Expert Workshop on the Measurement of Social Capital for by fuk43069

VIEWS: 0 PAGES: 34

									Expert Workshop
on the Measurement of Social
Capital for Public Policy
Synthesis Report
June 8, 2004




PRI Project
Social Capital as a Public Policy Tool


In collaboration with
Statistics Canada
Expert Workshop
on the Measurement of Social
Capital for Public Policy
Synthesis Report
June 8, 2004




PRI Project
Social Capital as a Public Policy Tool
    Discussion Paper




    This document is a synthesis report from the      was placed on the measurability of social net-
    Expert workshop on measuring social               works that provide access to resources and
    capital for public policy purposes, held          support, these being understood as the core
    on June 8th, 2004, at the Sheraton Hotel in       constituent elements of social capital. The
    Ottawa. The workshop was organized by the         exercise was to identify essential elements of
    Policy Research Initiative (PRI) as part of its   a “toolbox” for analyzing social capital for pur-
    interdepartmental project “Social Capital as      poses of developing and assessing government
    a Public Policy Tool”, in partnership with the    programs and policies.
    Social Sciences and Humanities Research
                                                      The present document provides an outline of
    Council (SSHRC) and in collaboration with
                                                      the presentations and discussions held by the
    Statistics Canada.
                                                      experts. The structure follows the program,
    The workshop brought together about seventy       except for Professor Derek Hum’s commen-
    academic and government experts to examine        tary, for which the complete version is repro-
    different empirical strategies and measurement    duced at the end of the report.
    tools used in social capital research. Emphasis




2
Genomics, Health and Society Emerging Issues for Public Policy




TABLE                         OF             CONTENTS
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5

Session 1:
Main approaches to analyzing social networks and their usefulness
for social capital-related public policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7

Session 2:
Measuring the constituent elements of social capital: what exactly
are we measuring, and how do we go about it? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11

Session 3:
Measuring social capital at the community level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15

Session 4:
Measuring the cause-effect link between social capital and socio-economic outcomes                                                                                                   . . . . . . . . .19

Conclusion:
Advantages and limitations of the main tools of measuring social capital
for the development and evaluation of federal policies and programs . . . . . . . . . . . . . . . . . . . . . . . . . . . .23

Measuring Social Capital for Public Policy: Comments and Reflections . . . . . . . . . . . . . . . . . .25

Appendix 1:
Workshop Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33

Appendix 2:
Workshop Information                          . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35




                                                                                                                                                                                                              3
Genomics, Health and Society Emerging Issues for Public Policy




INTRODUCTION
A network approach to facilitating                     network approach to social capital. The under-
and measuring social capital                           lying hypothesis is that social capital refers to
Jean-Pierre Voyer, Executive Director, Policy          the networks of social relations that provide
Research Initiative (PRI)                              access to resources and support. Understand-
                                                       ing how social relations and their dynamics
One year ago, the PRI launched its “Social Cap-        constitute an additional resource for individu-
ital as a Public Policy Tool” project in order to      als and communities and how these resources
assess the usefulness and application of social        can be created, accessed and productively uti-
capital for public policy development. Social          lized is an avenue which can potentially offer
capital is a subject that raises                                        substantial benefits from the
many questions for policy                                               public policy standpoint.
makers. For example: To what              To be useful, the
extent do existing inventories                                          In November 2003, the PRI in
                                          concept of social
of social capital serve to facili-                                      concert with the OECD and
                                            capital must be             other federal partners organ-
tate or impede key policy
objectives? What impact do
                                         operationalized in             ized an international confer-
government programs have                 ways that allow the            ence designed to bring about
on the creation, utilization and           identification of            a better understanding of the
performance of social capital?            policy levers that            role of social capital in the
How can the use of a social             enable its utilization.         integration of immigrants and
capital perspective improve                                             the management of diversity.
policy outcomes or make serv-                                           A pre-conference workshop
ice delivery more effective?                          was also held to inform participants of the cur-
                                                      rent status of data development in the fields
Efforts to harness the concept of social capital      of social capital and immigration. Two main
for public policy purposes have come up               messages emerged from this activity. First,
against certain conceptual and measurement            strategic issues should determine the way that
difficulties. The debate on the policy useful-        we operationalize social capital. The aspects
ness of this concept remains ongoing, and             of social capital that we want to measure and
scepticism continues to be voiced on this sub-        study cannot be established in the abstract:
ject. To be a useful tool, the concept of social      they have to be related to the strategic issues
capital, like human capital, has to be opera-         in which we are interested. Second, there is no
tionalized so that policy makers can precisely        lack of data on social capital, but rather a lack
identify what it is, explore its productive poten-    of coherence and integration in the production
tial for achieving broader policy objectives, and     and analysis of that data.
identify policy levers for affecting the many
ways it is created, accumulated and utilized.         In terms of official statistics, we now have
                                                      available a wide range of data directly or indi-
To overcome the conceptual and measurement            rectly related to social capital, but most have
ambiguities associated with the concept, the          not been developed on the basis of the same
PRI has introduced an operational definition          analytical framework. Consequently, existing
and an analytical framework for the study and         statistics can provide a basis for doing good
measurement of social capital for federal pol-        descriptive and comparative analysis, and they
icy purposes. This framework is based on a            also allow more or less solid correlations
                                                                                                           5
    Synthesis Report




