Social and Behavioral Sciences

Social Media’s Role in Career Development During Young Adulthood

Shanley Corvite, Information

As the next generation of young adults enters the workforce, it is essential to examine how they develop career interests, aspirations, and ultimately make career decisions. Social media has become a pervasive feature of everyday life, exposing young people to a wide range of ideas and narratives about work. This exposure may shape how they make sense of work, value different careers, and form expectations about the workplace. My dissertation comprises three studies that will identify the types of career-related messages on social media, examine how exposure to these messages shapes career interests and aspirations, and compare these messages with those from traditional sources of career information. By highlighting how social media use influences young people’s socialization into the working world, my dissertation builds on theory and empirical research in human-computer interaction and social computing, vocational psychology, and communication and media studies.

The Space Between Us: When, How, and Why Interpersonal Distance in Workplace Relationships Influences Individual and Relational Outcomes

Laurel Detert, Management and Organizations

Workplaces abound with opportunities for employees to develop close interpersonal relationships that benefit both the individuals and their organizations. However, employees report being lonelier than ever before at work and a growing body of research highlights the potential dark side of workplace relationships. I examine how the distinct types of interpersonal distance — the psychosocial, structural, and functional space between two or more individuals — that characterize a workplace relationship play an important and influential role in whether the relationship serves a functional purpose. Across three chapters, I unpack how and why specific dimensions of interpersonal distance in different workplace relationships (i.e., mentor-mentee, leader-follower, and team member relationships) influence individual, dyadic, and group outcomes (i.e., investment behaviors, interpersonal trust, emotion and conflict patterns, performance). With this work, I begin to integrate a more thorough study of interpersonal distance into the literature on workplace relationships to facilitate better relational dynamics.

Borrowing and Adjustment: Student Debt and Household Responses to Economic Change

Kelcie Ferrara-Gerson, Public Policy and Economics

This dissertation explores how student debt influences the ways households adjust to economic change. The first chapter shows that income-driven repayment programs, which adjust payments with income, can reduce the severity of economic downturns by stabilizing household spending. The second chapter studies how student debt shapes access to homeownership, finding that higher debt burdens make it harder to qualify for mortgages and build wealth. The third chapter examines how communities respond to trade shocks, showing that more exposed regions experience larger losses of educated workers, driven by both lower investment in education and greater out-migration. Together, these essays demonstrate how debt and policy shape economic resilience and access to opportunity across households and regions.

The (Functional) Echo Chamber Does Exist: How the Dynamics of Online Political Communication Can Polarize Users

Audrey Halversen, Communication and Media

The digital transformation of the political information environment has sparked concerns about online echo chambers: politically homogeneous online spaces that result in attitude reinforcement and polarization. Recently, a number of studies have indicated that these concerns are unnecessary, finding that many people are exposed to a heterogeneous mix of political information online. Yet, members of the U.S. public continue to believe that social media sites polarize users. In this three study, mixed methods dissertation, I offer an explanation for this discrepancy between public concerns and academic findings. I argue that extant work into online echo chambers has focused primarily on the degree to which people encounter pro- vs. counter-attitudinal news online, while paying less attention to a) qualitative characteristics of both pro- and counter-attitudinal content that may breed negative outcomes and b) political information that is generated, contextualized, and circulated not by news organizations but by ordinary users and influencers who teach their followers how to interpret political events and ideas. In this work, I propose that some social media users may belong to functional echo chambers: online communication environments that are not conducive to serious consideration of counter-attitudinal views, not because they lack counter-attitudinal information, but because that counter-attitudinal information is of low quality and/or has been recontextualized and criticized by attitudinally similar peers. Ultimately, I argue that people whose information environments resemble functional echo chambers may experience similar outcomes associated with traditional echo chambers, such as anger, affective polarization, attitude polarization, and belief superiority.

Enhancing Teacher Judgment with AI: How Mathematics Teachers Interpret and Respond to Simulated AI-Generated Student Work

Soobin Jeon, Educational Studies

There is growing recognition that the most profound gains from artificial intelligence (AI) will emerge not from automation but from collaboration, whereby the outcomes of human and AI partnerships exceed what either can achieve independently—yet realizing this potential remains challenging. This dissertation examines this issue in mathematics teaching, focusing on how AI can support teachers in anticipating, interpreting, and responding to students’ varied thinking in geometry, where ideas are expressed through diagrams and teachers must make complex judgments about what students understand based on their work. Using an interactive simulation, teachers engage in these three phases of professional judgment with AI-generated representations of student thinking. The findings will demonstrate how AI can function as a resource that leverages teacher expertise, supporting instructional decision-making while preserving the central role of human judgment.

