18,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Geisteswissenschaft
- Sprache: Englisch
In Communities and Networks, Katherine Giuffre takes the science of social network analysis and applies it to key issues of living in communities, especially in urban areas, exploring questions such as: How do communities shape our lives and identities? How do they foster either conformity or innovation? What holds communities together and what happens when they fragment or fall apart? How is community life changing in response to technological advances?
Refreshingly accessible and built on fascinating case examples, this unique book provides not only the theoretical grounding necessary to understand how and why the burgeoning area of social network analysis can be useful in studying communities, but also clear technical explanations of the tools of network analysis and how to gather and analyze real-world network data. Network analysis allows us to see community life in a new perspective, with sometimes surprising results and insights, and this book enables readers to gain a deeper understanding of social life and the relationships that build (and break) communities.
This engaging text will be an exciting new resource for upper-level undergraduate and beginning graduate students in a wide range of courses including social network analysis, community studies, urban studies, organizational studies, and quantitative methods.Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 389
Veröffentlichungsjahr: 2013
Ähnliche
Communities andNetworks
Communities andNetworks
Using Social NetworkAnalysis to Rethink Urban andCommunity Studies
KATHERINE GIUFFRE
polity
Copyright © Katherine Giuffre 2013
The right of Katherine Giuffre to be identified as Author of this Work has been asserted inaccordance with the UK Copyright, Designs and Patents Act 1988.
First published in 2013 by Polity Press
Polity Press65 Bridge StreetCambridge CB2 1UR, UK
Polity Press350 Main StreetMalden, MA 02148, USA
All rights reserved. Except for the quotation of short passages for the purpose of criticism and review, no part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the publisher.
ISBN: 978-0-7456-6461-3
A catalogue record for this book is available from the British Library.
The publisher has used its best endeavours to ensure that the URLs for external websites referred to in this book are correct and active at the time of going to press. However, the publisher has no responsibility for the websites and can make no guarantee that a site will remain live or that the content is or will remain appropriate.
Every effort has been made to trace all copyright holders, but if any have been inadvertently overlooked the publisher will be pleased to include any necessary credits in any subsequent reprint or edition.
For further information on Polity, visit our website: www.politybooks.com
Contents
Acknowledgments
1 What is network analysis and how can it be useful?
What is structure?
Networks as metaphors
Metaphor into method
What is a social network?
Further developments
Why use network analysis?
Plan of the book
Software
A closer look: basic network terms and definitions
Entering and displaying matrices in UCINET6
2 What is a community? Where does it come from?
Gemeinschaft und Gesellschaft
Durkheim: mechanical and organic solidarity
Simmel: individuality and social groups
“The duality of persons and groups”
Cohesion
A closer look: matrix multiplication
Matrix multiplication in UCINET6
3 What do communities do for us?
Gift exchanges as social support and obligation
The world the slaves made
Different types of support
Conjugal roles and social networks
Agency in networks
The Search for an Abortionist
A closer look: density
Calculating density in UCINET6
4 How do communities shape identity?
Community and ethnicity in Chicago
Minorities, majorities, and the weight of numbers
Assimilation in Chicago
Small worlds and six degrees of separation
“The strength of weak ties”
The disadvantages of strong communities
A closer look: balance
5 What happens when communities become fractured?
Witchcraft in Salem
Arson and the Indochinese in East Boston
Isolation and atomization in Germany
A closer look: centrality
How to calculate centrality in UCINET6
6 How do communities mobilize for collective action and social movements?
Riots in Washington
Urban revolt in Paris
Recruiting high-risk activists
Structuring connections
A closer look: structural equivalence
How to find structural equivalence with UCINET6
7 How do communities foster creativity and innovation?
Social networks and innovative thinking
The creative context of the city
Small world structures
The Apple example
A closer look: correspondence analysis
Doing correspondence analysis in UCINET6
8 How do new communities differ from traditional communities?
Rethinking the concept of community
Cyber utopians vs. cyber dystopians
A different kind of space
The structure of the web
A closer look: multidimensional scaling
Doing multidimensional scaling in UCINET6
Glossary of network terms
References
Index
Acknowledgments
I would like to thank, first of all, Jonathan Skerrett at Polity, for his enthusiasm for this project and his kindness, helpfulness, support, and insight in seeing this book through from the very beginning all the way to the finish.
