Stable exponential random graph models with non-parametric components for large dense networks

Exponential random graph models (ERGM) behave peculiar in large networks with thousand(s) of actors (nodes). Standard models containing 2-star or triangle counts as statistics are often unstable leading to completely full or empty networks. Moreover, numerical methods break down which makes it complicated to apply ERGMs to large networks. In this paper we propose two strategies to circumvent these obstacles. First, we use a subsampling scheme to obtain (conditionally) independent observations for model fitting and secondly, we show how linear statistics (like 2-stars etc.) can be replaced by smooth functional components. These two steps in combination allow to fit stable models to large network data, which is illustrated by a data example including a residual analysis.     

The identification of interlock groups in large corporate networks: convergent validation using divergent techniques

The present study attempts to assess the substantive and theoretical implications of using different analytical techniques to identify groups of highly interlocked corporations within large corporate networks. Specifically, it compares the interlock groups identified using direct factor analysis with those identified using hierarchical cluster analysis. A comparison of these two sets of empirical results suggests that these two analytical techniques yield surprisingly similar findings even though they employ disparate methodological rationales and computational algorithms. Nevertheless, some minor but significant differences in the results provided by these two techniques arise as a consequence of different assumptions implicit in these techniques concerning the possibility of overlap between interlock groups. Finally, the convergence of these two analyses confirms that interlock groups are concrete empirical aggregations within corporate networks which cannot be dismissed as the artifacts of particular analytical methodologies.     

Familial groups in social networks

Many structural definitions for social community have been proposed in attempt to characterize and further understand the structure of social relationships. Algorithms using quantitative concepts such as centrality measures, spectral methods and other clustering measures have been used to compute social communities. While these methods have had much success in extracting meaningful subgroups in social and biological (and other) networks, they do not necessarily reveal the defining structure of social attraction.     

Social positions in influence networks

In this article we derive implications about social positions from a formal theory of social influence. The formal theory describes how, in a group of actors with heterogeneous initial opinions, a network of interpersonal influences enters into the formation of actors' settled opinions. We derive the following conclusions about a special form of structural equivalence. If actors are structurally equivalent in the network of interpersonal influences, then any dissimilarity of their initial opinions is reduced by the social influence process. If the social positions of actors are identical, i.e. if they have identical initial opinions and are structurally equivalent in the influence network, then they have identical opinions at equilibrium. If actors are not structurally equivalent in the network of interpersonal influences, then the social influence process does not necessarily reduce dissimilarities of initial opinions. We extend our analysis to consider automorphic equivalence.     

Perception differences between groups of employees identifying the factors that influence a return to work after a work-related musculoskeletal injury

Employee health and productivity losses as a result of work-related injury are estimated to be US dollars 1.2 trillion annually to US companies. This is approximately 14.3% of the gross domestic product [6,8,11,35]. Workers' compensation, medical care, and short and long-term disability are a part of these costs. Controlling or eliminating these costs is a problem for US employers [3,6,14,21,29]. The study discussed in this article examined the perceptions of manufacturing employees in identifying factors that influence a return to work after a work-related musculoskeletal injury. The classification of employees who participated in this study were safety professionals, supervisors and workers from the manufacturing industry in central Kentucky. The worker group consisted of material handlers, assembly line workers and quality control inspectors. The participants completed a developed survey instrument, the Return to Work Perception Survey. This survey instrument examined the perception of the participants on factors related to return to work: company policies and procedures, job satisfaction, worker relationships and work environment. The results indicated safety professionals and supervisors perceptions differ from workers on the variables of job satisfaction, worker relationships and work environment. Their perceptions did not differ on the variable relating to company policies and procedures. In addition, the safety professional and supervisor groups rated the items addressing job satisfaction higher than did the worker group. The worker group did not differ from one another on any of the factors. Implications of this study for manufacturing companies suggest (a). identifying those issues contributing to employee job satisfaction, (b). developing a plan for achieving increased job satisfaction and employee recognition at the workplace among all workers, and (c). consider allowing employees to develop new capacities and new learning, thus fostering motivation and job satisfaction [18,20].     

Peer influence and drug use among adolescents in rural areas

A sample of 1,004 eighth and tenth grade students in twenty-three small Central/East Texas communities was assessed to determine 1) their perception of the number of their friends who use drugs, 2) the amount of information they received about drugs from their friends, and 3) the connection between those perceptions and drug use. A multiple regression model which included grade, gender, the degree to which friends are perceived to use drugs and the amount of information about drugs received from friends explained 39 percent of the variance in the degree to which rural adolescents were involved in drug use. An item specific analysis of the subcomponents of these composite variables explained 44 percent of the variance in the degree to which rural adolescents were involved in drug use. This same four-factor model accurately classified over 81 percent of non-drug-users and 67 percent of users using discriminant analysis. Students who perceived a higher degree of drug use among their friends and who received more information about drugs from their friends used drugs more frequently. Lower marijuana use in these rural areas as compared to the nation, both as a peer perception and as a fact, may protect these students to a degree from broader patterns of drug use. The findings of this study support the theory that peer pressure is related to drug abuse, even in rural areas.     

