Docstoc

FOR A SOCIAL NETWORK ANALYSIS OF COMPUTER NETWORKS

Document Sample
FOR A SOCIAL NETWORK ANALYSIS OF COMPUTER NETWORKS Powered By Docstoc
					                       FOR A SOCIAL                NETWORK              ANALYSIS          OF COMPUTER             NETWORKS:

                             A Sociological         Perspective        on Collaborative    Work and Virtual        Community

                                                                          Barry Wellman
          Centre for Urban and Community Studies, Univ. of Toronto, Toronto, Canada; well man @epas.utoronto.ca


                            ABSTRACT

     When computer networks link people as well as machines,                          directly with the outside). The distinction between group
     they become social networks. Social network analysis                             and network is important because it opens up consider-
     provides a useful approach to moving beyond the concept                          ation of how the characteristics of computer supported
     of “group” in studying virtual communities and computer                          social networks    affect the behavior    of the people using
     supported cooperative        work and telework.        Such computer             them and the social systems in which these networks are
      supported social networks         (CSSNS) sustain strong, inter-                embedded. My aim here is to show how social network
     mediate and weak ties that provide information               and social          analysis might    be useful for understanding      how people
      support in both specialized          and broadly-based        relation-         relate to each other through computer mediated communi-
      ships. They foster informal         workplace      communities      that        cation (see also Wellman      & Gulia    1996, Wellman,   et al.
      are usually partial     and narrowly-focused,         although     some         1996).
     do become encompassing        and broadly-based.   CSSNS                         Social network analysis conceives of social structure as
     connect workers within and between organizations who                             the patterned (organization of these network members and
     often are physically dispersed. The nature of the medium                         their relationships.   Social network    analysts work at
      both constrains and facilitates   management    control.                        describing underlying    patterns of social structure, explain-
      Although many relationships function offline as well as                         ing the impact of such social structures on other variables,
      online, CSSNS have developed their own norms and                                and accounting    for change in social structures. In the past
      structures. They have strong societal implications, foster-                     three decades., social network analysis has developed a
      ing situations that combine      global connectivity,  the                      range of concepts, findings and methods for detecting
      fragmentation of solidarities and the de-emphasis of local                      structural patterns, seeing how patterns of different types
      organization,    and the importance        of home bases.                       of relationships interrelate, analyzing the implications that
                                                                                      structural patterns for the behavior of network members,
                                                                                      and studying the impact on social structures of the charac-
                                 KEYWORDS
                                                                                      teristics of network members and their social relationships
      computer supported cooperative            work, informal       relation-        (Berkowitz 19:B2, Wellman 1988, Scott, 1991, Wasserman
      ships, telework,     electronic   mail, social networks                         & Faust 1993)1.
                                                                                      The social network approach provides ways for analysts to
       COMPUTER           SUPPORTED            SOCIAL       NETWORKS                  think about social relationships that are neither groups nor
      When a computer         network    connects people, it is a social              isolated duets. Instead of an either/or distinction between
      network. Just as a computer network is a set of machines                        group membership        and social isolation,   researchers can
      connected by a set of cables, a social network is a set of                      bring to bear in their analysis a set of structural variables,
                                                                                      such as the iiensity of a network, how tightly it is
      people (or organizations or other social entities) connected
      by a set of socially-meaningful     relationships.  Why do                      bounded, and whether it is diversified or constricted in its
      computer      scientists talk about “groupware”    when a                       size and heterogeneity,     how narrowly    specialized or
      computer     network supports all forms of social networks                      broadly multiplex are its relationships, and how indirect
      and not only groups? A group is only a special form of                          connections    and positions   in social networks affect
      social network, one which is densely-knit (most nodes are                       behavior. For example, the fact that Person A and Person
                                                                                      B interact online may have to be interpreted in the light of
      directly connected with each other) and tightly-bounded
      (most links stay within the network rather than connecting                      the offline repclrting relationship of Person B to Person C,
                                                                                      the company president. Thus thinking about relationships
 Permission to make digitrd/trard copies of all or part of this material for          in social networks rather than in groups can allow analysts
 personal or classroom use is granted without fee provided that the copies            to take into account the contexts within which relation-
are not made or diwibmed        for profit or commercial advantage, the copy-
 right notice, the title of the publication and its date appear, and notice is        ships operate.
given that copyright is by permission of the ACM, [nc. To copy otherwise,
to repubtish, to post on servers or to redktribute to lists, requires specific        The distinction between groups and social networks opens
permission and/or fee,                                                                up consideraticm of how the characteristics of computer
SIGCPR/ SIGMIS ’96, Denver Colorado USA
@ 1996 ACM o-89791-782-O/96/04.               .$3.50

