ABSTRACT: Selecting an appropriate method of clustering for network data a priori can be a frustrating and confusing process. To address the problem we build on an a posteriori approach developed by Grimmer and King (2011) that compares hundreds of possible clustering methods at once through concise and intuitive visualization. We adapt this general method to the context of social networks, extend it with additional visualization features designed to enhance interpretability, and describe its principled use, outlining steps for selecting a class of methods to compare, interpreting visual output, and making a final selection. The interactive method, implemented in R, is demonstrated using Zachary's karate club, a canonical dataset from the network literature.
ABSTRACT: Network analysis has become a popular tool to examine data from online social networks to politics to ecological systems. As more computing power has become available, new technology-driven methods and tools are being developed that can support larger and richer network data, including dynamic network analysis. This timely merger of abundant data and cutting edge techniques affords researchers the ability to better understand networks over time, accurately show how they evolve, find patterns of growth, or study models such as the diffusion of innovation. We combine traditional methods in social network analysis with new innovative visualizations and methods in dynamic network studies to explore an online tobacco-control community called GLOBALink, using almost twenty years of longitudinal data. We describe the methods used for the study, and perform an exploratory network study that links empirical results to real-world events.
ABSTRACT: Often objects are to be ranked. However, there is no measurable quantity available to express the ranking aim and to quantify it. The consequence is that indicators are selected, serving as proxies for the ranking aim. Although this set of indicators is of great importance for its own right, the most commonly used practice to obtain a ranking is an aggregation method. Any aggregation, however suffers from the effect of compensation, because the aggregation technique is in the broadest sense an averaging method. Here an alternative is suggested which avoids this averaging and which is derived from simple elements of the theory of partially ordered sets (posets). The central concept in partial order is the 'concept of comparison' and the most general outcome is a web of relations between objects according to their indicator values, respecting the ranking aim. As an example the 'Failed State Index' (FSI), annually prepared by the Fund of Peace is selected. The FSI is based on twelve individual contextual different indicators, subsequently transformed into a single composite indicator, by simple addition of the single indicator values. Such an operation leaves space for compensation effects, where one or more indicators level out the effect of others. Hence, a comparison between the single states (in total 177) based on their mutual FSI ranking has its limitations as the comparisons are made based on the composite indicator. We show that brain drain is one of the indicators in the FSI-study that plays a crucial role in the ranking, whereby the ranking aim is the stabilization of nations.
ABSTRACT: An affiliation network consists of actors and events. Actors are affiliated with each other by virtue of the events they mutually attend. This article introduces a family of affiliation measures that captures the extent of actors' affiliations in the network. At one extreme, one might have an actor who attended many events, but none of these events were attended by any of the other actors in the network. Although of high degree, in no reasonable interpretation would such an actor be considered highly affiliated with other actors in the network. At the other extreme, one might have an actor defined by a collection of events, all of which were attended by another actor(s), making the actor as enmeshed in the network as possible. Most actors will be between these extremes, with some events being shared by varying others, and some not. This article introduces a family of affiliation measures based on the entries of the co-occurrence matrix. After defining the measures, the cumulative distribution function of first-order affiliation is derived and expressed as a difference of binomials.
ABSTRACT: Building on world-systems theory, simulation models of 5-line intersocietal networks were generated in an effort to understand systemic power hierarchies. The societal nodes were exclusively connected by three types of interaction: migration, warfare, and unequal trade. These networks can be considered “mixed relation” networks due to the ways in which these types of ties combine positive and negative sanction flows. Insights from elementary theory were employed to understand how exclusion from these different types of ties might influence the resulting power distributions. Additionally, the resource carrying capacity of the nodes was varied by structural position in an effort to differentiate the influence of structural position and individual attributes on location in the hierarchy. It was determined that exclusion from interaction is likely a structural, scale invariant mechanism that helps to determine power distributions above and beyond the inherent attributes of network actors.