Norm formation in social influence networks

I propose a mechanism of norm formation and maintenance that combines classical theory in social psychology on attitudes and social comparisons with a formal network theory of social influence. Underlying the formation of norms is the ubiquitous belief that there is a correct response for every situation and an abiding interest for persons to base their responses on these correct foundations. Given such a belief, a normative evaluation of a feeling, thought or action is likely to arise when persons perceive that their positive or negative attitudinal evaluation is shared by one or more influential others. If interpersonal agreements validate attitudes and transform attitudes into norms, then the development of a theory of norm formation may draw on extant “combinatorial” theories of consensus production that describe how shared attitudes are produced and maintained in groups. The network theory of social influence that I employ is one such combinatorial approach. An important sociological implication of this network theory is that the content of norms must be consistent with the social stratification (or more general pattern of inequality) of interpersonal influences in a group. I illustrate the theory with an analysis of Roethlisberger and Dickson’s (1939) classic observations on the Bank Wiring Observation Room.     

Mixing patterns between age groups in social networks

We present a method for estimating transmission matrices that describe the mixing and the probability of infection between age groups. Transmission matrices can be used to estimate age-dependent forces of infection in age-structured, compartmental models for the study of infectious diseases. We analyze the social network generated by the synthetic population of Portland and extract mixing patterns. Our results show that the mixing within the population consists of two groups, children and adults. Children interact most frequently with other children close to their own age, while adults interact with a wider range of age groups and the durations of typical adult contacts are shorter than typical contacts between children. Furthermore, the transmission matrix shows that children are more likely to acquire infection than adults.     

Sampling and estimation in large social networks

An unknown network is modelled by a directed or undirected graph having vertices of different kinds. Partial information is available concerning the vertex labels and the edge occurrences within a simple random sample of vertices. Using this information we find unbiased estimators and variance estimators of such graph parameters which can be given as dyad or triad counts. In particular, we give approximate formulae pertaining to large networks.     

Partitioning large signed two-mode networks: Problems and prospects

While a substantial amount of attention within social network analysis (SNA) has been given to the study of one-mode networks, there is an increasing consideration of two-mode networks. Recent research on signed networks resulted in the relaxed structural balance (RSB) approach and its subsequent extension to signed two-mode networks involving social actors and social objects. We extend this approach to large signed two-mode networks, and address the methodological issues that arise. We develop tools to partition these types of networks and compare them with other approaches using a recently collected dataset of United Nations General Assembly roll call votes. Although our primary purpose is methodological, we take the first step towards bridging Heider's structural balance theory with recent theorizing in international relations on soft balancing of power processes.     

Alternative estimation methods for identifying contagion effects in dynamic social networks: A latent-space adjusted approach

Contagion effects, also known as peer effects or social influence process, have become more and more central to social science, especially with the availability of longitudinal social network data. However, contagion effects are usually difficult to identify, as they are often entangled with other factors, such as homophily in the selection process, the individual’s preference for the same social settings, etc. Methods currently available either do not solve these problems or require strong assumptions. Following Shalizi and Thomas (2011), I frame this difficulty as an omitted variable bias problem, and I propose several alternative estimation methods that have potentials to correctly identify contagion effects when there is an unobserved trait that co-determines the influence and the selection. The Monte-Carlo simulation results suggest that a latent-space adjusted estimator is especially promising. It outperforms other estimators that are traditionally used to deal with the unobserved variables, including a structural equation based estimator and an instrumental variable estimator.     

Clearing the FOG: Fuzzy,overlapping groups for social networks

Humans are well known to belong to many associative groups simultaneously, with various levels of affiliation. However, most group detection algorithms for social networks impose a strict partitioning on nodes, forcing entities to belong to a single group. Link analysis research has produced several methods which detect multiple memberships but force equal membership. This paper extends these approaches by introducing the FOG framework, a stochastic model and group detection algorithm for fuzzy, overlapping groups. We apply our algorithm to both link data and network data, where we use a random walk approach to generate rich links from networks. The results demonstrate that not only can fuzzy groups be located, but also that the strength of membership in a group and the fraction of individuals with exclusive membership are highly informative of emerging group dynamics.     