Why Do Governments Collect High Quality Data?

Rachel Kulikoff, Health Services Organization and Policy

Collecting data is a ubiquitous function of government, yet collecting high-quality data is resource-intensive and costly. This dissertation asks, when and why do governments collect high-quality data? How and why does the quality of this data vary? I use a comparative case study design within the United States; my empirical cases are maternal mortality data, the consumer price index, and the poverty threshold. I use mixed methods, conducting elite interviews, doing archival studies and textual analysis, as well as making use of quantitative epidemiological methods, to construct a novel measure of data quality. I argue that levels of bureaucratic autonomy, expertise, and the broader environment of political contestation influence changes in government data quality. This research on the political construction of government data fills an interdisciplinary blind spot that spans political science, health services research, and science and technology studies.

Adaptation to Environmental Extremes in a Great Ape (Gorilla beringei)

McKensey Miller, Anthropology

It has become essential to understand animal populations’ adaptive ability to predict how species may be impacted by the changing climate. One especially important aspect of adaptation to investigate is an organism’s ability to thermoregulate, a major determinant of survival in any given environment. This dissertation aims to understand how endangered mountain gorillas (Gorilla beringei) in Volcanoes National Park, Rwanda, behaviorally and physiologically thermoregulate in their extreme high elevation environment. The project demonstrates the use of novel techniques for collecting fine-scale environmental data to capture information about the conditions that populations and individuals directly experience and to which they respond. This research provides important opportunities for understanding how environmental variability is related to behavioral and physiological plasticity, which has broader implications for the study of primate evolution as well as the conservation of non-human primates in a rapidly changing climate.

When Being Alone Doesn’t Hurt: Demystifying the Puzzle of Solitude

Micaela Rodriguez (Roblin Fellow), Psychology

Decades of research have documented serious risks associated with being alone, such as loneliness and poor health. Yet many people seek time alone, and growing evidence suggests that it can promote well-being. In this dissertation, I address this puzzle by examining individual, contextual, and technological factors that shape the experiences of being alone. In Chapter 1, multi-method evidence demonstrates that people’s beliefs about being alone influence how lonely they feel. In Chapter 2, large-scale data (N>350,000 across 58 countries) reveal that the health implications of being alone vary widely and systematically by cultural context. In Chapter 3, I examine how an emerging technology redefining what it means to be alone—AI companionship—is affecting social and emotional life. These findings shed light on when, why, and for whom spending time alone promotes versus undermines well-being. I conclude by discussing implications for global efforts to combat loneliness and promote social connection.

Farming the Sun: Planning for Farmland Access Amid the Energy Transition

Lanika Sanders, Urban and Regional Planning

U.S. farmland faces mounting pressures from urban expansion and agricultural consolidation, constraining land access for small and midsized farmers, particularly those from marginalized and underresourced groups. As the climate crisis deepens, solar energy introduces a vital but competing use for farmland, exacerbating existing land access inequities. Planning scholarship and practice are well positioned to protect farmland access amid the renewable energy transition, yet little research examines how planners engage farmland retention amid rapid solar expansion. Employing a mixed-methods, multi-scalar design, this dissertation analyzes solar and agricultural policies, planning efforts, and funding mechanisms across federal, state, and local levels. Focusing on Minnesota as a prototypical case, the study traces barriers and opportunities for under-resourced farmers to leverage solar development and assesses how planners, developers, and policymakers shape equitable or inequitable outcomes. The findings inform strategies to align renewable energy goals with farmland retention and robust regional food systems.

Econometric Methods for Credible Policy Analysis in the Digital Age

Lonjezo Sithole, Economics

This dissertation enhances the credibility of modern empirical research by tackling two fundamental challenges: reliable use of complex high dimensional or unstructured data (including texts and images) and empirical testing of behavioral assumptions behind welfare analysis. It contributes: (i) a framework for distributional causal inference with unstructured data and sparse labels; (ii) a robust and safe way to integrate machine learning into judge/examiner designs; and (iii) a technique that ensures valid comparisons in high-dimensional observational studies by improving overlap in distributions of background characteristics between treated and control groups. It also provides the first fully nonparametric test for Slutsky symmetry, a core prediction of consumer theory that also underlies nonparametric welfare analysis. Together, these contributions provide a practical toolkit that delivers transparent and reliable evidence for economists and policymakers.