I would like to thank the following for permission to use copyrighted material: Social Forces for permission to use table 2.9 and figure 2.3 from Ronald Breiger’s “The duality of persons and groups,” University of Chicago Press for permission to use table 3.1 from Barry Wellman and Scot Wortley’s “Different strokes from different folks,” Nancy Howell for permission to use figure 3.5 from The Search for an Abortionist, Steven Borgatti for permission to use table 3.5 and table 3.6 from the UCINET data files, Harvard University Press for permission to use figure 5.1 from Paul Boyer and Stephen Nissenbaum’s Salem Possessed, the American Sociological Association for permission to use figure 6.1 from Roger Gould’s “Multiple networks and mobilization in the Paris Commune, 1871,” and the Annual Review of Sociology for permission to use figure 8.1 and figure 8.2 from Duncan Watts’s “The ‘new’ science of social networks.” In addition, Stephanie Denton of the US Bureau of Labor Statistics expeditiously provided the data used for the correspondence analysis in chapter 7. Three anonymous reviewers provided extremely helpful critical feedback and thoughtful advice on the manuscript.
Finally, I would like to thank the members of my family for being so enormously supportive. Aiden Giuffre gave the manuscript a close reading and careful editorial attention, Tristan Giuffre gave very valuable hugs when the going got tough, and Jonathan Poritz gave both of those and everything else from mathematical expertise to homemade pumpkin pie. He proved once again that he is not only decorative, but also functional. This book is dedicated with love to him.
1 What is network analysisand how can it beuseful?
Humans are inherently social beings. We have lived in communities as long as we have been human. We form communities and, in many ways, communities form us. Our lives in communities shape who we are and how we approach the world. Understanding how communities work is vital to understanding our lives, but communities are complex and analyzing them can be difficult. They are spaces both of cooperation and of conflict. They can give us both a sense of belonging and one of alienation. They can root us in a place and also allow us to transcend the confines of geography.
For example, in the 1690s the Puritan village of Salem, Massachusetts, erupted in a paroxysm of witch-hunting the likes of which had never before been seen in North America. In the 1930s, community after community of solid citizens in Germany turned on their neighbors and succumbed to Nazism. What was happening at the level of the community to make the citizens of these places behave in these ways?
In the early twentieth century, Jewish immigrants from Eastern Europe to Chicago gradually assimilated into the American culture. In the late twentieth century, however, Indochinese refugees to Boston were the victims of arson by their neighbors who wanted the Indochinese out of their community. How and why do different communities react differently to the presence of outsiders in their midst?
In the 1960s, the residents of the ghetto in Washington, DC, responded to the assassination of Rev. Martin Luther King, Jr., with days of riots and looting. In the 1980s, the residents of New York City responded to the AIDS crisis by building a broad network of organizations intended to respond to a wide-ranging population of people in need. How and why does crisis elicit such different responses from different communities?
Community members can provide social support for each other in a variety of ways and that support can play a part in, for example, the explosion of innovative thinking that characterized Silicon Valley at the end of the twentieth century. But community members can also exert enormous pressure on each other to conform to established traditions, customs, and ways of thinking or risk being ostracized from the group and all of the social goods that it provides. How do members of communities navigate this tension between support and suffocation? Are there new types of communities emerging in the twenty-first century that are better able to maintain the balance between having too much community support and having too little?
These examples illustrate some of the ways in which communities are complex, many-faceted phenomena. Analyzing communities and communal life, therefore, is fascinating, but can at times seem overwhelming. Social network analysis helps make sense of all of this. Using social network analysis to look at community structures can give us a new perspective, new insights, richer understanding, and surprising answers to questions like the ones above and many more. The rise of social network analysis in the past half century has opened up wide ranges for exploring the questions of how human life is organized on a variety of levels and in a variety of disciplines. This book will look at how incorporating network analysis into urban and community studies can help us to understand questions and issues in those fields and can lead us to think about problems in new and fruitful ways.
What is structure?
Network analysis is the study of structure. But what do network analysts mean when they use the term “structure”? Berkowitz writes that “[t]he idea that social systems may be structured in various ways is not new. In fact, all of the established social sciences have evolved some notion of structure. But, until recently, no field had taken the idea of a regular, persistent pattern in the behavior of the elementary parts of a social system and used it as a central or focal concept for understanding social life” (1982: 1, emphasis in original). Wasserman and Faust add, “Regularities or patterns in interactions give rise to structures” (1994: 6–7).