Functional specificity in discussion networks: The influence of general and problem-specific networks on health outcomes

Though commonly used in social science research to measure the influence of social networks, the link between discussing “important matters” within networks and the achievement of instrumental ends is frequently unspecified or untested. To address this gap, we explore the consequences of using the general “important matters” approach versus one in which the substance of discussions is directly linked to the outcomes of interest. The egocentric social networks of people experiencing an acute health crisis are examined. We identify the characteristics of and the degree of membership overlap between “important matters” and “health matters” discussants, and explore their relative power in explaining a wide range of health and health services-related outcomes. We find no evidence for the influence of “important matters” networks, while characteristics of “health matters” networks demonstrate a significant effect across all models.     

City networks and commodity chains: identifying global flows and local connections in Ho Chi Minh City

Any analytical framework for understanding actual forms of the intensified incorporation of cities into the world economy needs to go beyond the exclusive focus on advanced producer services, which is characteristic of most of the World City Network (WCN) approach. Simultaneously, an account of the role of advanced producer services will strengthen Global Commodity Chain (GCC) analysis. A combination of the literatures on WCN and GCC can contribute to a broader conceptualization of the connections and connectivities of global cities. In addition, a combined approach will improve our understanding of globalization processes within many so-called 'third-world' cities that are experiencing booms in export-oriented industrialization and in migration from rural hinterlands as they are being integrated into Global Commodity Chains. We illustrate our argument with insights from GCC analyses of the electronics industry located in Ho Chi Minh City and the agricultural sector in its rural hinterland, the Mekong Delta.     

Attitudes towards foreigners and Jews in Germany: identifying the determinants of xenophobia in a large opinion survey

The ultimate aim of opinion surveys is the provision of information on the distribution of preferences and perceptions at the individual level. Yet, eliciting this information from the data is typically difficult. This paper uses a structural model to explain the answers on a set of questions regarding the perception of foreigners and Jews by native Germans. In this model it is assumed that in addition to observable individual characteristics there exists an underlying unobserved attitude towards minorities which drives the distribution of answers by native respondents. This latent variable in turn is assumed to be influenced by a set of observable socio-economic characteristics of the individuals. In order to estimate this model it is necessary to impose strong identification restrictions. Estimation results show that education is the key correlate of the perception of foreigners and Jews in Germany.     

Assessing the bias in samples of large online networks

We consider the sampling bias introduced in the study of online networks when collecting data through publicly available APIs (application programming interfaces). We assess differences between three samples of Twitter activity; the empirical context is given by political protests taking place in May 2012. We track online communication around these protests for the period of one month, and reconstruct the network of mentions and re-tweets according to the search and the streaming APIs, and to different filtering parameters. We find that smaller samples do not offer an accurate picture of peripheral activity; we also find that the bias is greater for the network of mentions, partly because of the higher influence of snowballing in identifying relevant nodes. We discuss the implications of this bias for the study of diffusion dynamics and political communication through social media, and advocate the need for more uniform sampling procedures to study online communication.     

Disentangling social media influence in crises: Testing a four-factor model of social media influence with large data

Social media empower publics by providing a platform for their voices during crises. Digital-enabled platforms allow individuals to become influentials by sharing their insights and expertise with others. Confronted with the fast-paced and complex dynamics of crises, we lack a systematic conceptualization and a valid measure of social media influence in the crisis context. By integrating diverse perspectives on influence, we propose a new framework that theorizes different dimensions of social media influence based on publics’ communicative behaviors during crises. This integrated framework offers a refined conceptualization and measurement of social media influence in crises by incorporating the network perspective. We tested the framework with large-scale Twitter data from four crises. Results from multigroup CFA on Twitter influencers suggest that social media influence is composed of four factors: output, reactive outtake, proactive outtake, and network positioning. Each factor is associated with a distinct set of users’ behavioral indicators (e.g., retweet). Implications for crisis communication and public relations are discussed.     

Detecting large cohesive subgroups with high clustering coefficients in social networks

Clique relaxations are used in classical models of cohesive subgroups in social network analysis. Clustering coefficient was introduced more recently as a structural feature characterizing small-world networks. Noting that cohesive subgroups tend to have high clustering coefficients, this paper introduces a new clique relaxation, α-cluster, defined by enforcing a lower bound α on the clustering coefficient in the corresponding induced subgraph. Two variations of the clustering coefficient are considered, namely, the local and global clustering coefficient. Certain structural properties of α-clusters are analyzed and mathematical optimization models for determining α-clusters of the largest size in a network are developed and validated using several real-life social networks. In addition, a network clustering algorithm based on local α-clusters is proposed and successfully tested.