ABSTRACT: Advances in text analysis, particularly the ability to extract network based information from texts, is enabling researches to conduct detailed socio-cultural ethnographies rapidly by retrieving characteristic descriptions from texts and fusing the results from varied sources. We describe this process and illustrate it in the context of conflict in the Sudan. We show how network information can be extracted from vast quantities of unstructured texts-based information using computer assisted processes. This is illustrated by an examination of changes in the political networks in Sudan as extracted from the Sudan Tribune. We find that this approach enables rapid high level assessment of a socio-cultural environment, generates results that are viewed as accurate by subject matter experts, and match actual historical events. The relative value of this socio-cultural analysis approach is discussed.
ABSTRACT: In this paper, we employ archival materials from multiple institutional sources to reconstruct the dynamic network of interorganizational collaboration that emerged in response to the Hurri- cane Katrina disaster of late 2005. Over the period from initial storm formation through the first week following landfall in Louisiana, we record active participation by over 1,500 organizations in response activities. We here conduct an exploratory analysis of the growth and evolution of the network of collaboration among responding organizations, an identification of organizations that emerged as central actors in the response process, and the cohesive subgroups that crystal- lized within the larger network. Finally, we conclude with a discussion of several issues related to the use of archival methods in research on interorganizational networks in disaster settings, and to the use of automated methods for network extraction.
ABSTRACT: We propose a visual representation of bibliographic data based on shared references. Our method employs a distance metric that is derived from bibliographic coupling and then subjected to fast approximate multidimensional scaling. Its utility is demon- strated by an explorative analysis of social network publications that, most notably, depicts the genesis of an area now commonly referred to as network science. However, the example also illustrates some common pitfalls in bibliometric analysis.
ABSTRACT: This paper introduces a new computer-based visualization method, the parallel arc diagram (PAD), which is capable of uniquely representing 2-mode temporal relationships in a manner that assists in highlighting simple features of the network. The PAD approach relies on a computer’s ability to render link lines adjacent to each other with orderly precision, resulting in features that facilitate preattentive processing of simple network characteristics and providing the ability to discern patterns of interactions over time. PADs supplement existing methods such as node-link diagrams by offering a simple alternative visualization without the computational complexity of graph layout algorithms and the additional issues that animation introduces. This paper subjectively evaluates the PAD approach using low level task taxonomies developed for assessing adjacency matrix and node-link visualization effectiveness. We argue based on those taxonomies that the PAD approach is as effective or in some cases more effective than existing approaches except for tasks requiring the identification of structural groups or middle-man nodes. This paper also demonstrates how the PAD approach can be utilized in a software application. The TIPAD (Temporal Interactive Parallel Arc Diagram) uses character participation in movie scenes as a test-bed for exploring social interactions over time and provides the ability to compare a PAD based visualization with traditional visualizations of the same network.
ABSTRACT: In the present study, the social linkages of street-involved youth and correlates of infection with chlamydia and gonorrhea are explored. This is the first study to assess the social linkages of street- involved youth using RDS. Eleven street-involved youth aged 14 to 24 were selected as seeds to recruit their peers into the study using RDS (N=169). Study staff administered a questionnaire, obtained a urine specimen, and provided recruitment coupons to participants. A week later, participants were provided with test results and treatment if necessary. RDS Analysis Tool was used to assess the effectiveness of RDS and define the social linkages. A Fisher‘s Exact test was used to identify any correlates of infection. Gender was the only variable that correlated with infection status (22 percent of females vs. 8 percent of males). A high proportion of male participants had never been tested before. Despite the fact that most female participants had been tested before, high infection rates indicate that more accessible and frequent testing is required. Street-involved youth are connected socially to those who share similar health related behaviors. There is a need for increased testing options and opportunities for street-involved youth.
