elections across several issue domains: the economy, crime and education. Using a dataset which combines gubernatorial press releases and text from local newspaper articles spanning roughly a decade, this project will examine how the interplay between governors and journalists produces news and, in turn, how this coverage alters public opinion.
Advisors: Shanto Iyengar (communication), Justin Grimmer (political science)
Project: Expert crowdsourcing with flash teams
Department: Management Science & Engineering
Advisors: Pamela Hinds (MS&E), Melissa Valentine (MS&E), Michael Bernstein (computer science)
Over the last decade, web-mediated marketplaces have emerged and enabled a new type of computational labor economy that has broadened crowdsourcing from volunteerism (e.g., Wikipedia) to paid labor. To date, most crowdsourcing platforms, such as Amazon Mechanical Turk, are designed to recruit amateurs to accomplish microtasks that require little, such as document editing and translation. Current teams of crowd experts are neither effective nor efficient at completing complex interdependent tasks due to a lack of generalizable techniques for guiding expert crowds. Overcoming these challenges requires a multidisciplinary approach with contributions from both (1) behavioral scientists, such as methods and theories from organizational behavior, who understand how to manage team structures, how people are motivated and the way work is organized; and (2) computer scientists, who understand and can develop the necessary technology to investigate and extend our capability in this area. This project aims to fill this critical gap by building on research in organizational theory and computer science to develop a theoretical model and authoring platform designed to support the creation, coordination, and effectiveness of teams of interdependent expert crowds, referred to as flash teams.
Project: Fighting labor market inequality: a social network perspective
Advisors: Matt Jackson (economics), Amir Goldberg (Graduate School of Business), Mark Granovetter (sociology)
This project is one of the first empirical studies on social network and labor market outcomes. To date, researchers’ understanding of the topic has been severely limited by data availability. With a unique dataset from the Chinese internet giant Tencent, consisting of 300 million users (i.e. over 90 percent of the Chinese internet population) and 1.5 billion social interactions, this is the first time that any researchers have come close to understanding an entire population’s dynamics. Specifically, it affords us the opportunity to measure precisely how social network’s impact on labor market outcome depends upon various contexts. This project will be the first step of a long-term project whose ultimate objective is to generate network-topology based labor market policies.
2013 CSS Fellows
Project: The New Globalization of Science
Advisers: Dan McFarland (Education); Paolo Parigi (Sociology)
At no other point in history has science been as much of a “global enterprise” as it is today, characterized by the dramatic increase in the mobility of scientists between countries. This “global science” is the driver behind solutions to the most daunting challenges of the 21st century. The resources, knowledge, and expertise necessary for these pursuits are embodied by the scientists scattered across the world. To that end, my broader research agenda is to better understand how modern science is shaped by globalization, specifically the global mobility of scientists. I will explore how mobility shapes scientists in four ways by examining their (1) careers, (2) collaboration patterns, (3) citation patterns, and (4) research foci in an emerging order of global science. I will conduct a comprehensive study using data from Thomas Reuters’s Web of Science, which curates nearly every scholarly publication. I will use tools from data mining, social network analysis, and the emerging field of computational linguistics to study how different degrees of mobility affect scientists, with policy implications for higher education both in the U.S. and worldwide.
Project: Learning Analytics for Smarter Psychological Interventions
Advisers: Carol Dweck (Psychology); Roy Pea (Education)
Online college courses fail to meet their full potential when they fail to support students who underperform or drop out due to a fixed mindset - the mistaken and self-limiting belief that their intelligence is unchangeable. Psychologists have developed short interventions to both influence mindset and improve learning outcomes, but they lack detailed evidence of the specific behaviors that bridge the gap between the two. I propose to use tools from learning analytics to identify the student behaviors that link self-reported mindset with learning outcomes. I will use my results to implement smarter psychological interventions that can deliver a message to the right student at the right time.
Project: Computational Field Experiments
Advisers: Dan McFarland (education); Sharique Hasan (Business)
Our research will investigate the tradeoff between innovation and coordination in order to better understand optimal venture design, especially in the context of developing economies. To do so, we will be conducting a novel field experiment by running an entrepreneurial incubator, Innovate Delhi, with Indraprastha Institute of Information Technology in New Delhi (IIIT-D). We hope to move beyond the standard field experiment and conduct the first computational field experiment.