What this means in the context of network analysis is that the persistent patterns of relations among the participants in a system become the core of the analysis. It is the relationships between the members of the system rather than the individual attributes of those members that are the key component of understanding the system. For example, network analysts would focus on the web of relations among various artists and galleries (rather than, say, the talent of individual artists) in order to understand why some artists find success in the art world and others do not (Giuffre 1999). The logic behind this thinking is part of our everyday understanding of how the world operates when we say things like, “It’s not what you know; it’s who you know.” Network analysts do not stop at merely “who you know,” however, but investigate much more deeply into the persistent patterns of those relationships. These patterns of relationships are what we call “structure” (Wasserman & Faust 1994: 3).
Much social science research in a variety of fields focuses exclusively on individual attributes (like talent, in the example above) and ignores important information about the patterns of relations among the members of the system (Wasserman & Faust 1994: 6–7). Network analysts, however, argue that “structural relations are often more important for understanding observed behaviors than are such attributes as age, gender, values, and ideology” (Knoke & Yang 2008: 4).
Social network analysis, then, concentrates on relations among the members of a system rather than on the individual attributes of those members. This is because patterns of relations have consequences. Again, this is the kind of everyday thinking we employ when we say that people looking for jobs should “activate their networks,” or that the “environment” in Silicon Valley is conducive to producing innovative thinking, or that the American economy suffered as the country lost its “position of dominance” in the system of global trade, or even that a friend going through a difficult time needs others to rally around with emotional support. In fact, one of the earliest examples of the type of work that we now think of as network analysis was Moreno’s 1934 study of how the “social configurations” surrounding individuals affected their psychological well-being (Moreno 1934). More generally, “[t]he central objectives of network analysis are to measure and represent these structural relations accurately, and to explain both why they occur and what are their consequences” (Knoke & Yang 2008: 4).
Because the relations among the members of a system can have profound consequences, it is important to understand how those relations fit together – how they are structured. Network analysts, therefore, gather data on the relations among the actors as their primary source of information. This relational data, Scott writes, is “the contacts, ties and connections, the group attachments and meetings, which relate one agent to another and so cannot be reduced to the properties of the individual agents themselves. Relations are not the property of agents; these relations connect pairs of agents into larger relational systems” (2000: 3).
Relational data is not about the individual members of a system (who have relations), but about relations (which occur among members of a system.) This difference may seem trivial (or even non-existent), but it is really a change in worldview, bringing the importance of the structure of relations to the fore in our understanding of how the world works. Focusing on the structure of the relations rather than on the attributes of the parties in those relations is the key to understanding network analysis. Wasserman and Faust note that “[o]f critical importance for the development of methods for social network analysis is the fact that the unit of analysis in network analysis is not the individual, but an entity consisting of a collection of individuals and the linkages among them” (1994: 4–5).
Networks as metaphors
Although the methods by which we can analyze relational data are relatively new, relational thinking has a long history in the social sciences. As far back as 1845, for example, Marx wrote in the Theses on Feuerbach: “VI: Feuerbach resolves the religious essence into the human essence. But the human essence is no abstraction inherent in each single individual. In its reality it is the ensemble of social relations” (Marx 1978 [1845]: 145). That is, Marx argues that it is our relations with others – our real, lived relations – that make us who we are.
In the early twentieth century, the work of Georg Simmel pushed the relational thinking behind network analysis to new heights. Simmel, one of the founders of the field of sociology, wrote prolifically on the relationship between individuality and social forms, especially during the transition to modern urban life. Simmel’s idea of the development of individuality (discussed in more detail in chapter 2) is based on a notion of the dynamic between social circles and individuals with each forming and being formed by the other. Simmel’s study of these social forms – which he referred to as the “geometry” of social relations – was the basis of “formal sociology.” The “forms of sociation” are made by individuals who are tied together in relations. Simmel particularly concentrates on exchange as the form of interaction through which society is formed; religion, economy, and politics, he argues, are all based on exchange. “Exchange,” he writes, “is the purest and most concentrated form of all human interactions in which serious interests are at stake” (Simmel 1971: 43). Social circles are themselves formed by these interactions – we are linked together through exchanges. We can see here the basis of the idea that networks are formed by relations between actors and that these networks have consequences. For Simmel, the creation of society itself is the result of these exchanges.
Simmel argues that exchange “lifts the individual thing and its significance for the individual man out of their singularity, not into the sphere of the abstract but into the liveliness of interaction” (1971: 69). This relational thinking, especially the primacy given to the role of exchanges between individuals, played a key role in the later development of network analysis. The “geometric” mindset brought relational thinking more clearly into focus.