ABSTRACT: In this paper we propose relational hyperlink analysis (RHA) as a distinct approach for empirical social science research into hyperlink networks on the World Wide Web. We demonstrate this approach, which employs the ideas and techniques of social network analysis (in particular, exponential random graph modeling), in a study of the hyperlinking behaviors of Australian asylum seeker advocacy groups. We show that compared with the commonly-used hyperlink counts regression approach, relational hyperlink analysis can lead to fundamentally different conclusions about the social processes underpinning hyperlinking behavior. In particular, in trying to understand why social ties are formed, counts regressions may over-estimate the role of actor attributes in the formation of hyperlinks when endogenous, purely structural network effects are not taken into account. Our analysis involves an innovative joint use of two software programs: VOSON, for the automated retrieval and processing of considerable quantities of hyperlink data, and LPNet, for the statistical modeling of social network data. Together, VOSON and LPNet enable new and unique research into social networks in the online world, and our paper highlights the importance of complementary research tools for social science research into the web.
ABSTRACT: Bonacich (1987) suggested a family of centrality measures that provide a useful way of modeling questions of power and network constraint. However, the literature offers little guidance regarding the choice of β, the parameter which alters the way the measure accounts for the effect of having powerful contacts in one's network. In this paper I explore the way the choice of the β parameter affects the power indices the Bonacich measure generates. I consider three network properties which might affect the way the choice of β influences the Bonacich power indices. I find that in high density networks with few internal 'chains' and few pendants, the choice of β is largely immaterial. Conversely, in sparse networks, those with a high proportion of pendant nodes, or those with many chains, the value of β has a substantial effect on the power indices the measure generates. Next I consider whether power indices produced by interior values of β might be represented as a linear combination of "pure" vectors, those generated with values of β at either end of the parameter range and β = 0. I find that in the vast majority of cases a linear combination of "pure" vectors power is equivalent to using indices produced by interior values of β, making the choice of β largely moot. Finally, in the unlikely case that this disaggregation is inappropriate, I discuss the question of determining an appropriate value of β empirically.
ABSTRACT: Changes in observed social networks may signal an underlying change within an organization, and may even predict significant events or behaviors. The breakdown of a team's effectiveness, the emergence of informal leaders, or the preparation of an attack by a clandestine network may all be associated with changes in the patterns of interactions between group members. The ability to systematically, statistically, effectively and efficiently detect these changes has the potential to enable the anticipation, early warning, and faster response to both positive and negative organizational activities. By applying statistical process control techniques to social networks we can rapidly detect changes in these networks. Herein we describe this methodology and then illustrate it using four data sets, of which the first is the Newcomb fraternity data, the second set of data is collected on a group of mid-career pU.S. Army officers in a week long training exercise, the third is the perceived connections among members of al Qaeda based on open source, and the fourth data set is simulated using multi-agent simulation. The results indicate that this approach is able to detect change even with the high levels of uncertainty inherent in these data.
ABSTRACT: Our goal in this paper is to explore two generic approaches to disrupting dark networks: kinetic and non-kinetic. The kinetic approach involves aggressive and offensive measures to eliminate or capture network members and their supporters, while the non-kinetic approach involves the use of subtle, non-coercive means for combating dark networks. Two strategies derive from the kinetic approach: Targeting and Capacity-building. Four strategies derive from the non-kinetic approach: Institution-Building, Psychological Operations, Information Operations and Rehabilitation. We use network data from Noordin Topbent. Using this strategic framework as a backdrop, we strongly advise the use of SNA metrics in developing alterative counter-terrorism strategies that are context-dependent rather than letting SNA metrics define and drive a particular strategy.
ABSTRACT: Labor mobility, both as a mechanism of knowledge diffusion and as a kind of social glue that holds together small production communities operative within a given territory, deserves serious consideration. In this context, focusing on a specific industrial cluster in Ankara, this paper reveals the extent and characteristics of the social networks created by the mobile laborers in order to understand the interconnections between social context, knowledge spillovers, innovation and labor mobility. For this purpose a step-wise algorithm is employed in order to identify social sub-groups by employing social network analysis and by drawing on the flow data constructed for this study. What is evident from this study is that the social network created by the mobility of laborers in Siteler, an industrial cluster specialized in furniture production, reveals a topography of social relations that cannot be split into equally large blocks but eventually parceled out to micro parts consisting of generally 2 or 3 firms. Interestingly, the contexts of innovation also unveil that innovative firms tend to be located at an intermediate position, not an upper and central position, within the topography of the network.