                                                                                  1
supported social networks affect the behavior of the people                  q
                                                                                       How does participation               in computer networks affect
using them and the social systems in which these networks                              patterns of social relationships                 in social networks.
are embedded. (Although both “group” and “network” are                       My       objectives        are to discuss how               computer          networks
really social networks, the linguistic distinction is well-                  support a variety of social relationships and social net-
established in everyday discourse.) Just as a local area                     works at work and in the community. I examine the extent
network is only one kind of a computer network, a group                      to which people work and find community on computer
is only one kind of a social network. More precisely, a                      supported social networks (CSSNS), and I show how some
group is asocial network whose ties are tightly-bounded                      of the key concepts of social network                             analysis apply to
within a delimited set and are densely-knit so that almost                   understanding          relationships          on-line.     What is the network
all network members are directly linked with each other.                     composition   and structure of CSSNS like, with their
To be sure, there are densely-knit   and tightly-bounded                     weaker constraints of distance and time, their easy connec-
work groups and community groups. Yet there are other                        tivity    and limited          social presence?
kinds of work and community networks whose relation-
                                                                             I focus on the implications for computer supported social
ships are sparsely-knit    with only a minority       of members of
                                                                             networks of two opposing ideal types: dense, bounded
the workplace or community directly connected with each
                                                                             groups and sparse, unbounded                      networks.         Although        these
other. These relationships tend to ramify out in many
                                                                             are often seen as opposite ideal types, a good deal of
directions like an expanding spider’s web rather than
                                                                             research has demonstrated that in contemporary society,
curling back on themselves into a densely-knit tangle.
                                                                             both      dense,      bounded         groups       and      sparse, unbounded
For example, people who interact together -- at work, in                     networks         exist simultaneously            at work and in communi-
a cafk or on an internet discussion group -- can be studied                  ties. Indeed, the same persons may be involved                              in both, as
as either a group or a social network. Those who study                       they iterate between different workgroups and communi-
them as groups assume that they know the membership                          ties, or as the workgroups and communities themselves
and boundaries of the groups. They might ask how                             change in response to external situations and internal
important each group is to its members, how the groups                       dynamics. Nevertheless, there are systematic ways in
are governed      and make decisions,        and how the groups              which involvement in dense, bounded groups of work and
control members. Yet in all but laboratory       situations                  community    are different from involvement  in sparse,
researchers will be faced with the real-world problem that                   unbounded          networks         in which        the active players               shift
members are entering and leaving a group over time. By                       frequently.
contrast, those who study such entities as social networks
                                                                             Casting analyses of CSSNS in social network terms gives
can treat their membership      and boundaries    as open
                                                                             the benefit of linking   studies of CSCWS and virtual
questions. For example, frequent participation       in an
                                                                             communities   to a century’s research into the nature of
internet discussion group might be treated as the basis for
                                                                             community; work and social organization. This provides
membership       but so might be the indirect connections         (and
                                                                             an antidote to the parochial, unscholarly assumption of
resource flows) that discussion group members provide to
                                                                             some computer scientists that life online has no connection
others outside     the group.       The patterns     of relationships
                                                                             to life offline.       Despite the inherently               social nature of the
becomes a research question rather than a given.
                                                                             subject,      sociological         research has not been well                       repre-
By definition,     people who use computer            networks   have        sented.
social relationships with each other that are embedded in
social networks. Yet much of the analysis of online
relationships via the study of “computer-mediated   commu-                                         RELATIONSHIPS                      ONLINE
nication” has focused on two-person dyads rather than on                     Social      Support
the broader social networks in which these dyads are
                                                                             As is the case for relationships  in “real life,” social
connected. People’s relationships with other strongly affect
                                                                             support flows through CSSNS, with much support involv-
their social resources, mobility, happiness, work habits and
                                                                             ing the exchange of information.      The nature of the
many   other   important   things    about   them.   Hence   it makes
                                                                             medium supports a focus on information     exchanges, as
sense to use social network analysis to understand the
                                                                             people can easily post a question or comment and receive
patterns of relationships that people have online in addi-
                                                                             information        in return.     Broadcasting           queries    through     CSSNS
tion to fine-grained analyses of online dyads. In thinking
                                                                             increases        the chances        of finding      information          quickly       and
about how to use the social network approach to study
                                                                             alters     the    distribution       patterns      of     information.         It    gives
patterns of interaction online, two linked questions are
                                                                             those      working      in small       or distant         sites    better     access    to
especially pertinent:
                                                                             experienced,         skilled     people.      The widespread          dissemination
e      How do different social network patterns of relation-                 of advice        provided        informally       through         CSSNS       may     also
       ships affect interactions, online and offline.                        threaten      professionals’        claims      for monopolies           on expertise
                                                                             (Foderaro         1995).



