Personal networks among people experiencing homelessness are vital but often overlooked. Drawing on a unique three-year dataset (2022-2024) from over 3,000 unhoused individuals in King County, Washington, this study analyzes the structure and change of four different personal networks: acquaintances, close friends, kinship (household), and peer referrals. Using this respondent-driven sample data set, we find that the mean size of the acquaintance networks ranged from 22 to 27 people, while the mean number of close friendships declined from 4.9 to 4.19. This drop, despite population growth, suggests a decrease in network density and potentially an increase in social isolation or a shift in composition (e.g., more people experiencing homelessness for the first time). However, we do see growth in the kinship networks, which grew slightly, indicating a greater prevalence of family cohomelessness or reliance on kin. These shifts highlight the need for policies that foster social connection and community stability.

Open political dialogue is regarded as foundational to democratic health and our social fabric. Here, we study political dialogue by examining with whom we prefer to talk about politics and why. In so doing, we develop and test hypotheses about what random encounters (e.g., meeting similar versus dissimilar others, meeting friends or colleagues) foster political dialogue, pose and test conjectures about what features of extended networks facilitate political debate, and present a new unique factorial survey experiment to answer our questions. We incorporated this factorial survey experiment within the NEtherlands Longitudinal Lifecourse Study 2022 (NELLS) and presented to a large sample of Dutch citizens – including Dutch ethnic majority members as well as minoritized Dutch with a Turkish or Moroccan heritage – a choice to engage in political talk or not. Hierarchical linear models reveal that relationship strength, rather than identity similarity (e.g., gender, ethnic), is the primary driver of opting for political dialogue. However, in encounters lacking prior relationship history, gender similarity increases the willingness to engage, and similar political views do relate to engaging in substantive debate modeled dichotomously. Ethnic diversity within extended networks fosters political discussion, while network size has a nonlinear association – medium-sized networks are more conducive to dialogue than very small or very large ones. These findings contribute to debates on political polarization by highlighting the relational conditions that encourage political exchange. • Political conversations are key to democratic health and our social fabric. • We hypothesize that dyadic similarity, tie strength, network size, and network diversity facilitate political conversation. • We study political conversations through a unique survey experiment. • Political conversations are more likely among strong ties, medium-sized extended networks, and with more ethnic diversity. • These findings have implications for our understanding of choice homophily, extended network structure, and polarization.

Aggregate relational data (ARD) on individuals’ ties to groups in a population offer valuable insights into the size of personal networks and the extent of segregation in contact with those groups. However, existing ARD models face two key limitations. First, they mask heterogeneity in network sizes across individuals who may differ markedly in relevant characteristics despite exhibiting similar response patterns on the ARD instrument. Second, although existing models can measure the overall level of segregation in contact with groups, they cannot reveal the determinants of segregation. To address these limitations, we introduce the Covariate Model, a regression-based framework that incorporates respondent covariates into analyses of ARD. We illustrate this model using ARD on contact with occupational categories. In addition to obtaining more substantively plausible network size estimates than existing approaches, the Covariate Model uncovers novel segregation patterns. For example, covariates — driven primarily by educational differences — account for a considerable portion of the segregation in contact with Higher Service occupations (e.g., lawyers and professors) but contribute little to explaining barriers to interaction with other occupational classes. By modeling the determinants of contact patterns in acquaintanceship networks, the Covariate Model extends the analytical reach of ARD and opens new avenues for research on social capital and segregation. • Covariate Model uncovers heterogeneity in network size masked by existing models. • Covariate Model reveals the determinants of segregation in contact with subgroups. • Covariate Model provides improved fit over existing models. • Covariate Model is widely applicable to research on social capital and segregation.

Bullying is a complex social construct, and informants (e.g., bullies, victims, and peer observers) may have differing perceptions about bully-victim relationships. This study examines how informant perspectives differ using network data obtained from 438 fifth-grade students ( M age = 11.19 years old, 46 % girls) in 13 Indonesian elementary school classrooms. Using a cross-informant framework, we investigated how self- and peer-reported bully-victim relationships overlapped as a function of the sex of bullies and victims, friendship ties, and relational schemas (i.e., mental network heuristics). Results from a multiplex exponential random graph model revealed significant agreement between self- and peer-reports. There was greater agreement when bully-victim relationships occurred between non-friends. When self- and peer-reports disagreed, peers identified more instances of boys engaging in bullying than girls, as well as more cross-sex than same-sex bully-victim relationships. Self-reports more often identified bully-victim relationships between friends than between non-friends. Post-hoc analyses revealed that bullies and their friends often had conflicting views of their friendship. Additionally, peers reported more victims per bully and fewer bullies per victim when compared to self-reports. These findings contribute to the understanding of the network structure of bully-victim perceptions and offer practical implications for identifying bully-victim relationships. • Consensus on bully-victim relationships: Preadolescent bullies, victims, and peers often agreed on those ties. • Asymmetric awareness: When views differed, peers saw more victims per bully and fewer bullies per victim than self-reports. • Reciprocal bully-victim relationships: These ties were more often mutual than expected by random networks. • Moderation by gender: Peers reported more bullying by boys and more cross-sex bully-victim ties than self-reports. • Relational misperceptions within friendships: Bullies often saw victims as friends; this bullying was less visible to peers.

Spending concentration, political influence, and collusion violate rules and principles of open and fair public procurement, leading to corrupt contract allocation. This study adopts stochastic actor-oriented models to test the evolution of these forms of procurement corruption risks in a longitudinal network study of 33579 construction contracts pertaining to Guatemalan local governments from 2012 to 2020. We identify a range of network configurations, based on past empirical research and theory, that capture different patterns of suspicious micro tendencies suggestive of corruption. We show how these micro tendencies shift in strength according to changes in electoral cycles and anti-corruption interventions, thus shedding light on how interventions may temporarily impact corrupt behavior, and how it may adapt and persist after a period of transition. The results indicate that collusion and spending concentration play significant roles in sustaining the risk of corrupt contract allocation, and that this behavior is able to rebound even after the introduction of anti-corruption interventions and new political regimes. The findings underscore the importance of local interventions and advocate for network approaches to enhance transparency, accountability, and long-term anti-corruption efforts.