ABSTRACT: This study examines global patterns of stability and change within six longitudinal samples of English-language weblogs (or "blogs") during the 2004 U.S. Presidential election campaign. Using distance-based methods of graph comparison, we explore the evolution of the blog-blog citation networks for each sample during the period. In addition to describing the qualitative dynamics of the blog networks, we relate major campaign events (e.g., party political conventions and debates) to the observed pace of change. As we demonstrate, such events are associated with substantial divolatility is also shown to have strong seasonal and endogenous components. Our findings suggest that external factors (both regular and episodic) may be important drivers of network dynamics.
ABSTRACT: Whereas much theoretical and empirical work concerning exchange networks examines the payoffs to actors in particular positions, a much smaller body of work focuses on exchange patterns (i.e., who exchanges with whom). The problem we address in this paper is that the existing methods to predict patterns of exchange are limited by computational cost to relatively small networks. We develop a computationally efficient method for locating and predicting patterns of exchange. The theory on which we draw argues that each actor.s resistance to unfavorable offers is based on relative position within the exchange network. Using experimental data we first show that the order of exchanges does affect payoffs as the network structure changes over time. We then attempt to predict the ordering of exchanges for a number of networks using three existing methods and one new method derived from our theory. As a test of the predicted orderings, we compare each method.s predictions for six exchange networks against experimental results. In general, the results suggest that our more parsimonious method performs quite well relative to the more computationally complex prediction methods.
ABSTRACT: Analysis of social network data is often hampered by non-response and missing data. Recent studies show the negative effects of missing actors and ties on the structural properties of social networks. This means that the results of social network analyses can be severely biased if missing ties were ignored and only complete cases were analyzed. To overcome the problems created by missing data, several treatment methods are proposed in the literature: model-based methods within the framework of exponential random graph models, and imputation methods. In this paper we focus on the latter group of methods, and investigate the use of some simple imputation procedures to handle missing network data. The results of a simulation study show that ignoring the missing data can have large negative effects on structural properties of the network. Missing data treatment based on simple imputation procedures, however, does also have large negative effects and simple imputations can only successfully correct for non-response in a few specific situations.
ABSTRACT: When ideas and tools move from one field to another, the movement is generally from the natural to the social sciences. In recent years, however, there has been a major movement in the opposite direction. The idea of centrality and the tools for its measurement were originally developed in the social science field of social network analysis. But currently the concept and tools of centrality are being used widely in physics and biology. This paper examines how and why that.wrong way.movement developed, its extent and its consequences for the fields involved.
ABSTRACT: This paper presents the design and results of a statistical and visual analysis of a dynamic signed network. In addition to prevalent approaches to longitudinal networks, which analyze series of cross-sectional data, this paper focuses on network data measured in continuous time in order to explain the signs of lines rather than their occurrence. As a consequence, current stochastic actor-oriented models for network change cannot be applied. Instead, multilevel logistic regression analysis is used for uncovering the main statistical regularities of network evolution. Visualization by means of animated Scalable Vector Graphics with several options for interaction allows for in-depth inspection of network evolution and offers detailed information on the people involved in the network. The substantive focus of the paper is on the evaluations and complex labeling process among literary authors and critics illustrating the interplay between identity (literary style) and structure. It is hypothesized that actors do not just evaluate their immediate ego-network; they also try to survey and interpret the overall structure of the network and derive part of their identity from it. The latter, however, is a collective process involving communication, e.g., publicly labeling groups of actors in the network and adapting behavior to labels that have previously been assigned to actors. Including perceptions of overall network structure and classifications in dynamic network models would extend current actor-oriented models.