                                                                         2
The flow of information          through     CSSNS itself generates     Specialized             and Multiplex         Ties
access to new information.         Online information     flows spill   The nature c~f CSSNS encourages                      both specialized        and
over unexpectedly         through message forwarding,      providing    multiplex       relationships.        The structure of online networks
access to more people and new social circles, thus increas-             encourages specialized relationships                  because it supports a
ing the probability       of finding     those who can solve prob-      market apprctach to finding social resources in virtual
lems (Kraut & Attewell     1993). People often bump into                communities. With more ease than in almost all real life
new information or new sources of information       uninten-            situations, people can shop around for resources within the
tionally through “leaky ... quasi social networlds” (Brent              safety and comfort of their homes or offices, and with
1994: online). Information  obtained serendipitously helps              reduced search and travel time. The Usenet alone houses
solve problems      before they occur and helps keep people             more         than     3,500     newsgroups     (Kling     1995)       to which
aware of organizational news. Weak online ties are bridges              anyone may subscribe, with diverse foci including politics,
between diverse sources of information.                                 technical problems, therapeutics, socializing and recre-
If CSSNS were solely a means of information                exchange,    ation. The narrow focus of newsgroups,                        distribution   lists
then they would mostly contain                  narrow, specialized     and chat lines allows people to take risks in specialized
relationships. However, information            is only one of many      relationships that may only exist in a single, partial online
social resources exchanged online. Despite the limited                  group. Some CSSNS even allow people to be anonymous
social presence of CMC, people find emotional support,                  or use nicknames when they want to speak freely or try on
instrumental aid, companionship and a sense of belonging                different      personas (Hiltz         & Turoff      1993).
through the normal course of CSSNS of work and commu-                   In addition to the online support of specialization, the
nity, even when they are composed of persons they hardly                inclusion of email addresses in most message headers
know (Wellman        & Gulia 1996). Although         providing   such   provides the basis for more multiplex      relationships to
types of support often does not require major investments               develop. In the absence of social and physical cues, people
of time, money or energy, CSSN members have also                        are able to get to know each other on the basis of their
mobilized    goods, services and long term emotional         support    communication and decide later to broaden the relationship
to help each other (e.g. Lewis 1994). For example, an                   or move it offline (Rheingold 1993). Thus more than half
informal support group sprang up inadvertently when the                 of the recovering addicts on electronic support groups also
“Young Scientists’ Network” aimed primarily at providing                contact each other by phone or in-person (King 1994).
physicists with job hunting            tips and news stories. Simi-     Soon after an especially intense computerized conference,
larly, the “Systers” mailing            list, originally designed for   many “of the ptiicipants       altered their business and
female computer scientists to exchange information, has                 vacation       travel     plans to include        a face-to-face        meeting
become a forum for companionship     and social support                 with one another” (Hiltz and Turoff 1993: 114). This
(Sproull & Faraj          1995). The members of a computer              expansion of online ties to offline relationships is often
science laboratory        exchange a good deal of emotional             ignored in CSCW research
support by email.         Because much of their time is spent
online, it is natural for them to discuss problems with
online coworkers and for their coworkers to respond via                 Strong         Ties
email to these on-screen messages (Haythornthwaite,                     Can the medium support the message? Or as some pundits
Wellman & Mantei 1995).                                                 have feared, does the limited social presence of CMC
Some CSSNS go beyond the serendipitous exchange of                      work against the maintenance of socially-close,                      strong ties
online support to be explicitly-established support groups              on CSSNS? M:any online ties do meet most of the standard
for the provision of emotional aid, group membership, and               criteria for strong ties (Wellman & Wortley 1990). They
information  about medical treatment and other matters                  facilitate     frequent,        reciprocal,   companionable          and often
(Foderaro 1995, King 1994). For example, Peter and                      supportive and multiplex contact. Moreover, the placeless-
Trudy Johnson-Lenz (1990, 1994) have organized online                   ness of CSSNI interactions facilitates long term contact
groups for twenty years, working to build self-awareness,               without the loss of relationships that often accompanies
mutually supportive activities, social change, and a sense              residential         mobility.    Virtual   communities        are quite volun-
of collective well-being.   Their software tools, such as               ta~, while CSSN participation varies between voluntary
passing around sacred “talking sticks,” rearrange commu-                and mandatory in CSCW and telework (Hiltz & Turoff
nication    structures,    vary exchange      settings, mark group      1993, Johnsonl-Lenz               & Johnson-Lenz          1994, Rheingold
rhythms and encourage lurkers to express themselves.                     1993). Certainly there are many stories of great involve-
                                                                        ment in online relationships. Community members came
                                                                        to regard each other as their closest friends even though
                                                                        they seldom or never met in-person (Hiltz & Turoff 1993).
                                                                        Network members tend to base their feelings of closeness
on shared interests rather than on shared social characteris-                    expertise, earning respect and status, and responding to
tics such as gender and SES. That the siren call of CSSNS                        norms of mutual aid (e.g. Constant, Sproull & Kiesler
sometimes lures participants away from “real-life” argues                        1994, Kraut & Attewell       1993, Kollock      & Smith 1996). In
for    the potential     strength      of   online    relationships       and    some organizations,       employees        are encouraged    to help
networks.                                                                        each other or to direct those in need to others who could
                                                                                 help.  Such processes also arise in densely-knit virtual
In addition to the truly strong ties online, many ties are
                                                                                 communities and are common among frequent contributors
what Wireman (1984) has called “intimate        secondary
                                                                                 to computerized  conferences. People having a strong
relationships”: moderately strong, frequent and supportive
                                                                                 attachment   to an electronic     group will     be more likely    to
but operating only in one specialized domain. Over time,
                                                                                 participate and provide assistance to others. For example,
some of these relationships become more personal and
                                                                                 computer hackers involved in illegal activities are reluctant
intimate. Perhaps the limited social presence and asynchro-
                                                                                 to change their pseudonyms because the status they gain
nicity of CMC only slows the development of intimacy,
                                                                                 through online demonstrations   of technical                expertise
with    online   interactions       eventually     developing    to be as
                                                                                 accrues to that pseudonym (Meyer 1989).
sociable and intimate        as in-person        ones (Walther     1995).
                                                                                 Some commentators        have warned about the consequences
In part, concerns about whether online ties can be strong
                                                                                 of making connections on CSSNS teeming with strangers
ties are wrongly       specified.    Although      CSSNS do transcend
                                                                                 whose biographies, social positions and social networks
time and space, not all ties are either totally                  online    or
                                                                                 are unknown     (Stoll   1995). Nevertheless       CSSN members
offline. Much online contact is between people who see
                                                                                 tend to trust strangers, much like people gave rides to
each other in person and live locally. At work, computer
                                                                                 hitchhikers in the flowerchild  days of the 1960s. This
scientists intermingle in-person and email communication,
                                                                                 willingness to engage with strangers online contrasts with
At some offices, employees chat privately by email while
                                                                                 in-person situations where bystanders are often reluctant
they work silently side-by-side (Garton 1995). In such
                                                                                 to intervene and help strangers (Latan6 & Darley 1976).
situations, conversations started on one medium continue
                                                                                 Yet bystanders are more apt to intervene when they are
on others. As with the telephone and the fax (Wellman &
                                                                                 the only ones around and they can withdraw            easily in case
Tindall 1993), the lower social presence of CMC may be
                                                                                 of trouble. Analogously,  online requests for aid are read
sufficient to maintain          strong ties between         persons who
                                                                                 by people alone at their screens. Even if the request is to
know each other well.
                                                                                 a newsgroup and not by personal email, as far as the
                                                                                 recipient of the request knows, slhe is the only one who
Weak Ties                                                                        could provide aid. Individual acts can aggregate to sustain
Although     online ties are strong, it is nevertheless true that                a large community        because each act is seen by the entire
CSSNS are excellent          supporters of weaker ties. There are                group and perpetuates a norm of mutual              aid (Rheingold
low logistical and social costs to participating in CSSNS,                       1993, Lewi{ 1994).
People can participate within the comfort and safety of                          By contrast, most of the anti-social behavior online has
their own homes or offices, at any time and at their own                         consisted of rude, uninhibited remarks and hostile “flam-
convenience.  Limited   social cues online encourages                            ing” (Hiltz, Johnson & Agle 1978, Kiesler, Sproull &
 contact between weak ties. Very often, the only social                          EccIes 1985, Siegal et al 1986, Sproull & Kiesler 1991,
 characteristic that people learn about each other online are                    Lea & Spears 1992, Walther 1994). The limited social
 their email addresses which provide very little information.                    presence of CMC encourages the misinterpretation        of
 The egalitarian nature of many CSSNS encourages re-                             remarks, the isolated and anonymous possibilities       of
 sponses to requests. It also generates a culture of its own,                    communication       hinder    the social control      of anti-social
 as when humorous           stories sweep CSSNS, possibly foster-                 behavior, while the asynchronous nature of most conversa-
 ing a revival    of folk humor.                                                  tions hinders the immediate        repair of damages, stressing
 On the face of it, CSSNS should not support much                                 and even disrupting      relationships.
 reciprocity. Many online ties are between persons who
 have never met face to face, are weakly tied, socially and
                                                                                               SOCIAL       NETWORKS            ONLINE
 physically distant, and not bound into densely-knit work
                                                                                  Range
 or community structures. Computerized conferences allow
 free-rider “lurkers”  to read others’ messages invisibly                         The range of a network describes how large and diverse
 without    contributing.       Nevertheless,        there is evidence      of    is the population within its boundaries. Although contem-
 reciprocal supportiveness on CSSNS, even between weak                            porary people in the western world may know one thou-
 ties (Hiltz, Johnson & Turoff     1986, Walther    1994).                        sand others, they actively maintain only about twenty
 Providing reciprocal support and information online is a                         community ties (Wellman 1992). With larger size comes
 means of increasing         self-esteem,        demonstrating    technical       the population basis for more heterogeneity in the social
characteristics      of network       members        (Wellman       & Potter        line work team formed more subcommittees                 than an off-
1996) and more complexity              in the structural patterning        of       line team and was better able to involve its members in its
ties in networks. Dense, bounded networks almost always                             activities   (Bikson    & Eveland       1990).
have a small range because a large network becomes                                  Where the climate fosters open communication,                the lack
unbounded relatively quickly. As the number of network                              of status cues fosters connections           across hierarchical     or
members increases the population basis for more diversity                           other forms of status barriers           (Sproull   & Kiesler    1991,
within the network increases. Moreover, asthe number of                             Eveland & Bikson 1988). The relative lack of social
network members increases arithmetically,     the number of                         presence online fosters relationships with network mem-
connections required to sustain full connectivity increases                         bers who have more diverse social characteristics               than is
geometrically.                                                                      normally encountered in person. It also gives participants
Where dense, tight networks with small range are good for                           more control over the timing and content of their self-
conserving existing resources, sparse, loosely-bounded                              disclosures    (Walther     1995). This allows       relationships   to
networks       with a large range are good for obtaining                            develop      on the basis of shared interests rather than be
additional       resources. Sparsely-knit, loosely-bounded                          stunted at the onset by differences in social status (Hiltz
networks with a large range provide the number of                                   & Turoff 1993). The phenomena is a technologically-
network members and the diversity for people to obtain a                            supported continuation   of a long term shift to social
wide range of resources. Such large and heterogeneous                               networks organized by shared interests rather than by
networks       tend to be structurally         complex.       Instead of a          shared locality (Wellman    1992). When a computerized
uniform distribution   of relationships (as in densely-knit                         conference focuses on concerns that are held dear, physi-
networks), ties in networks with a large range tend to form                         cally-distant members may have more in common than
clusters that only a few ties bridge                 (Granovetter     1973).        they do with their neighbors. There is strong commitment
Hence,     resources       (such    as information)        tend     to flow         to online      groups     when   they     are seen as Ionglasting
unevenly       through these networks          and the network        mem-          (Walther     1994), However, heavy involvement          in a comput-
bers whose ties are bridges tend to accrue both power and                           erized conference can nourish CSSNS with homogeneous
suspicion as gatekeepers to other clusters.                                         interests (Lea & Spears 1992). Moreover,            the quite similar
The architecture of CSSNS may nourish two contradictory                             social characteristics   of those currently   on a CSSN
                                                                                    constrains the diversity of the participants encountered.
trends for the range of virtual communities. First, CSSNS
foster membership in multiple, partial communities. People                          The proliferation    of virtual communities online may
often belong to several computerized      conferences, and                          produce a counter-trend to the contemporary privatization
they can easily send out messages to separate personal                              of community. People in the western world are spending
distribution      lists   for   different    kinds     of conversations.            less time in public places waiting         for friends to wander by
Moreover,      they can vary in their involvements              in different        and to introduce friends to other friends (Wellman              1992).
communities, participating actively in some and occasion-                           Community has moved indoors to private homes from its
ally in others. Second, the ease of responding to entire                            former semi-pulblic, accessible milieus such as cafe%,parks
groups     and forwarding          messages to others            foster   the       and pubs. This dispersion         and privatization      means that
folding    in of online networks            into broader communities.               people must actively contact community         members to
Some intense role-playing environments resemble village-                            remain in touchl instead of visiting a caf6 and waiting for
like structures when they capture their members’ attention.                         acquaintances to drop by. By contrast, computerized
Third, the asynchronous nature and distance-free cost                               conferences support connections with large numbers of
structure of computer           networks     transcends spatial limits,             people, providing possibilities for reversing the trend to
allowing     people to communicate           over different     time zones          less public contact. Because all members of newsgroups
and enabling more active contact in otherwise latent                                and discussion groups can read all messages -- just as in
relationships. Fourth, computer networks enable people to                           a caf6 conversation -- groups of people can talk to each
maintain      a larger number of ties on-line than they proba-                      other casually and get to know the friends of their friends.
bly   could     face-to-face.      Computerized         conferences       and       “The keyboard is my caf6,” William       Mitchell enthuses
distribution lists provide permeable,  shifting sets of                             (1995: 7). Moreover, each participant’s personal social
members, with more intense relationships continued by                               network connects specialized, partial social networks,
private email. Fifth, the resulting expansion of the size and                       providing cross-cutting links between otherwise discon-
proximity of one’s on-line “communication    audience” can                          nected groups.
increase the diversity of people encountered (Lea &
Spears 1995). For example, four-fifths of the email of one
large, physically dispersed organization were from elec-
tronic groups and not individuals (Finholt & Sproull 1990,
Sproull & Kiesler         1991). In another organization,            an on-