But at first, Berkowitz notes, networks “were employed as little more than metaphors for the things social scientists were really trying to deal with: a friendship group was like a ‘star’ with one central point; a work group was like a small ‘pyramid’; or the spread of a rumor was like a ‘chain’” (1982: 2). While these metaphors helped social scientists conceptualize more clearly about the phenomena that they were studying, to truly be able to analyze in detail the spread of actual information through actual networks, social scientists needed more than just compelling metaphors; they needed method.
The relational thinking exemplified by Simmel was part of many disciplines in the social sciences, but it was an anthropologist, John A. Barnes, who is usually credited with first using the term “social network” in 1954 (Wasserman & Faust 1994, Knoke & Yang 2008) in his study of a Norwegian island parish. Barnes drew on Moreno’s work on “social configurations” and emotional well-being (mentioned above) and, specifically, on Moreno’s ground-breaking tool for analyzing these configurations: the sociogram (Moreno 1934). It was the advent of the sociogram that allowed network thinking to move from metaphor to method.
Metaphor into method
A sociogram1 is a picture of a network of relations, where the members of the network are represented by points and the relations between them are represented by lines connecting the points. The map in the back of an airline’s in-flight magazine showing the airports connected by the airline’s flights is an example of a sociogram, for instance. So is the organizational chart of a company showing the chain of command in decision making or a family tree showing kinship connections. Sociograms are familiar to us now – so familiar, in fact, that it is difficult for us to conceive of the revolutionary impact that Moreno’s work had in the social sciences. “Before Moreno, people had spoken of ‘webs’ of connection, the ‘social fabric’ and, on occasion, of ‘networks’ of relations, but no one had attempted to systemize this metaphor into an analytical diagram” (Scott 2000: 9–10).
Once the analytical diagram had been developed, it was available for analysis along a number of different lines. The mathematics for the analysis came from graph theory. Graph theory is the mathematical study of graphs, which (not to be confused with bar graphs, pie charts, and other types of graphs that we think of colloquially) are simple structures consisting of a set of vertices (represented by points in a sociogram) some of which are connected by edges (the lines connecting the points in a sociogram.) Edges may have additional characteristics such as direction (going from one vertex to another and not simply connecting two vertices) or color or weight. Topological techniques in graph theory yield results about the coarse properties of graphs, such as when they are connected, have certain special paths (Eulerian or Hamiltonian, for example), or can be colored with a certain number of colors. Probabalistic, linear algebraic, discrete geometric, and other, purely numerical techniques can answer questions about densities of edges (that is, the proportion of actual ties which exist in the network) in certain graphs, the rate of propagation of some kind of signal through a graph, or of the solutions of differential equations defined on a graph, for example. Often a graph is used as a simplified version of a complex situation (such as in geometry or differential equations) where results about the graph will give approximate answers to the full, original problem. Graph theory has been extensively used in computer science to model communications networks, the connections on a single computer chip, the relationships between the components of a large software system, and so on. Graph theorists can analyze not only a sociogram, but also the translation of a sociogram into a matrix (see below). Graph theory provided the mathematical method to analyze social networks.
What is a social network?
A social network is a type of graph – a set of vertices and edges. Or, less abstractly, a social network is composed of a set of actors and the relations among them. What, then, is an “actor” and what constitutes a “relation” between actors?
“Actors” can be any social entity that is engaged in interaction with others of its type – individual persons can be actors in a network, and so can small groups like families, larger groups like civic organizations, bigger groups like corporations or even nation-states. Actors can be much more than people sharing friendships mediated through a “social networking” site like Facebook. Some examples would be workers in a tailoring shop in Zambia who attempt to organize a strike (Kapferer 1972), composers working in the Hollywood film industry (Faulkner 1987), monks forming cliques in a monastery (Sampson 1969, White et al. 1976), families vying for political dominance in Renaissance Florence (Padgett & Ansell 1993), corporations connected by shared board members (Useem 1978), public and private agencies engaged in interagency collaboration to improve school safety (Cross et al. 2009), political parties forming coalitions (Centeno 2002), or nations importing guest workers from other nations (Massey et al. 2002).
Actors are all members of the system being analyzed, but they do not necessarily all have relations with each other. Some composers working in the Hollywood film industry may work together on projects repeatedly; some may never have any contact with each other. Some countries may be active trading partners; some may have no relations of any kind. Moreover, although network analysts use the term “actors” for the members of the network, that does not necessarily mean that they “act” or that they have agency. Family members in a kinship network “act” merely by being born. Some artists are “actors” in the art world by failing to get gallery representation – by being ignored and unable to form ties. Actors are represented as points or “nodes” in a sociogram. The lines that connect the actors represent their “relations”.