ABSTRACT: Social roles in online discussion forums can be described by patterned characteristics of communication between network members which we conceive of as 'structural signatures.' This paper uses visualization methods to reveal these structural signatures and regression analysis to confirm the relationship between these signatures and their associated roles in Usenet newsgroups. Our analysis focuses on distinguishing the signatures of one role from others, the role of "answer people." Answer people are individuals whose dominant behavior is to respond to questions posed by other users. We found that answer people predominantly contribute one or a few messages to discussions initiated by others, are disproportionately tied to relative isolates, have few intense ties and have few triangles in their local networks. OLS regression shows that these signatures are strongly correlated with role behavior and, in combination, provide a strongly predictive model for identifying role behavior (R2=.72). To conclude, we consider strategies for further improving the identification of role behavior in online discussion settings and consider how the development of a taxonomy of author types could be extended to a taxonomy of newsgroups in particular and discussion systems in general.
ABSTRACT: Despite cross-disciplinary interest in social influence among adolescent peer groups, significant variations in collecting and analyzing peer network data have not been explored, so it is difficult to disentangle substantive and methodological differences in peer influence studies. We analyze two types of network data (self-reported friendships and multi-informant reports of children who hang around together a lot) with three methods of identifying group structures (two graph theoretic approaches and principal components analysis) to explore substantive differences in results. We then link these differences back to underlying features of the networks, allowing greater insight into the general problem of identifying groups in network data. We find that different analytic approaches applied to the same network data produced moderately concordant group solutions, with higher concordances for multi-informant data. The same analytic approaches applied to different relational data (on the same nodes) produced weaker concordance, suggesting that the underlying data structure may be more salient than analytic approach in accounting for different results across studies. Behavioral similarity among group members was greatest for approaches that rest directly on density of direct ties.
ABSTRACT: If graph drawing is to become a methodological tool instead of an illustrative art, many concerns need to be overcome. We discuss the problems of social network visualization, and particularly, problems of dynamic network visualization. We consider issues that arise from the aggregation of continuous-time relational data ("streaming" interactions) into a series of networks. We describe our experience developing SoNIA (Social Network Image Animator, http://sonia.stanford.edu) as a prototype platform for testing and comparing layouts and techniques, and as a tool for browsing attribute-rich network data and for animating network dynamics over time. We also discuss strengths and weakness of existing layout algorithms and suggest ways to adapt them to sequential layout tasks. As such, we propose a framework for visualizing social networks and their dynamics, and we present a tool that enables debate and reflection on the quality of visualizations used in empirical research.
ABSTRACT: A set of measures for Simmelian tie strength, Simmelian brokerage, and, being Simmelianly brokered are introduced. The measures are derived from interpretations of a quote from Simmel (1950). The theoretically most informative measure of Simmelian brokerage is based on a complex value measure of Simmelian tie strength reflected in an Hermitian matrix. Also measures based on weight matrices and hypergraphs are discussed. A maximum for the number of ties one node could Simmelian broker in a network of n nodes is determined.
ABSTRACT: In this paper we present an approach to Social Network Analysis, based on statistical analysis of conceptual distance between people. In particular, we introduce the concept of valued centrality and a generalisation of geodesic distance which we call link distance. We examine a number of benefits of the link distance concept, including ease of visualisation and applicability of common statistical methods. Using a case study, we demonstrate how examining the statistical relationships between link distance and other forms of conceptual distance can offer insights into the nature of communication within an organisation. Thus an integration of the graph-theoretic techniques traditional in Social Network Analysis, and the statistical techniques traditional in other Social Sciences, leads to a combined technique which integrates the strengths of both approaches.
ABSTRACT: Given the increasing threat of terrorism and spread of terrorist organizations, it is of vital importance to understand the properties of such organizations and to devise successful strategies for destabilizing them or decreasing their efficiency. However, intelligence information on these organizations is often incomplete, inaccurate or simply not available. This makes the study of terrorist networks and the evaluation of destabilization strategies difficult. In this paper, we propose a computational methodology for realistically simulating terrorist networks and evaluating alternative destabilization strategies. We proceed to use this methodology to evaluate and conduct sensitivity analysis of the impact of various destabilization strategies under varying information surveillance regimes. We find that destabilization strategies that focus on the isolation of individuals who are highly central are ineffective in the long run as the network will heal itself as individuals who are nearly structurally equivalent to the isolated individuals "move in" and fill the communication gaps.