                                                                                5
Density                                                                            Boundedness
Densely-knit       social networks are those in which almost all                   Although      all studies have to start somewhere with some
direct links between network members actually exist.                               populations, many network analyses do not treat formal
When they are tightly bounded (see below), they are                                group boundaries as truly social boundaries, be they
groups. Such dense, bounded                groups have considerable                departments       in     organizations          or    officially-designated
communication            among members.         Within    the group, ease          neighborhoods          in cities.     Instead        they trace the social
of access promotes connections      among all members.                             relationships    of those they are studying, wherever these
Almost all contacts between all persons in the network are                         relationships    go and whomever they are with. Only then
made and usually made frequently.       By contrast, few                           do network       analysts look to see if such relationships
members of sparsely-knit,    loosely-bounded  networks                             actually cross formal group boundaries. In this way,
communicate directly and frequently with each other.                               formal boundaries become important analytic variables
                                                                                   rather than a priori         analytic constraints.
Computer       mediated communication     supports both dense,
bounded        groups   and sparsely-knit,    loosely-bounded                      Boundedness refers to the proportion                  of network members’
networks.      Dense groups are supported when all partici-                        ties that stay within        the boundaries of the social networks
pants in computerized          conferences read and respond to all                 (Laumann,       Marsden       & Prensky         1983). Networks          can be
communications           and so are directly          connected      to each       bounded groups or permeable ramifying                    networks in which
other on-line.      Focused task groups, role-playing                MUDS          people can reach out widely to connect with others. In
(multi-user      dimensions) and some newsgroups resemble                          tightly-bounded networks (almost) all of the relationships
village-like     structures when they capture their participants’                  remain within the population. Those that cross the popula-
attention (Kollock   & Smith 1996). Thus work groups                               tion’s boundaries are likely to be relationships maintained
using computer mediated communication        have a higher                         by a few gatekeepers, such as workgroup managers, who
level of communication  than those that do not (Bikson &                           have routinized        contact on behalf of the group with the
Eveland 1990), although a high level of on-line communi-                           rest of the organization,    neighborhood, etc. Tightly-
cation may reduce the use of face-to-face                     and telephone        bounded groups better control the flow of information,
contact    (Finholt       &   Sproull    1990,      Finholt,     Sproull   &       disease and social resources, as Laumann and associates
Kiesler 1990). Moreover,           the forward-and-copy          features of       (1994) found          with   respect to the spread of AIDS                    in
most CSSNS can increase network density. Forwarding                                America.
communications to third parties provides indirect connec-                          Computer        networks       support     dense,        bounded     groups,
tions between previously-disconnected      people, as they                         through    computerized        con ferencing          and distribution     lists
become aware of each other’s mutual interest. The ease of                          in which all can send and receive messages with all.
reply can transform an indirect tie to a direct tie.                               Indeed, many organizations put tight boundaries around
Densely-knit,       on-line    groups are also apt to be mutually                  their computer networks to inhibit communication    from
supportive, with exchanges of help often forming      a                            leaving the organizational     precincts and to forestall
complex web of assistance among several group members.                             intrusion. Technical constraints may also reinforce tight
Motivations        for    providing     assistance       on    a computer          boundaries.      At     present,      multimedia         systems    such      as
network       are partially    founded     on norms of generalized                 Telepresence, only have the ability to support videoconfer-
reciprocity      and group       citizenship.       People     who   have a        encing among a small set of participants who also have
strong attachment to the on-line group will be more likely                         the necessary equipment. Participants are aware of who is
to participate  and assist others, even total strangers                            available for interaction through slow-scan video snapshots
(Kollock   & Smith 1996, Constant, Sproull & Kiesler                               and can then use computer commands to establish video
forthcoming).                                                                      links with another person or even a small group (Mantei,
                                                                                   et al. 1991, Buxton          1992).
Computer networks are well configured to support partici-
pation in sparsely-knit, loosely-bounded networks. People                          By contrast, the members of loosely-bounded       networks
can quickly send private email to anyone whose address                             have many ties with people who are not members of this
they know, they can belong to a number of computerized                             particular network. Their orientation to a network will not
conferences, and they can easily send messages to separate                         be as intense. Because so many ties go outside the
personal distribution         lists for different     kinds of conversa-           network, it is likely that the network will be sparsely-knit.
tions and activities. Moreover, they can vary their involve-                       For example, purchasing agents and sales people often are
ment in different work groups and communities, participat-                         on the organizational margins, developing strong bonds
ing actively in some and occasionally in others.                                   with people they buy from or sell to, at the cost of their
                                                                                   loyalty to their own organization (Bristor 1987, Dorsey
                                                                                    1994). Thus CSSNS link groups as well as people, for
                                                                                   when ties connect two groups, they provide intergroup as
                                                                                   well as interpersonal links. (Breiger 1974, calls this “the