Actors are tied together by specific types of relations. These ties can be almost anything. When we think of social networking sites, we often think of individual people being tied together through friendship ties, but individuals could also be linked together by kinship, by belonging to an organization together (as in Useem’s [1978] corporate interlocks), by attending events together (as Davis et al.’s [1941] club women did), by disliking each other (as some of the monks in Sampson’s [Sampson 1969, White et al. 1976] monastery did), or by a whole host of other types of relations. On a larger scale, nations, for example, could be tied together by shared trade or diplomatic relations, by links of tourism, by sending and receiving guest workers and so on. Once we define what particular type of tie we are studying, we can then see which pairs of actors in the network are linked together by sharing a tie of that type (Wasserman & Faust 1994: 18–20). Two actors in a network that are tied together are called a “dyad” and the dyad is the most basic building block of a network.
Like the edges that graph theorists studied, ties can have properties such as direction and strength. A tie is directionless when it is mutual; I am related to my cousin in the same way that she is related to me, for example. But some ties have direction – they are sent from one node to another. An employer pays an employee and not the other way around. A boss directs an underling. Even ties such as those of friendship or affection may have direction – think, for example, of unrequited love. Ties can also have strength. I can be acquainted with someone, be friends with him, or be the very best of friends with him. Nations may engage in no trade with each other, engage in some trade with each other, or be primary trading partners.
It is important to remember that these properties are properties of the tie, not of the actor. It is only by looking at the flow of goods and money between two countries that we can tell if they are strong partners or not.
Further developments
Simmel’s formal sociology and the relational thinking that underlay it had an enormous impact on the development of social science in the twentieth century. Further advances were made in the 1950s by a group of anthropologists at Manchester University, including John Barnes (who first used the term “social network,” as mentioned above), Clyde Mitchell, and Elizabeth Bott (whose work will be discussed in much greater detail in chapter 3). These researchers concentrated on specific case studies and developed many network analytical concepts, tools, and terms to help them describe and explain the social structures that they uncovered. These empirical studies laid the groundwork for further methodological and theoretical developments that began to appear in the 1960s and 1970s – especially from Harrison White and his students at Harvard.
It is impossible to overstate the importance of White to the development of social network analysis. There were three especially important parts of White’s work: the development of algebraic methods for dealing with structure, the development of multidimensional scaling, and the training of a generation of important network researchers (Scott 2000, Berkowitz 1982.)
Using algebraic methods, White and François Lorrain developed a technique for identifying structurally equivalent nodes. (We will discuss structural equivalence in depth in chapter 6.) Using structural equivalence to reduce networks to models allows researchers to compare structures and positions across different networks. “Lorrain and White’s method was able to realize, for the first time, all of the power implicit in the social network concept. First, it operated simultaneously on both nodes and relations. . . . Second, it enabled researchers to deal with a given network at all levels of abstraction” (Berkowitz 1982: 5). The second technique, multidimensional scaling (which we will discuss in detail in chapter 8), is a method for mapping social distances onto geographic space. Like structural equivalence, multidimensional scaling allows researchers to build models based on the actual relations between actors in the network rather than having to impose a priori categories and understandings on the social world before beginning to analyze it. Network analysts now had methods by which they could compare positions across networks, compare network structures, build models to explain and understand action, and so on.
White was also responsible for training future generations of network researchers who produced some of the most important and ground-breaking network studies. For example, Granovetter’s study of the strength of weak ties (discussed in chapter 4), Lee’s study of women seeking illegal abortions (discussed in chapter 3), Wellman’s studies of urban community networks (discussed in chapters 3 and 8), and Breiger’s work on the duality of persons and groups (discussed in chapter 2) were all encouraged and informed by White. These are only a few of the brilliant researchers who began their work in network analysis with White.
This very brief overview is certainly not a full discussion of the development and usefulness of social network analysis, but has been restricted mostly to those aspects of it that are most important for this book. The shift in worldview that undergirds network analytical thinking – the shift from analyzing attributes to analyzing relations – has had an impact on every discipline in the social sciences, as well as in many fields in the natural sciences and the humanities. More complete discussions can be found in Berkowitz (1982), Collins (1988), Watts (1999a and 2003), Scott (2000), Barabási (2003), Freeman (2004), Azarian (2005), and Knoke and Yang (2008), just to name a few. The most in-depth and comprehensive book is Wasserman and Faust (1994).