ABSTRACT: The order in which people freely recall a set of words, persons' names, or other items indicates how they organize those items in memory. An individual's cognitive structure of the persons with whom he or she has some particular relation can be described by noticing how he or she associates from one person to the next during recall. Using improved statistical measurement, we review in detail five studies that systematically examined associative patterns in the recall of persons and evaluate competing hypotheses about the nature of these patterns. Individuals in these studies recalled their acquaintances, coworkers, and friends. Across studies, the results consistently show that persons recalled adjacently or successively are perceived to interact more with each other than those not recalled adjacently. No other factor describes associative patterns as well as this notion of perceived social proximity. These results, along with related research, imply the influence of social networks on memory for persons and suggest a universal feature of human social cognition.
ABSTRACT: In this study, we explore the combined effects of layout and motion on viewers' perceptions of social network data. We ask viewers to interpret the overall network and we ask domain specific questions about managing change within a departmental team to understand how network display influences viewers' overall perception of networks. We find that motion has a positive effect on the accuracy of viewers' perceptions of change in status from formal to informal networks. We also find no main effect for hierarchical versus spatially central layout on viewers' accuracy. There is a significant interaction effect of motion and graph layout on viewers perception of change. Finally, we find that when viewers are asked to make interpretations of the overall graph, they bring their own pre-existing graphical vocabulary that may influence their interpretation.
ABSTRACT: This paper explores the use of visualization methods at two different levels of analytical detail (e.g., moving from a time-independent aggregated view to a more detailed time series analysis) of data derived from Reuters news tickers between September 11 and November 15, 2001. This data comprises a large, complex set of related words that presents an interesting analytical challenge. The relationships among words in the news articles were extracted using Centering Resonance Analysis (CRA). In this paper, we present two secondary analyses of the CRA networks. The first analysis involves a two-dimensional layered grid approach using convex hulls for examining the intersection and union of sets of words in the text in a two-dimensional projection. Additional SVG versions of these images allow exploration of the solutions interactively. The second approach involves the use of three-dimensional interactive visualizations and centrality analysis using the molecular visualization program Mage. A major guiding question of this work is the following: Can we gain knowledge about the conflicts within and dynamics of the "Bush team" from a visual exploration and analysis of the CRA data over the 66 days of the text? The analysis of the news text using these approaches suggests that Powell has played less of a role during the 66 days of the analysis then might be expected given his formal status as chief foreign policy advisor to the President.
ABSTRACT: This paper contributes to an ongoing debate in International Political Economy about the appropriateness of globalization, regionalization and macroeconomic imbalance theory by identifying quantitative estimates for all three tendencies from world trade data. This is achieved with a series of gravity models enhanced stepwise by the mapping of the estimation errors of a given model on representations of the overall structure of trade. This not only allows the identification of imperfections in a given model but also permits the further improvement of the models since any systematic regional organization in the error-terms can be identified. The results of the most elaborated model indicate that single factor explanations of global economic integration are presumably misleading. Instead, each of three explanations captures only part of the ongoing changes, as they can be identified under a comparative static perspective from world trade data.
ABSTRACT: The research we report here tests the "Freeman-Linton Hypothesis" which we take as arguing that the structure of a set of relational ties over a population is more strongly determined by type of relation than it is by the type of species from which the population is drawn. Testing this hypothesis requires characterizing networks in terms of the structural properties they exhibit and comparing networks based on these properties. We introduce the idea of a structural signature to refer to the profile of effects of a set of structural properties used to characterize a network. We use methodology described in Faust and Skvoretz (forthcoming) for comparing networks from diverse settings, including different animal species, relational contents, and sizes of the communities involved. Our empirical base consists of 80 networks from three kinds of species (humans, non-human primates, non-primate mammals) and covering distinct types of relations such as influence, grooming, and agonistic encounters. The methods we use allow us to scale networks according to the degree of similarity in their structuring and then to identify sources of their similarities. Our work counts as a replication of a previous study that outlined the general methodology. However, as compared to the previous study, the current one finds less support for the Freeman-Linton Hypothesis.