Reality Reckoning: Reality TV After 2020

Olivia Stowell, Communication and Media

Reality Reckoning investigates how post-2020 U.S. reality TV negotiates changes in legislation and policy, global media flows, emergent technologies, and cultural conversations about identity, representation, and power. As the George Floyd uprising, the #MeToo movement, the so-called “culture wars,” the COVID-19 pandemic, the streaming wars, and more have destabilized both conventions of television production and conceptions of race, gender, and labor, I argue that reality TV functions as a crucial site by which everyday people encounter ideology and politics. By close reading reality programs alongside trade press coverage of casting protocols, landmark legal battles related to employment in the TV industry, and the design of streaming video platforms, Reality Reckoning reveals how shifts in discourses of race, gender, and labor produce shifts in media industry production, which in turn produce new textual politics in contemporary media.

Awakening a Feminist Consciousness: Reimagining Chineseness with Women’s Entrepreneurship in a Shifting Global Order

Huiran Yi, Information

This dissertation examines how young Chinese women re-politicalize feminism on digital platforms and through their entrepreneurial labor amid widespread anti-feminist sentiment. Based on 18 months of ethnographic research with the entrepreneurial and grassroots communities that young Chinese women build in the United States, I trace the emergence of a new, digital feminist ideal on prominent Chinese platforms and how these feminist values in turn shape the businesses and communities that young women establish. I theorize these women’s technological and entrepreneurial practices as a transnational “recursive public” that builds cultural infrastructure, enabling their feminist pursuits. The formation of this new feminist subjectivity takes place through the circulation of ideas and imaginaries between the online and offline, between China and the United States. My research advances our understanding of the role young women play not only in shaping contemporary ideas about Chineseness but also in crafting transnational mobilities amid global geopolitical tensions.

Working and Learning in Hard Times: A Causal Analysis of How Employment Downturns Shape Community College Students’ Trajectories

Chenjun Yu, Higher Education

Despite community college (CC) students’ extensive workforce participation and vulnerability to economic shocks, we lack causal evidence on how employment downturns affect their educational and employment outcomes. This gap is critical: two-thirds of working CC students hold jobs unrelated to their studies, creating tensions between financial necessity and academic success that intensify during recessions. This dissertation examines how the Great Recession impacted working CC students’ college and job trajectories. Drawing on Minnesota administrative data linking over 300,000 students’ enrollment and employment records with local economic indicators, I address: (1) how do downturns affect students’ work intensity and program choices and (2) how do downturns alter relationships between work intensity and credential completion and earnings? I employ difference-in-differences with shift-share instruments to establish causal effects across student populations and institutional contexts. Preliminary analyses reveal substantial variation in how recessions affect students across programs and regions, informing targeted support strategies during economic disruption.

Balancing Privacy Protection and Statistical Inference in Microdata: The Role of Synthetic Data

Chendi Zhao, Survey and Data Science

As public demand for access to social microdata grows, statistical agencies increasingly rely on disclosure limitation methods to protect respondent confidentiality, with synthetic data emerging as a widely used approach. Designing and applying these methods is complex, as they must protect confidentiality while preserving valid statistical relationships. This dissertation makes three contributions to the evaluation and use of synthetic data methods for disclosure protection. First, it provides an empirical assessment of how synthetic data compares with commonly used traditional methods in terms of their impact on small area estimation accuracy. Second, it proposes two fairness-constrained synthetic data generation approaches that aim to maintain privacy protection and statistical accuracy, particularly for underrepresented groups. Third, it extends synthetic data methods to longitudinal survey settings, developing strategies to preserve coherent patterns across waves. Together, these studies provide empirical evidence and guidance for balancing confidentiality and analytic validity in microdata protection and release.

Dancing with Metrics: Digitalization of Women’s Labor in China

Jun Zhou (Roblin Fellow), Sociology

Since the reform era, Chinese capitalism has remade women’s labor in its own image. The socialist factory demanded their bodies. The post-reform service economy demanded their affect. Now, platform capitalism demands something new: their faces, voices, and personalities—rendered into data and optimized in real time. My dissertation examines this transformation through China’s trillion-dollar live-commerce industry, where 38 million streamers sell on camera while algorithms govern their every move. Drawing on multiple years of ethnographic fieldwork, 141 interviews with streamers, managers, and platform designers, and textual data, I ask why the same industry produces billionaire entrepreneurs and mass precarity simultaneously. I show that power over women’s labor has shifted into platform infrastructure—not as a technical black box, but as sociotechnical systems that sort workers into accumulation or dependency. This research reveals how digitalization produce new forms of labor stratification, with implications for class formation and social reproduction under digital capitalism.