Why use network analysis?
Network analysis provides an important and interesting lens with which to look at communities because it is concerned, most fundamentally, with the relations among the people and groups who make up those communities. It is these relations themselves which are at the heart of what it means to belong to a community – rather than to merely coexist in the same general geographic area as other individuals. The nature of the types of relations that we have with others affects how we behave, what we believe, how we understand the world around us and how we navigate through it, what constraints we labor under, and what avenues of opportunity are open to us.
The nature of our relations with others is fundamental to our lives. Network analysis allows us to examine those relations directly. By doing so, we can gain a clearer picture of human life in communities and a richer understanding of our world. We can also find surprising insights into questions that might otherwise seem intractable. The virulence of the Salem witch crisis, for example, seems mysterious and unexplainable – a bizarre aberration among the sober-minded Puritans. But when we look at the structure of the relations of the inhabitants of Salem Village during the end of the seventeenth century (as we will do in chapter 5), we will see how the underlying logic of the social networks at play made the tragedy of 1692 not only understandable, but almost inevitable.
In this book, we will see how using the perspective provided by social network analysis – focusing on the relations among people rather than on their individual characteristics – can give us new insights into the working of our communities. This deeper understanding may even help us be better able to have a positive impact on our own communities.
Plan of the book
This book examines the ways in which different types of community structures allow for different possibilities for individual and group actions – deviance and conformity, successful challenges to outside authority and failures, the emergence of innovation, etc. Chapter 2 will look at some of the fundamental building blocks of communities and examine the ways in which communities changed with the rise of modernity. Chapter 3 will address issues of communal support systems and the resources which communities provide to us. Chapter 4 focuses on the ways in which we are shaped by our communities and made to conform to communal expectations. Chapter 5 will turn to fractured communities and explore both why they break apart and what the consequences of that might be. Chapter 6, on the other hand, examines ways in which communities can come together and foster collective action among the members. Chapter 7 examines how different types of community structures either do or do not generate creativity and innovation. Finally, chapter 8 turns to the rise of computer-mediated communication and “virtual” communities to see if and in what ways they differ from more traditional communities.
This book uses network analysis to delve into the issues of community structure, formation, and dissolution. We will see how network thinking can be used to explore issues about communities in fruitful and interesting ways and discuss how applying a network analytical perspective to these questions allows new insights to emerge. At the end of each chapter, there is a section called “A Closer Look” that discusses in more depth one technical aspect of network analysis so that we can understand what the analysis is doing, as well as understanding why and in what circumstances this particular aspect of network analysis would be a useful tool.
Software
The development of network analytical tools was certainly facilitated by the increasing access to computers that researchers have had in the past half century, and there are many network analysis software packages that are commercially available, including UCINET, Pajek, NetMiner, STRUCTURE, MultiNet, and StOCNET (Knoke & Yang 2008: 2). In order to understand the “Closer Look” sections, though, it is not necessary to have access to any type of network analysis software. These sections are meant to explain how the various techniques work and in what situations a researcher would want to use them. However, most of the tools and techniques are complex enough to require computer power in order to actually use them for analysis. UCINET (Borgatti et al. 2002) is an excellent general purpose program that allows users access to all of the techniques covered in this book (and more). I will use UCINET to give examples of data analysis throughout the book.
A closer look: basic network terms and definitions
It is important to understand the basic terms and definitions used by network analysts to describe the world. First of all, a network is a set of relations between actors – and by actors we can mean individuals, groups, organizations, even entire nations. Friends in an informal friendship group can be actors in a network – so can nations in trade relationships. We call these actors nodes in the network. The important point is that the actors are in a situation or system where they may or may not have relations with each other.
These relations between nodes are called ties. Ties are the connections that actors have with each other. The content of the ties can be lots of things. You could have a tie of kinship with people, for example. These people are your relatives. Or you could have ties of affection or of enmity. You could have ties of the exchange of favors or loans of money. You could have ties of sharing memberships in clubs or organizations or of attending certain events together. Ties are the connections that the actors could have with each other. Not every member of a social network will be tied to every other member of that network by a particular type of tie – all that matters is that they could conceivably be tied together.
Ties can have direction. That is, you could have a tie of liking someone who does not like you back. Lines of authority usually have direction. Other ties are directionless, such as ties of kinship or shared membership in organizations. Ties of “knowing” are often directionless.