ABSTRACT: This paper is about estimating the parameters of the exponential random graph model, also known as the p* model, using frequentist Markov chain Monte Carlo (MCMC) methods. The exponential random graph model is simulated using Gibbs or Metropolis-Hastings sampling. The estimation procedures considered are based on the Robbins-Monro algorithm for approximating a solution to the likelihood equation.
A major problem with exponential random graph models resides in the fact that such models can have, for certain parameter values, bimodal (or multimodal) distributions for the sufficient statistics such as the number of ties. The bimodality of the exponential graph distribution for certain parameter values seems a severe limitation to its practical usefulness.
The possibility of bi- or multimodality is reflected in the possibility that the outcome space is divided into two (or more) regions such that the more usual type of MCMC algorithms, updating only single relations, dyads, or triplets, have extremely long sojourn times within such regions, and a negligible probability to move from one region to another. In such situations, convergence to the target distribution is extremely slow. To be useful, MCMC algorithms must be able to make transitions from a given graph to a very different graph. It is proposed to include transitions to the graph complement as updating steps to improve the speed of convergence to the target distribution. Estimation procedures implementing these ideas work satisfactorily for some data sets and model specifications, but not for all.
ABSTRACT: Most personal (egocentric) network studies describe networks using measures that are not structural, opting instead for attribute-based analyses that summarize the relationships of the respondent to network members. Those researchers that have used structural measures have done so on networks of less than 10 members who represent the network core. Although much has been learned by focusing on attribute-based analyses of personal network data, the application of structural analyses that are traditionally used on whole (sociocentric) network data may prove fruitful. The utility of this approach becomes apparent when the sample of network members elicited is relatively large.
Forty-six respondents free-listed 60 network members and evaluated tie strength between all 1,770 unique pairs of members. Graph-based measures of cohesion and subgroups revealed variability in the personal network structure. Non-hierarchical clustering generated subgroups that were subsequently verified by respondents as meaningful. Further analysis of the correlation between subgroup types and overlap between subgroups demonstrates how the analysis of each network can be summarized across subjects. Four case studies are presented to illustrate the richness of the data and the value of contrasting individual matrix results to the norm as defined by all 45 subjects.
ABSTRACT: Networks of exchange opportunities can evolve as dissatisfied agents search for new partners. Are there stable networks whose participants do not look for new potential partners? What do these networks look like? How is the outcome of this evolutionary process related to the beginning network? Computer simulations are used to explore how networks of exchange opportunities evolve when agents can change positions. These simulations suggest that only networks with high degrees of power imbalance are unstable and that there are three forms for stable networks: equal power, indeterminate power, and inconsistent power (corelessnesss).
ABSTRACT: We propose a novel visualization approach that facilitates graphical exploration and communication of relative actor status in social networks. The main idea is to map, in a drawing of the entire network, actor status scores to vertical coordinates. The resulting problem of determining horizontal positions of actors and routing of connecting lines such that the overall layout is readable is algorithmically difficult, yet well-studied in the literature on graph drawing. We outline a customized approach. The advantages of our method are illustrated in a study of policy making structures from the privatization processes of former East German industrial conglomerates, in which the visual approach led to additional findings that are unlikely to have been revealed using non-visual means of analysis.