                                                                               6
duality      of persons and groups”).           Moreover,       networks       are        groups, . . . geographically dispersed yet coordinated”
scalable as “networks of networks” (Craven and Wellman                                    (Carley & Wendt 1991, p. 407). They also blur organiza-
1973): interpersonal, intergroup, interorganizational, and                                tional   and locational        boundaries,      supporting        “invisible
international.          Thus if two persons who are members of                            colleges”     of dispersed professionals        (Constant, Sproull &
two different       workgroups        or organizations       are linked, their            Kiesler 1994, Eveland & Bikson, 1988, Hesse et al 1993,
interpersonal tie is also an intergroup                or interorganizational             Hiltz & Turoff 1993, Kling 1996, Meyer 1989, Carley
one. The logic            and the analysis             are quite    similar      to       1990, Kaufer & Carley 1993, Huff et al 1989, Kaye 1992,
computer networks:     Connectivity  between local area                                   Rice & Steinfield 1994, Walsh& Bayama 1996, Steinfield
networks creates an organizational network, while connec-                                 1986, Sproull & Kiesler, 1991 Walsh & Bayama 1996).
tions to the internet support inter-organizational                      computer          The blurred       or shifted      organizational,      work      group and
networks.                                                                                 community boundaries supported by computer networks
Many computer networks                support loosely-bounded               social        aids organizations to develop work teams that span unit
networks because both private email and computerized                                      boundaries,      facilitates   interorganizational      coordination       of
conferences makes it easy and inexpensive to maintain                                     joint projects, easily link buyers and sellers in different
connections with distant acquaintances and form new ties                                   organizations, and fosters nonlocal communities.      They
with strangers. Such ties are usually                   better connected to                help people to maintain a stockpile of potentially useful
other, more diverse social circles and hence are more apt                                 contacts      outside   of their     work     group,    organization       or
to be sources of new information                       (Granovetter   1982).              community       that can provide information,           instrumental      aid
Information may come unsolicited                       through distribution               and emotional       support. For example, more than half of the
lists, computerized         conferences, forwarded            messages from               email messages in one organization               were from unknown
friends who “thought           you might like to know about this,”                        people, different buildings, or people external to the work
and direct email from strangers (“a mutual friend recom-                                  group or the chain of command (Finholt & Sproull 1990,
mended that I contact you”). The lower social presence of                                 Sproull & Kicsler        1991).
email -- as compared to in-person meetings or telephone                                   Shifting boundaries not only characterize networked virtual
conversations -- makes it easier to contact strangers                                     organizations but links between organizations. Interorgani-
because there is less concern about rude intrusion or                                     zational CSSNS can help an organization in negotiations
interpersonal risk (Stoll 1995).                                                          between       buyers    and sellers     and in coordinating            joint
In many organizations,               work outcomes depend more on                         projects.     They also help managers and professionals
the ability       of people      and groups            to bridge        cognitive         maintain      a large network of potentially useful contacts,
distances than on people and other resources being located                                stockpiling     network capital for when they need to obtain
together (Mowshowitz      1994). Such organizations     are                               information     externally. It also helps employees to maintain
managed by network             and not by matrix.            They are based               a sense of ccmnection with former                   colleagues     and can
on multiple  reporting relationships and shifting work                                    provide support during job changes.
groups, each consuming only a part of a worker’s time.                                    At the extreme is the internet, an unbounded network
From an organizational perspective, such dispersed work                                   whose population boundary approaches infinity and is so
teams require social as well as technical support (Wellman                                in flux that it cannot be analyzed at any one time. Al-
et al 1994, Garton 1995). Studies of collaboration                        among           though it is inherently           impossible to map all internet
scientific occupational communities suggest that an initial                               relationships (or Web              site hyperlinks), ego-centered
period of physical proximity is necessary to build trust and                              analyses can trace the nature of the connections                       of a
come to consensus on the focus of the proposed project                                    sample of internet users. Another way to study the internet
(Carley and Wendt            1991). Such collaborations             may need              (and other      unbounded         networks)    is to trace flows           of
different forms of CMC support at different points in the                                 resources. For example, Dantowitz       and I (1996) are
project. For example, work groups tightly focused on a                                    currently replicating Milgram’s “small world” study ( 1967)
single project need different types of CSCW support than                                  by asking randomly-selected internet users to try to contact
individuals switching among multiple tasks and relation-                                  a target person whom they may or may not know. If they
ships (Mantei and Wellman 1995).                                                          do not know the person, they are asked to forward our
Social processes can make                    loosely-bounded   networks                   message to someone in their social network and also copy
develop tighter boundaries.                   Computer    networks  can                   the message to, us. This is enabling us to trace the flow of
integrate new workers into communication                        channels and              messages through the internet, seeing how close two nodes
culture and increase commitment         (although an initial                              are to each other in terms of the links needed to connect
period of physical proximity may be necessary to build                                    them     see also Schwartz        & Wood      1992)
trust and consensus). Such networks have knit dispersed
professionals, technicians, administrators or sales person-
nel   into    “highly     cohesive     and    highly      cooperative         . . .