Figure 1.1 Sociogram: nodes and ties
Ties also have strength. There are a couple of different ways in which we can think about the strength of ties. One way is by the intensity of the relationship. You could love someone or you could like them or you could even just merely tolerate them without feeling too much about it. Or we could measure the strength of a tie by how many different types of content that the tie contains. For example, you might be roommates with someone and that is the full extent of your relationship. We would call that tie uniplex, meaning that it has only one dimension. Or you might be roommates with someone and also be that person’s friend and take classes together and work at the same job and even be related to that person. This would be a multiplex relationship and might be stronger than the uniplex relationship.
Nodes and ties can be put together to show a map of relationships called a sociogram. Figure 1.1 is an example sociogram. The circles represent the nodes (each of which is given an identifying letter as its “name”) and the lines represent the ties. Dark lines are strong ties, dashed lines are weak ties, and regular lines are ties of medium strength. Ties with an arrow on one end of them indicate a tie with direction, so that although A likes B in this example, B does not return the favor (if this is a sociogram of “liking.”) Ties that do not have arrows on them are directionless ties, which could mean that they are reciprocated (C likes D and D likes C back) or that the type of tie (such as being cousins) does not lend itself to having a direction.
Table 1.1 Matrix of figure 1.1
Sociograms are a good way of looking at the picture of the social structure, but they are not easy to analyze, especially when they get very large. Imagine a sociogram of the political networks in the US House of Representatives. It would be a tangled, unreadable mess. A better form to use for the analysis of the network data is a matrix.
A matrix is a grid of numbers arranged in rows and columns. Table 1.1 is the matrix translation of the sociogram in figure 1.1. The rows and columns are labeled with the names of the actors in the network. We will name this matrix “M.”
To understand how to read the information presented in the matrix, we need to know a few more terms. First, the matrix is arranged in rows and columns. Rows are read from left to right and columns are read from top to bottom. The intersection of each row and column is called a cell and each cell has a value. The values in the example matrix here are either 1 or 0, depending on whether or not the two actors whose row and column intersect in that cell have a tie or not. If the actor in the row of the cell sends a tie to the actor in the column of the cell, the cell value is 1. Otherwise, it is 0.
If we wish, we can indicate the strength of the ties in the matrix. Instead of using 1s and 0s to merely indicate the presence or absence of a tie between two nodes, we can use the cell values as a measure of strength. If the ties are multiplex, the values in the cells could, for example, be the number of different types of relations that the nodes share. Or suppose we were measuring trade relations among nations; the cell values could be the annual amount of exports between pairs of nodes expressed as dollar amounts. Or we could use a simpler metric merely indicating “strong,” “medium,” or “weak” ties. In table 1.2, we can see the matrix translation of figure 1.1, but now the cell value is 3 if the tie is strong, 2 if the tie is medium strength, or 1 if the tie is weak. Nodes which have no tie between them are still coded 0.
Along with having a value, each cell also has an address. It is important to remember that the matrix is always read in the order of row and then column. The cell addresses follow this pattern as well. The cell whose address is M2,3 is the cell that is in the matrix that we have named M and is the intersection of row 2 and column 3. The cell whose address is M is the cell in the matrix that we have named M3,2 and is the intersection of row 3 and column 2. You can see that in the example of table 1.1, those cells have different values. This is because this example matrix is not symmetric; some nodes send ties to others in the matrix and those ties are not reciprocated. B, for example, sends a tie to C, but C does not send a tie to B. When noting ties that have direction in a matrix, the actors who are listed as the headings in the rows are always considered the “senders” of the ties and the actors who are listed as the headings in the columns are always considered the “receivers.” So we can read across the B row and see that there is a 1 in the C column, indicating that B has sent a tie to C. But when we read across the C row, we see that there is a 0 in the B column indicating that C has not chosen to send a tie to B.
Table 1.2 Matrix of figure 1.1 with tie strength
The main diagonal of the matrix is the cells where the sending and the receiving node are the same. That is, M1,1, M2,2, M3,3, M4,4, and M5,5 in table 1.1 are the cells of the main diagonal. As you can see, the values of each of these cells is 0. That is because we did not allow nodes to have ties to themselves in this example. You could imagine, however, situations in which nodes would be tied to themselves. Suppose, for example, the nodes were the attendees at a cocktail party. We could imagine the host circulating among them with a tray of hors d’oeuvres, handing out some to his guests but also eating some himself. The tie would be receiving an hors d’oeuvre; some ties would be from the host to his guests and some would be from the host to himself. When nodes have ties to themselves, those ties are usually indicated in the sociogram by an arrow curving back to the node. In the matrix, the value of the main diagonal cell for that node would no longer be zero. Figure 1.2 and table 1.3 are the sociogram and matrix examples of this. As you can see, both B and D are now tied to themselves and the value of their cells in the main diagonal is 1.