Project: Media Day: Using Computational Techniques to Study Emotion Regulation Through Media Choices Over an Entire Day
Advisers: Byron Reeves (Communication); James Gross (Psychology); Paulo Blikstein (Education)
Understanding the extent to which users actively process and proactively control sequence and pace of media choices has important implications for theories and design in communication, human-computer interaction, and psychology and is the basis of this research. I will use computational techniques to 1) create a database linking emotion (e.g. skin conductance) to corresponding screen content over the course of a day; 2) utilize machine learning to detect and classify salient features of screenshots (e.g. presence of faces, luminance, text vs. images); and 3) to test the theories “in the field” by analyzing the effects of precise changes in screen experience on arousal and mood over the course of a day.
2012 CSS Fellows
Project: Benevolent Rejection
Department: Management Science and Engineering
Advisers: Kathy Eisenhardt, Riitta Katila, Steve Barley (Management Science and Engineering,) Amir Goldberg and Sharique Hasan (Graduate School of Business)
Abstract: With the advent of the Internet, organizations gain access to large and diverse audiences that greatly expand the pool of outsiders it interacts with and the ideas it sources from them. This has activated a latent, unresolved issue: the organization cannot act on most of these ideas. Inevitably, organizations either ignore or reject most of the ideas. This leads to my research question: How do users react to being rejected. Organizations need a way to "benevolently reject" suggestions they receive - that is, a way to reject the suggestion without rejecting its author and endangering the relationship. To examine this question I have built a dataset that includes 24,067 organizations' efforts to collect user 702,729 suggestions between November 2007 and June 2011. My research contributes to a better understanding of the effect rejections have on the evolution of the dyadic tie between the organization and the outside actor.
Project: Applying Network Analysis and Machine Learning Techniques to Exploit Social Eye-Tracking Data
Department: Graduate School of Education and Computer Science
Advisers: Roy Pea, Dan Schwartz, Paulo Blikstein (Education) and Andrew Ng and Jure Leskovec (Computer Science)
Abstract: I plan to use computational techniques to make sense of existing eye-tracking data. I previously conducted a study where dyads remotely worked on a set of contrasting cases (Schwartz & Bransford, 1998) to learn how the human brain processes visual information. In one condition, members of the dyads saw the gaze of their partner on the screen; in a control group, they did not have access to this information. This intervention helped students achieve a higher quality of collaboration and a higher learning gain compared to the control group. To my knowledge, this is the first study that was able to highlight the learning benefits of this kind of “gaze-awareness” tool. The eye-tracking data of this study, however, has been largely unexploited so far; I plan to develop skills in 1)network analysis, to describe how this effect took place; and 2), in machine learning, to predict how well future students will do based on their gaze patterns.
Project: Distortions from reality: A computational approach to the study of issue salience and coverage in news
Advisers: Justin Grimmer (Political Science) and Dan Jurafsky (Linguistics)
Abstract: The news media are an essential institution for a well-functioning democracy, informing the people about national issues so that they can make informed political decisions. Therefore, understanding the process by which news content is generated is highly important to society, particularly given the accusation that news content is intentionally distorted from reality in order to lead the people to interpret issues and events in a politically strategic way. However, an alternate account for news content can be found in the market model of news, where a news outlet’s economic considerations predominantly drive editorial decisions in generating content. These accounts both presume that the following qualities indicate distortion in the news: 1) proportion of coverage and 2) elements of language style. I intend to use computational methods and tools devised for investigating large-scale text corpora in order to validate or reject these theories. A computational approach towards evaluating distortion in news content is superior to previous attempts to study this issue, which relied upon the use of small samples of news data. This is because studying news data of an appropriate size for this question would be infeasible for humans, a shortcoming that computational methods lack. Furthermore, advances in natural language processing, both at the sentence- and document-level of analysis, allow for the quantitative study of coverage and style in news content.