Social Networks最新文献

The exponential-family random graph models (ERGMs) have emerged as an important framework for modeling social networks for a wide variety of relational types. ERGMs for valued networks are less well-developed than their unvalued counterparts, and pose particular computational challenges. Network data with edge values on the non-negative integers (count-valued networks) is an important such case, with examples ranging from the magnitude of migration and trade flows between places to the frequency of interactions and encounters between individuals. Here, we propose an efficient parallelizable subsampled maximum pseudo-likelihood estimation (MPLE) scheme for count-valued ERGMs, and compare its performance with existing Contrastive Divergence (CD) and Monte Carlo Maximum Likelihood Estimation (MCMLE) approaches via a simulation study based on migration flow networks in two U.S. states. Our results suggest that edge value variance is a key factor in method performance, while network size mainly influences their relative merits in computational time. For small-variance networks, all methods perform well in point estimations while CD greatly overestimates uncertainties, and MPLE underestimates them for dependence terms; all methods have fast estimation for small networks, but CD and subsampled multi-core MPLE provides speed advantages as network size increases. For large-variance networks, both MPLE and MCMLE offer high-quality estimates of coefficients and their uncertainty, but MPLE is significantly faster than MCMLE; MPLE is also a better seeding method for MCMLE than CD, as the latter makes MCMLE more prone to convergence failure. The study suggests that MCMLE and MPLE should be the default approach to estimate ERGMs for small-variance and large-variance valued networks, respectively. We also offer further suggestions regarding choice of computational method for valued ERGMs based on data structure, available computational resources and analytical goals.

Access to social capital is stratified by socioeconomic status and has been cross-sectionally linked to involvement in voluntary organizations. Yet, we know little about the origin and interplay of these empirical regularities. Regression analyses on German panel data (SC6-NEPS) reveal that people rich in social capital join organizations more often (selection). Furthermore, joiners access more and higher-status social capital after joining (socializing opportunities). Low-status individuals disproportionally extend their reach towards higher positions through involvement but join less often. Compared to a counterfactual situation in which nobody joins, current involvement patterns marginally reduce some socioeconomic inequalities in access to social capital.

In a previous paper, Kovacs (2010) proposed a generalized relational similarity measure based on iterated correlations of entities in a network calibrated by their relational similarity to other entities. Here I show that, in the case of two-mode network data, Kovacs’s approach can be simplified and generalized similarities calculated non-iteratively. The basic idea is to rely on initial similarities calculated from transforming the two-mode data into one-mode projections using the familiar duality approach due to Breiger (1974). I refer to this as two-mode relational similarities and show, using the Southern Women’s data and data from Senate voting in the 112th U.S. Congress, that it yields results substantively indistinguishable from Kovacs’s iterative strategy.

Given increased political polarization and racial tension in the wake of the 2016 presidential election in the United States, this study examines dropped ties in personal networks at that time based on political and racial identities. We employed data from the 2015–2018 UCNets study (n = 1159), a longitudinal, representative data set of the San Francisco Bay Area. In late 2015 and early 2016 it generated personal network data via multiple name generators, eliciting alters whom respondents socialized with, confided in, received advice from, exchanged social support with, and found difficult. Using multilevel multinomial logit models, we then examined various reasons for tie dissolution immediately following the inauguration of Trump in early 2017. The results show that among young adults, politically dissimilar alters were more likely to be dropped due to disagreements. With respect to racial homophily, we found that interracial dyads were more likely to be dropped because of drifting apart or some other reason for both younger and older cohorts. Overall, there is some support for the notion that dropped ties due to political disagreements did occur immediately following the 2016 election, but the results highlight the continuing significance of race in personal networks.

As a specific art convention, graduating music students in college often invite network members to attend their degree recitals. This customary practice of network mobilization involves two types of tie effects that seemingly conflicted with each other, which became more acute when the practice abruptly changed during the Covid-19 pandemic. Based on invitation and attendance records collected from seventy-one recitals in a span of two decades before (N = 4866) and during the pandemic (N = 428) in Taiwan, multilevel analysis helped untangle which concert and tie features contributed to successful invitations at both the recital and invitation levels. Recruiting a larger proportion of weak ties helped boost the overall attendance at the recital level, while strong ties ensured positive responses to individual invitations in terms of both meeting at the recital hall and recalling the recital, particularly before the pandemic. More importantly, certain cross-level effects changed during the pandemic while others remained intact.