Figure 1.2 Sociogram: nodes and ties, including ties to self
You may also notice that symmetric matrices have the same values in corresponding cells above and below the main diagonal, while asymmetric matrices, such as the one in figures 1.1 and 1.2, do not.
These are some of the basic terms with which we will be working in the following chapters. More terms will appear as we move through the concepts, but these are enough to get us started.
Entering and displaying matrices in UCINET6
It is quite simple to enter network data into a matrix in UCINET6 (Borgatti et al. 2002.) Once you are in the program, click on the “Matrix Spreadsheet Editor” icon at the top of the screen. (You can also access this editor by using the dropdown menu under “Data” and then “Data editors.”) This will open a spreadsheet into which you can enter your data just as it would be in a matrix like the ones we have been using in this chapter. Notice that you have the option of entering data in either “Normal” or “Symmetric” mode. If you are entering a square, symmetric matrix (where there are no directional ties), you can save time by choosing Symmetric mode and then you will only have to enter the data for each pair of relations into one cell. For example, in Normal mode, if A and B are tied together, you would need to enter a “1” in the A,B cell and also into the B,A cell. But in Symmetric mode, the UCINET program will automatically enter a “1” into the B,A cell for you once you have entered it into A,B. If you are not entering data into a square, symmetric matrix, use the Normal mode.
Table 1.3 Matrix of figure 1.2
The blue boxes at the top and side of the spreadsheet are for row and column labels. In a network of individuals, these would be their names or ID numbers. In most matrices, the labels are the names or IDs of the nodes, but in some matrices there may be other labels. For example, if you were analyzing a network made up of women attending social events together (as in Davis et al. 1941), the rows would be the names of the women and the columns would be labels for the events. This would not be a square matrix (unless by chance there happened to be the same number of women in the group as there were events held). The cell entries would be a “1” if the woman in the row had attended the event in the column and a “0” otherwise. We will discuss how to analyze these types of matrices in chapter 2.
When entering the data, you can save time and can perhaps work more accurately by entering only the 1s into your matrix. It is often much easier to note only the ties that exist. But you cannot properly analyze matrices that have empty cells. Use the “Fill” button at the top of the Spreadsheet Editor to fill all empty cells with 0s once you have finished entering your data. Trim the matrix to the appropriate dimensions by entering the number of rows and the number of columns in the boxes on the right-hand side of the Spreadsheet Editor, then click on an empty cell in your matrix. Then click on the “Fill” button at the top of the screen. This will automatically fill all empty cells with 0s.
Save your matrix by clicking on “File” at the top of the screen and choosing “Save as” from the dropdown menu. Give your dataset a name in the “File name” window and then click “Save.” Your dataset is now saved and you can close the Spreadsheet Editor. You can add more data to your matrix or make changes in it by opening it in the Spreadsheet Editor at any time in the future. Open the “Matrix Spreadsheet Editor” just as before, click on the “open file” icon and click on your data file to open it.
To see the matrix on which you want to work, instead of “Matrix Spreadsheet Editor,” choose the “display” icon and click on your data file. (Again, this function can also be accessed by using the drop-down menu under “Data.”) Note also that UCINET comes preloaded with data files that we can use as examples. These are stored in the “datafiles” file folder. (If you have accessed the Display function by using the dropdown menu under “Data,” just click on the “. . .” button next to the “Data Set Filename” window in order to get access to the list of datasets.) Double click on “datafiles” and scroll down to “sampson,” for instance, and double click on it. What you should see are a series of matrices showing the relations between the monks whom Sampson studied (1969). Matrix #1 shows the ties of “liking” between the monks at Time 1. Notice that this is a square matrix (with the names of the monks in the same order down the row labels and across the column labels), but that it is not symmetric. Some monks said they liked others who did not choose them back. The cell values also indicate the strength of the tie.
If you close the Sampson dataset and open “davis,” you will see a different kind of matrix. The rows are 18 women from Davis et al.’s study Deep South (1941) and the columns are 14 social events that the women could have attended. A “1” in a cell means that the woman in the row attended the event in the column; a “0” means that they did not. We will analyze this matrix in chapter 2.
When you are finished with your work and have saved it, you can exit UCINET by choosing the “Exit” icon from the top of the screen or by choosing “Exit” from the dropdown menu under “File.”