ABSTRACT: In the last fifteen years, ecosystem ecologists have developed a theoretical approach and a set of computational methods called ecological network analysis (Ulanowicz, 1986; Kay et al. 1996). Ecological network analysis is based on input/output models of energy or material flows (e.g., carbon compound flows) through a trophic network (e.g., a food web describing which species eats which other species). Mathematically and conceptually, this ecological network analysis approach is strikingly similar to work in the field of social network analysis, particularly the influence models of Hubbell (1965), Katz (1963), and Friedkin and Johnsen (1990). In food web research, Yodzis and Winemiller (1999), have recently proposed a new way to operationalize the concept of a "trophospecies", which is a set of species with similar foods or predators. Their definition turns out to be identical to the notion of structural equivalence (Lorrain and White, 1971) in social network analysis, particularly as conceived by Burt (1976) and Burt and Talmud (1993). The striking convergence to date of the fields of ecology and sociology via independent invention of network concepts suggests that there may be considerable value in cross-fertilization of the two fields. With this paper we hope to begin a dialogue between the two fields, by applying advanced social role theory and methods to the study of food webs. In social network analysis, the introduction of the notion of structural equivalence thirty years ago was followed by the development of regular coloration (White & Reitz, 1983; Everett & Borgatti, 1991), an important advance over structural equivalence for modeling social roles. The objective of our paper is to answer a call in the ecological literature for greater clarity in thinking about the role of species in ecosystems (Simberloff and Dayan, 1991), by applying the notion of regular coloration to food webs.
ABSTRACT: This chapter discusses theoretical sociology in historical perspective: From the classic tradition to postclassical efforts of synthesis that culminated in multiple paradigms, to the situation today in which theorists are more and more constructing formal models as essential components of their methodology. The classical phase is treated very briefly and the discussion of the postclassical phase is limited to two major theorists, Parsons and Homans, in terms of their common focus on the Durkheimian problem of social integration. The bulk of the chapter deals with developments in recent theoretical sociology. I describe models of structure and of process before defining two types of models that combine a structural focus with process analysis. Finally, I set out a general perspective on theoretical model building and conclude with a discussion of standards in the assessment of such work.
ABSTRACT: This paper examines the degree to which the constraints imposed by various social contexts influence social interaction. We draw on two data sets. In each, we compare the patterning of interaction of the same individuals across different contexts. If minimal constraints are imposed, then the interaction patterns among the individuals in the two contexts should be similar. But if one of the contexts involves major constraints, then interaction patterns in the two should differ. The results suggest further that the constraints found in any context are not unlimited in their impact. Moreover, individuals who can, apparently do manipulate the context to minimize the constraint imposed by the context.
ABSTRACT: Data structures comprising many binary variables can be represented graphically in various ways. Depending on the purpose different plots might be useful. Here two ways of showing associations between variables and implications between variables are discussed. The methods are based on conditional independence graphs and lattices of maximal cluster-property pairs. Applications to multivariate samples and network data are briefly discussed.
ABSTRACT: In this article, we discuss the concept of social integration and its implications for health. We provide both an overview of the social epidemiology and a review of theories of how participation in a diverse social network might influence health. We also present evidence from a prospective study of social network diversity (number of social roles) and susceptibility to the common cold in people experimentally exposed to a cold virus. We found that the greater the social diversity, the lesser the susceptibility to infectious illness. However, our attempts to isolate the pathways through which social diversity was associated with susceptibility (health practices, hormones, immune function) were unsuccessful. The relation was independent of the number of people in the social network, and of personality characteristics thought to influence social participation.
ABSTRACT: We present an overview of eigen analysis methods and their applications to network analysis. We consider several network analysis programs/procedures (Correspondence Analysis, NEGOPY, CONCOR, CONVAR, Bonacich centrality) that are at their core eigendecomposition methods. We discuss the various matrix representations of networks used by these procedures and we give particular attention to a variety of centering and normalizing procedures that are carried out prior to the analysis. We compare three types of iterative procedures with the standard SVD in terms of pragmatic concerns and the results produced by each method. We show how the initial matrix representations and the adjustments made between iterations influence the results obtained. Finally, we show that the eigen perspective clearly highlights the similarities and differences between different network analysis procedures.
ABSTRACT: This paper documents the use of pictorial images in social network analysis. It shows that such images are critical both in helping investigators to understand network data and to communicate that understanding to others.
The paper reviews the long history of image use in the field. It begins with illustrations of the earliest hand-drawn images in which points were placed by using ad hoc rules. It examines the development of systematic procedures for locating points. It goes on to discuss how computers have been used to actually produce drawings of networks, both for printing and for display on computer screens. Finally, it illustrates some of the newest procedures for producing web-based pictures that allow viewers to interact with the network data and to explore their structural properties.