Despite a growing recognition of the importance of insurgent group structure on conflict outcomes, there is very little empirical research thereon. Though this problem is rooted in the inaccessibility of data on militant group structure, insurgents frequently publish large volumes of image data on the internet. In this paper, I develop a new methodology that leverages this abundant but underutilized source of data by automating the creation of a social network graph based on co-appearance in photographs using deep learning. Using a trove of 20,645 obituary images published online by the PKK, a Kurdish militant group in Turkey, I demonstrate that an individual’s centrality in the resulting co-appearance network is closely correlated with their rank in the insurgent group.

While previous research indicates that students benefit from their peers’ resources, little is known about access to social capital in the school context. Therefore, this study examines differential access to social capital – measured by friends’ socioeconomic status (SES), the number of books they have at home, and their reading habits – in secondary schools in Germany, the Netherlands, and Sweden. Relying on a large-scale dataset, I investigate the association between socioeconomic status, minority status, and social capital using complete friendship network information. I argue that social capital access is connected to a two-stage process consisting of school sorting and friendship selection. To differentiate between these two processes, I apply within-between random effects (REWB). The models show that friendship selection is much less relevant for access to social capital than school sorting. Results indicate that while high-SES students have better access to social capital across dimensions, access patterns for minority students are more nuanced.

Despite the progress in pharmaceutical and epidemiological tools for combating HIV spread, HIV stigma remains a significant social barrier impeding treatment and prevention efforts, potentially reducing the effectiveness of interventions to reduce HIV transmission. In this paper, we propose a novel approach to defining and estimating HIV stigmatization through the structure of sexual relations, as opposed to attitudes. We conceptualize structural stigma as arising from two mechanisms: (1) a reduced propensity towards HIV serodiscordant partnerships (exclusion); and (2) a reduced propensity towards partnerships with seroconcordant individuals who themselves have serodiscordant partnerships (ostracism). Both mechanisms can be assessed from observed partnership network data using exponential family random graph models (ERGMs). We demonstrate our approach on a sexual contact network of black men who have sex with men in the South Side of Chicago. We find a tendency for serodiscordant sexual relationships to be suppressed in the network ($\theta$ = −0.69, p $<$ .05), as well as a suppressive tendency for HIV negative YBMSM to have sex with other HIV negative people in serodiscordant relationships ($\theta$ = −0.96, p $<$ .05) suggesting that structural HIV stigma is present in this network. Potential relationships with attitudinal stigma and implications for epidemiological strategies for reducing HIV stigma are discussed.

Researchers often use pooled exponential random graph models (ERGM) to analyze samples of networks. However, pooled ERGM—here, understood to include both meta-regression and combined estimation on a stacked adjacency matrix—may be biased if there is heterogeneity in the latent error variance (‘scaling’) of each lower-level model. This study explores the implications of scaling for pooled ERGM analysis. We illustrate that scaling can produce bias in pooled ERGM coefficients that is more severe than in single-network ERGM and we introduce two methods for reducing this bias. Simulations suggest that scaling bias can be large enough to alter conclusions about pooled ERGM coefficient size, significance, and direction, but can be substantially reduced by estimating the marginal effect within a block diagonal or random effects meta-regression framework. We illustrate each method in an empirical example using Add Health data on 15 in-school friendship networks. Results from the application illustrate that many substantive conclusions vary depending on choice of pooling method and interpretational quantity.

In this study, we explore the relationship between inter-organizational network dynamics and innovation outcomes. We focus on node turnover and argue that both cluster and broker dynamics can range from low (stable) to high (volatile), resulting in differentiated outcomes. The data comprises 318 consortium members participating in 104 R&D consortia forged in a 23-year period in the Dutch water sector. Our analysis reveals two equifinal combinations (stable brokers – volatile clusters and volatile brokers – stable clusters) that both generate significantly higher innovation outcomes compared to networks with low, moderate, or high dynamics across the entire network.