I am a sociologist focusing on social inequality, social networks, and family demography. I apply and develop quantitative and computational methods to study fundamental sociological questions in new ways and gain new insights across a variety of research fields. My current work examines how friendship, family, political, and climate dynamics influence individual life chances and aggregate patterns of inequality. I focus on how individuals’ selection into and interactions within these social and natural environments collectively shape emergent dynamics of social stratification. Methodologically, I specialize in modeling causal effects, relational dynamics, multilevel phenomena, and heterogeneity within and between groups.
Friendship dynamics as drivers of social inequality
Adolescent friendship networks are characterized by low interaction across racial and socioeconomic lines. While the determinants of racial segregation are relatively well-established, the mechanisms driving socioeconomic segregation remain less understood.
In my first dissertation chapter, I examine socioeconomic segregation in friendship networks in US high schools, drawing on sociological theories that emphasize the role of structural barriers and boundary-making among groups with intersecting attributes. I evaluate the explanatory power of these perspectives with data from the Add Health study and a novel methodological approach that integrates demographic sampling weights and decomposition techniques into statistical network modeling. This approach allows me to disentangle the relative contributions of students’ residential locations, course and extracurricular selections, and racial and socioeconomic preferences to the pattern of socioeconomic segregation in high school friendship networks. The results from this study make three substantive contributions.
First, parents’ decisions on residential location and school enrollment produce compositional differences between schools that account for about half of the total segregation. The other half is attributable to students’ friendship choices within schools. This finding implies that existing desegregation efforts (e.g., busing, zoning) are insufficient to fully integrate friendship networks as they only address between-school segregation. Second, within schools, much of what prior research attributed to socioeconomic homophily is driven by students’ course selections and racial homophily. This result suggests that SES-integrated friendship networks will be difficult to achieve without addressing tracking practices and racial homophily within schools. Third, remaining socioeconomic homophily is characterized by unilateral closure among high-SES students rather than mutual avoidance between socioeconomic groups. Therefore, socioeconomic homophily differs qualitatively from racial homophily, which is marked by mutual avoidance. Together, these results clarify how socioeconomic segregation in high school friendship networks emerges from both structural barriers and intentional and unintentional boundary-making within an institutional structure that creates compositional disparities across schools and classrooms.
This paper has received three best paper awards and an R&R at the American Sociological Review. During my doctoral studies at Cornell University, I secured a three-year, $240,000 grant from the National Science Foundation to develop this project. As part of this project, I also co-authored a review paper synthesizing the state of the art in causal network analysis. The paper, published in the Annual Review of Sociology, is a comprehensive resource for researchers seeking to advance from descriptive to causal analyses with social network data.
Family dynamics as drivers of social inequality
Like friendship dynamics, family dynamics can also influence socioeconomic outcomes. A second line of my research examines their aggregate consequences for social inequality. Specifically, I investigate the “motherhood penalty”—a prominent concept in sociology and demography that describes the drop in earnings women experience after giving birth. While existing studies show that the motherhood penalty on women’s earnings has declined unevenly across economic strata, the consequences of these varying trends on overall earnings inequality remain unclear.
In my second dissertation chapter, I address this gap using Census data to analyze how the heterogeneous decline in the motherhood penalty over the past four decades has shaped earnings inequality within and between economic strata. I propose a new decomposition method to accomplish this methodologically.
The approach extends the classic variance decomposition (Western and Bloome 2009) to a causal framework, enabling researchers to analyze the distributional consequences of heterogeneous treatment effects in a population that can be separated into subgroups (e.g., race, class, or gender). Unlike methods focusing solely on mean differences between groups, my approach also considers variability within groups, offering a more comprehensive assessment of treatment effects on inequality.
The results of this study indicate that the uneven decline in the motherhood penalty since the 1980s has had a dual effect: it has increased earnings inequality between economic strata while decreasing inequality within them. Because the within-group effect is more pronounced than the between-group effect, the changes in the motherhood penalty have led to a net reduction in overall inequality. This pattern highlights the complex ways family dynamics can influence inequality and underscores the importance of examining both within-group and between-group disparities.