                                                                                      7
GLOBAL         NETWORKS          AND LITTLE             BOXES                   Some pundits have feared the advent of CSSNS (e.g., Stoll
                                                                                1995). They see the desirable form of community as dense
The evidence        suggests that despite       their    limited   social
                                                                                bounded groups and work while fearing that technological
presence, CSSNS successfully         maintain     strong, supportive
                                                                                change, industrialization             and urbanization    have so isolated
ties of work and community, and they foster an increase
                                                                                people to create alienated              individuals,     standing alone in
in the number and diversity   of weak ties. They are
                                                                                mass societies (a longstanding tradition in social commen-
especially suited to maintaining intermediate-strength  ties
                                                                                tary, e.g., Nisbet 1962). The evidence shows that although
between people who cannot see each other frequently.
                                                                                computer networks are affecting social networks, they are
Online relationships are based more on shared interests
                                                                                by no means destroying social networks. Rather, such
and less on shared social characteristics. Although many
                                                                                transformations are the intensification  and ramification of
relationships function offline as well as online, CSSNS are
                                                                                continuing social trends in the nature of work and commu-
developing norms and structures of their own. They are
                                                                                nity.
not just pale imitations of “real life.”
                                                                                Bellcore’s     Vice-President           of Research reports that when
These relationships      form social networks that need not be
                                                                                “scientists    talk       about the evolution          of the information
social    groups.     How    such   computer       supported       social
                                                                                infrastructure,...        [we don’t] talk about... the technology. We
networks vary in their size, heterogeneity,    density and
                                                                                talk about ethics, law, policy and sociology .... It is a social
boundedness both reflects the social systems in which they
                                                                                invention”   (Lucky 1995: 205). To date, there has been
are embedded and the interactions            of people within       these
                                                                                little sociological  study of computer supported social
social networks.
                                                                                networks. Yet as my computer science colleague William
The evidence suggests that organizational               boundaries    are       Buxton has told me, “the computer science is easy; the
becoming more permeable just as community boundaries                            sociology      is hard.”
already have. The combination  of high involvement in
CSSNS, powerful   search engines and the linking   of
organizational       networks    to the internet         enables   many                                 ACKNOWLEDGEMENTS
workers       to   connect   with    relevant     others     elsewhere,         Work     for this paper has been supported                  by the Social
wherever they are and whomever they work for. As                                Sciences and Humanities                 Research Council         of Canada,
organizations grow towards their information and commu-                         Bell Canada, and the Information                   Technology      Research
nication sources. CSSNS should affect changes in organi-                        Centre. I am grateful for the advice provided by my
zational structures.                                                            colleagues in the Computer Supported Social Network
What are the implications of such changes for the societies                     project: Janet Salaff, Dimitrina Dimitrova, Laura Garton,
within which they are proliferating?  Social networks are                       Milena Gulia, Deborah Hardwick,        Caroline Haythorn-
simultaneously becoming more global and more local, as                          thwaite and Emmanuel Koku. More detailed versions of
connectivity  and domestic  matters intersect. Global                           the discussion here appear in Garton & Wellman 1995,
connectivity de-emphasizes the importance                 of locality for       Wellman,       et al. 1996, Wellman              1996, and Wellman          &
work and community; online relationships                  may be more           Gulia 1996.
stimulating        than suburban    neighborhoods         and alienated
offices. Even more than before, on the information
                                                                                                                  REFERENCES
highway each person is at the center of a unique personal
and work-group.    Moreover,   the domestic environment                         Berkowitz,  S.D. (1982). An Introduction to Structural
around the workstation is becoming a vital home base for                        Analysis: The Network    Approach to Social Research.
the neo-Silas Marners sitting in front of their screens day                     Toronto: Butterworth.
and night. Their well-feathered            nests have gone online.              Bikson,       T., & Eveland,           J. D. (1990).      The Interplay     of
The privatization       of relationships    (Wellman       1994), online        Work        Group        Structures     and Computer        Support.    In J.
as well as offline, has affected community, organizational                       Galegher,         R.    Kraut,    &    C. Egido       (Eds.),   hztellectual
and coworker solidarity. Virtual communities are acceler-                        Teamwork:   Social and Technological      Foundations of
ating the ways in which people operate at the centers of                         Cooperative Work (pp. 245-290). Hillsdale, NJ: Lawrence
partial, personal communities,       switching    rapidly and                    Erlbaum Associates.
frequently between groups of ties. Whether working at                            Brent      D.A.        (1994).    Information     Technology       and the
home or at an office workstation, many workers have an                           Breakdown          of “Places”         of Knowledge.       E.loumal     4(4):
enhanced ability to move between relationships. At the                           online
same time, their m-ore individualistic    behavior means the
                                                                                 Breiger,     R. (1974). The Duality             of Persons and Groups.
weakening of the solidarity that comes from working in
                                                                                 Social Forces, 53, 181-90.
large groups.




                                                                            8
Bristor, J. (1987). “Buying Networks: A Model of Posi                                      Haythornthwaite, C., Wellman, B., & Mantei, M. (1995).
tional Influence in Organizational Buying.” Ph.D. thesis,                                  Work Relationships and Media Use: A Social Network
School of Business, University               of Michigan.                                  Analysis.    Group Decision and Negotiation,              4(3), 193-211.

Buxton, B. (1992). Telepresence: Integrating                     Shared Task               Hesse B., Sproull         L., Kiesler S., Walsh J. 1993. Returns
and Personal       Spaces. Paper presented                  at the Graphics                to Science Computer Networks in Oceanography.                    Commu-
Interface ’92 Conference,             Vancouver.                                           nications    of the ACM 36(8): 90-101

Carley, K. (1990). Structural             Constraints       on Communica-                  Hiltz, S. R., Johnson, K., & Agle, G. (1978). Replicating
tion: The Diffusion            of the Homophoric     Signal Analysis                       Bales Problem Solving Experiments on a Computerized
Technique Through              Scientific Fields. Journal of Mathe-                        Conference:       A Pilot     Study      (Research      Report   No.     8).
matical Sociology,        15(3-4), 207-246.                                                Computerized Conferencing and Communications                     Center,
                                                                                           New Jersey l[nstitute of Technology.
Carley,      K., & Wendt,            K. (1991).     Electronic         Mail     and
Scientific    Communication.            Knowledge,         12(4), 406-440.                 Hiltz, S. R., Johnson, K., & Turoff, M. (1986). Experi-
                                                                                           ments in Group Decision Making: Communication Process
Constant, D., Kiesler S., Sprotdl L. (1994). What’s Mine
                                                                                           and      Outcome     in     Face-to-Face       Versus     Computerized
is Ours, Or is It? A Study of Attitudes about Information
                                                                                           Conferences.       Human       Communication         Research,       J3(2),
Sharing. Information           Systems Research 5(4):400-21
                                                                                           225-252.
Constant, D., Sproull, L., & Kiesler, S. (forthcoming). The
                                                                                           Hiltz,    S. R., & Turoff,      M. (1993).      The Network      Nation.
Kindness of Strangers: The Usefulness of Electronic Weak
                                                                                           (second cd,). Cambridge,           MA:    MIT    Press.
Ties for Technical        Advice.       Organization        Science.
                                                                                           Johnson-Lenz, P., & Johnson-Lenz, T. (1990). Islands of
Craven, P., & Wellman, B. (1973).                     The Network             City.
                                                                                           Safety for Unlocking Human Potential (Research Report
Sociological Inquiry 43:57-88.
                                                                                           No. 3). Awakening Technology.
Dantowitz,      A., & Wellman,           B. (1996). The Small World
                                                                                           Johnson-Len;t,      P., & Johnson-Lenz,         T. (1993). Community
of the Internet, Presented to the Canadian                     Sociology        and
                                                                                           Brain-Mind:        Groupware        Tools for     Healthy    Civic     Lije.
Anthropology      Association.          St. Catherine,         Ont.
                                                                                           Paper presenled at the Healthy Communities                  Networking
Dorsey, D. (1994). The Force. New York:                         Ballantine.                Summit,      San Francisco, CA.
Eveland, J. D., & Bikson, T. (1988). Work Group Struc-                                     Kaufer,     D., & Carley,         K. (1993).     Communication         at a
tures and Computer Support. ACM Transactions                           on Oj$ce            Distance:     The Injluence       of Print on Sociocultural      Organi-
Information      Systems, 6, 354-79.                                                       zation and Change. Hillsdale,            NJ: Lawrence       Erlbaum.
Finholt, T., & Sproull, L. S. (1990). Electronic                       Groups at           Kiesler, S., Sproull, L. (1988). Technological   and Social
Work. organization    Science, 1(1), 41-64.                                                Change in Organizational Communication       Environments.
Finholt, T., Sproull, L., & Kiesler, S. (1990). Communica-                                 Department of Social and Decision Sciences, Carnegie
tion and Performance in Ad Hoc Task Groups. In J.                                          Mellon      University.
Galegher, R. Kraut, & C. Egido (Eds.), intellectual                                        Kiesler,     S., Sproull,     L., Eccles,      J. (1985).    Pool Halls,
Teamwork:        Social        and    Technological         Founalztions          of       Chips, and War Games: Women                    in the Culture of Com-
Cooperative      Work (pp. 291-325).              Hillsdale,    NJ: Erlbaum.               puting. Psychology          of Women Quarterly          9:451-462
Foderaro,      L. (1995).       Seekers of Self-help            Finding        it on       King,     S. (1994). Analysis        of Electronic      Support Groups
Line. New York Times, March 23.                                                            for Recovering        Addicts.      Interpersonal       Computing       and
Garton, L. (1995). An Empirical     Analysis of Desktop                                    Technology,       ;!(3), 47-56.
Videoconferencing   and Other Media       in a Spatially                                   Kling, R. (1995). Social Relationships   in Electronic
-Distributed Work Group. Centre for Information Technol-                                   Forums: Hangouts, Salons, Workplaces and Communities.
ogy Innovation.                                                                            (30 pp). University of California, Irvine: Department of
Garton,      L., & Wellman,            B. (1995).       Social        Impacts     of       Information  and Computer Science and Public Policy
Electronic      Mail      in    Organizations:         A     Review       of     the       Research Organization.
Research Literature. Communication                  Yearbook, 18,434-53.                   Kling,     R. (19!)6). Computerization          and Controversy,        2nd
Granovetter,      M.      (1973).      The   Strength          of Weak         Ties.       ed. San Diegoc Academic             Press.
American       Journal    of Sociology,       78, 1360-80.                                 Kollock, P., & Smith, M. A. (1996). Managing the Virtual
Granovetter,      M. (1982).The Strength of Weak Ties: A                                   Commons: Cooperation and Conflict in Computer Com-
Network    Theory  Revisited.   In P. Marsden & N. Lin                                     munities. In S. Herring (Ed.), Computer-Mediated Commu-
(Eds.), Social Structure      and Network   Analysis (pp.                                  nication     (Forthcoming).       Amsterdam:      John Benjamins.
105- 130). Beverly Hills, CA: Sage.