My method has broader applications beyond the motherhood penalty. It enables researchers to examine treatment effects on variance-based inequality statistics and decompose these effects into within-group and between-group components. Researchers can study inequality at specific time points, over time, and even under counterfactual scenarios (e.g., policies or interventions affecting pre-treatment inequality, treatment effects, or relative group sizes). I am preparing a manuscript for submission to a methodology journal where I formally develop the model, validate its performance through simulations, and provide an R package for implementation.
Political dynamics as drivers of social inequality
Research on inequality rarely translates into the implementation of equitable policies. A third area of my research explores the political barriers hindering progress toward social equity. One such barrier is government instability. Countries facing frequent cabinet collapses often become unstable and rudderless, fueling social unrest, eroding trust in institutions, and exacerbating inequality.
In my third dissertation chapter, which won the McGinnis Best Methods Paper Award from the Cornell Department of Sociology, I examine how the stability of coalition governments in parliamentary democracies is influenced by the interplay of political parties that comprise them.
This research involves a complex multilevel problem where outcomes at a higher level (government stability) are influenced by units at a lower level (political parties). While multilevel models exist for analyzing how lower-level outcomes (e.g., student achievement) are influenced by higher-level units (e.g., schools), we lack methods for studying micro-macro links where the conventional multilevel setup is reversed. Building on Snijders (2016), I develop such a model by incorporating parametrizable aggregation functions into the regression equation to model how parties and their interrelationships collectively shape governments. The model supports various aggregation functions (e.g., min, max, mean, sum) as it employs Bayesian estimation, which eliminates the need for analytically tractable likelihood functions.
I use my model to examine the relationship between party organization and the risk of coalition government breakdowns. The results suggest that parties with decentralized decision-making processes are less flexible in managing coalition governments, which increases the likelihood of breakdowns of coalitions they are part of. The aggregation analysis further reveals that this effect is moderated by coalitions’ interdependence structures: the more evenly distributed parliamentary seat shares among coalition parties, the stronger the impact of party organization on the risk of government breakdown. Consequently, while grassroots democratic parties promote inclusiveness, they also increase government instability, especially in coalitions where power is evenly distributed.
My method has broader applications beyond coalition research. It demonstrates that aggregation functions can be empirically derived when their specific forms are a priori unknown. In network and spatial analysis, weight matrices define the relative influence of neighboring units on focal units. My model enables researchers to estimate these weights instead of specifying fixed values. When constructing indices, researchers often aggregate data across different levels using predefined functions. My model allows them to test the hypothesized aggregation functions against empirical data.
In conclusion, the method provides an empirical-statistical approach to studying micro-macro links and contributes to sociological methodology for multilevel analyses. I am preparing a manuscript for submission to a methodology journal in which I formally develop the model, validate its performance through simulations, and provide an R package for its implementation.
Climate dynamics as drivers of social inequality
Climate change is not just a biophysical and technological challenge but also a complex social issue, with both social causes and consequences. Emerging sociological research on the topic indicates, for example, that extreme weather events (e.g., floods, droughts, heatwaves) worsen inequalities by disproportionately affecting vulnerable communities and countries.
During my postdoc at Princeton University, I will expand my research on drivers of inequality to include climate dynamics as a fourth pillar. In collaboration with Filiz Garip, I will analyze the climate-inequality nexus through a network lens. I will use high-resolution weather data to track the evolution of microclimates across 170 communities in Mexico since the 1980s. This will allow me to create bipartite networks linking communities to different types of extreme events and address the following key questions: How have networks of communities that experience similar weather events evolved over time? Do communities with similar microclimates experience different levels of vulnerability to extreme events, and what are the resulting patterns of inequality within and between these communities? I intend to submit a grant proposal to NSF to fund this project and support student training.
As part of this research line, I have collaborated with the OECD International Programme for Action on Climate (IPAC) to develop an R Shiny application to monitor countries’ climate actions and policies. This app, which was presented at COP27, can be accessed here. More information can be found here.