                                                                                       9
Kraut,      R,, Attewell,               P. 1993.     Electronic       Mail     and Organi                (Eds.),    Telematics and Work. (pp. 109-136). East Sussex,
zational     Knowledge.                 Working      paper,    Department         of Social              England:     Lea.
and Decision             Sciences         Carnegie        Mellon      University.                        Scott, J. (1991). Social          Network Analysis.         London:       Sage.
Latan6,      B., & Darley, J. (1976). Help in a Crisis:                                 By-              Siegel, J., Dubrovsky, V., Kiesler, S., & McGuire, T. W.
stander       Response to an Emergency. Morristown,                                     NJ:              (1986). Group Processes in Computer-Mediated   Communi-
General Learning                   Press.                                                                cation.    Organizational         Behaviour       and Human           Decision
Laumann,           E., Gagnon,              J., Michael,            R. & Michaels,         S.            Processes, 37, 157-87.
(1994). The Social Organization   of Sexuality: Sexual                                                   Sproull,     L., & Faraj, S. (1995). Atheism,                   Sex and Data-
Practices in the United States. Chicago: University of                                                   bases: The Net as Social Technology.                     In B. Kahin & J.
Chicago Press.                                                                                           Keller     (Eds.), Public       Access to the Internet            (pp. 62-81).
Laumann,           E., Marsden, P. & Prensky, D. (1983). R. Burt                                         Cambridge,      MA:    MIT       Press.
& M. Minor                   (Eds.), in Applied             Network          Analysis   (pp.             Sproull, L., & Kiesler, S. (1991). Connections. Cambridge,
18-34).      Beverly           Hills,     CA:     Sage.                                                  MA: MIT Press.
Lea, M., & Spears, R. (1992).                             Paralanguage and Social                        Steinfield, C. (1986). Computer-Mediated                    Communication
Perception in Computer-Mediated                            Communication. Journal                        in an Organizational Setting: Explaining                   Task-related and
of Organizational                  Computing,         2(3-4), 321-341.                                   Socio-emotional           Uses.     Communication               Yearbook,         9,
Lea, M., & Spears, R. (1995). Love at First Byte? Build-                                                 777-804.
ing Personal Relationships Over Computer Networks. In                                                    Stoll, C. (1995). Silicon           Snake Oil:         Second     Thoughts        on
J. T. Wood &S. Duck (Eds.), Understudied                                 Relationships:                  the Information       Highway.        New      York:    Doubleday.
Off the Beaten Track (pp. 197-233). Thousand Oaks, CA:
                                                                                                         Walsh, J., & Bayama, T. (1996). Computer Networks                            and
Sage.
                                                                                                         Scientific    Work. Social Studies of Science, forthcoming.
Lewis, P. (1994). Strangers, Not Their Computers, Build
                                                                                                         Walther,      J. B.   (1994).      Anticipated         Ongoing      Interaction
A Network in Time of Grief. New York Times, March 8:
                                                                                                         Versus     Channel    Effects      on Relational         Communication            in
Al, D2
                                                                                                         Computer-Mediated    Interaction.                Human      Communication
Lucky,       R. (1995). What Technology                             Alone      Cannot Do.                Research, 20(4), 473-501.
Scientific American                     273(3):204.
                                                                                                         Walther,      J. B. (1995).        Relational      Aspects of Computer
Mantei, M., Baecker, R., Sellen, A., Buxton, W., Milligan,                                               -Mediated       Communication:             Experimental          Observations
T., & Wellman,                  B. (1991).         Experiences         in the Use of a                   Over Time. Organization              Science, in press.
Media Space. Paper presented at the CHI ‘91 Conference.
                                                                                                         Wasserman,          S. &     Faust,       K.   (1993).     Social      Network
Mantei,       M., & Wellman,                     B. (1995). From Groupware                     to        Analysis:      Methods      and Applications.            Cambridge:        Cam-
Netware: Implications for CSCW Design. (12 pp). Univer-                                                  bridge University         Press.
sity of Toronto: Department of Computer Science and
                                                                                                         Wellman, B. (1988). Structural Analysis: From Method
Department of Sociology.
                                                                                                         and Metaphor to Theory and Substance. In B. Wellman &
Meyer,       G. (1989).             The     Social     Organization           of the Com-                S. Berkowitz (Eds.), Social Structures: A Network Ap-
puter       Underground.                  Master’s        thesis.     Northern      Illinois             proach, (pp. 19-6 1). Cambridge: Cambridge University
University.                                                                                              Press.
Mitchell,          W.        (1995).      City    of Bits:     Space, Time and the                       Wellman,       B. (1992). Which Types of Ties and Networks
Infobahn.          Cambridge,             MA:      MIT      Press.                                       Give What Kinds of Social Support? Advances in Group
Mowshowitz,               A. ( 1994). Virtual             Organization: A Vision of                      Processes, 9, 207-235.
Management               in the Information               Age. Information Society                        Wellman, B. (1994). I was a Teenage Network Analyst:
 10:267-88.                                                                                               The Route from The Bronx to the Information Highway.
Nisbet,       R.    (1962).             Community          and Power.          New York:                  Connections, 17(2), 28-4.5.
 Oxford       University            Press.                                                                Wellman, B. (1996). An Electronic Group is Virtually a
 Rheingold,             H.     (1993).       The Virtual Community: Home-                                 Social Network. In S, Kiesler (Ed.), Research Milestones
 steading          on        the       Electronic   Frontier. Reading, MA:                                on the Information Highway. (forthcoming). Hillsdale, NJ:
 Addison- Wesley.                                                                                         Lawrence      Erlbaum,

 Rice, R., & Steinfeld,                     C. (1994). Experiences               with New                 Wellman,      B., & Gulia,       M, (1996).      Net Surfers Don’t Ride
 Forms of Organizational  Communication     via Electronic                                                Alone. In P. Kollock and M. Smith (Eds.), Communities
 Mail and Voice Messaging. In J. E. Andriessen & R. Roe                                                   in Cyberspace (forthcoming).  Berkeley: University  of
                                                                                                          California Press.



                                                                                                    10
Wellman, B., & Potter, S. (1996). The Elements of                 Social Networks. Annual Review of Sociology 24: (forth
Personal Communities. In B. Wellman (Ed.), Networks in            coming).
lhe  Global  Village.   (forthcoming.) Boulder, CO:               Wellman, B., Tindall, D. (1993). Reach Out and Touch
Westview.                                                         Some Bodies: How Social Networks Connect Telephone
Wellman, B., Salaff, J., Dimitrova, D., Garton, L. &              Networks. Progress in Communication Science 12:63-93
Haythornthwaite, C. (1994). The Virtual Reality of Virtual        WeUman, B., & Wortley, S. (1990). Different Strokes
Organizations, Presented to the Annual Meeting of the             From Different Folks: Community Ties and Social Sup-
American Sociological Association. Los Angeles.                   port. American Journal of Sociology, 96, 558-88.
Wellman, B., Salaff, J., Dimitrova, D., Garton, L. Gulia,         Wireman, P. (1984). Urban Neighborhoods, Networks, and
M. & Haythornthwaite, C. (1996). Computer Networks as             Families.Lexington, MA: Lexington Books,




                                                             11