    between some of the dimensions of social capi-           the foundations of a rigorous empirical
    tal. But the absence of an integrated set of             approach to research on social capital. From
    strategic data on key elements of social capital         an empirical standpoint, using a network
    remains an important limit to existing surveys.          approach to study social capital takes advan-
    Thus, it is still difficult to explore relations of      tage of a rich, tested body of research that is
    causality that would demonstrate how social              useful for public policy purposes. Network-
    capital is related to socio-economic or health           based analysis is a field of research that is
    outcomes. Since empirical investigation of               supported by solid theory and utilizes refined
    social capital is still in its early stages, it has      measurement and analysis tools.
    also been suggested that we
                                                                               The expert workshop had the
    take advantage of more
                                                A network approach             following objectives:
    exploratory research such
    as qualitative analysis, case                 to studying social           • to offer an overview of the
    studies and social experimen-                                                 different approaches used to
                                                 capital allows us to
    tation, or try out new tools                                                  measure social capital from
                                             take advantage of a rich,
    such as those used in network                                                 a social net-works perspec-
                                              tested body of research             tive, based on a relevant
    analysis. This exploratory,
                                              that is useful for public           selection of Canadian
    inductive work will serve to
    expand our knowledge so that                   policy purposes.               research projects;
    more solid, major surveys can                                              • to discuss the analytical
    be built that can eventually give us a better               potential and public policy relevance of
    understanding of the relationship between                   different strategies of empirical investigation
    social networks, social capital and individual              of social capital as applied to specific issues,
    and collective outcomes.                                    both individual and collective; and

    Based on these observations, the PRI organ-             • to explore in greater detail a number of
    ized an expert workshop on measuring social               tools for measuring social capital and evalu-
    capital for public policy. Some of the top                ate the possibilities and limitations of apply-
    experts in the measurement of social capital              ing them in a public policy context.
    and social networks contributed their knowl-
    edge and experience toward the goal of laying




6
Expert Workshop On The Measurement Of Social Capital For Public Policy




SESSION 1:
Main approaches to analyzing social networks and their usefulness
for social capital-related public policy
Two main approaches to network analysis are          Understanding differences in the structures of
generally differentiated: the morphological          various network ties is an increasingly power-
approach, whereby social capital is a product        ful tool for explaining variations in resources,
of the structure of linkages, more specifically      social behavior and socio-economic outcomes.
of the configuration of social networks, and the     Network structures may vary in a number of
transactional approach, whereby social capital is    ways. For example, they may be quite dense,
instead associated with interactions among           or they may be sparsely knit. They may be
individuals, that is the relational dynamic.         tightly bounded, or be permeated with links to
                                                     many other networks.
The first block of presenta-
tions was designed to answer                                          Network analysis may be
the following questions:               Network analysis is            interested in understanding
• What are the fundamental            the study of existing           the structure or pattern of a
  characteristics of each of          relationships and the           whole network, or it may be
  the approaches?                                                     interested in studying the
                                       resources to which
                                                                      personal social networks
• How are they useful for                they give access.
                                                                      available to specific nodes
  studying social capital?
                                                                      (individuals, organizations,
• What are their respective implications             firms, etc.), sometimes referred to as an ego-
  for selection of the constituent elements          centered approach.
  of social capital to be measured?
• How are they different in terms of tech-           Studying a whole social network involves iden-
  niques of empirical investigation?                 tifying and analyzing a comprehensive set of
                                                     relationships in an entire social network (vil-
• Is it possible to benefit from both
                                                     lage, organization, kinship, etc.). This can be
  approaches by combining the study of
                                                     quite a challenge as it requires a roster of the
  network structures with the relational
                                                     entire population of the network, it requires
  dynamic?
                                                     that there be a social boundary or limit to the
                                                     network, and analysis may be severely
Measuring social capital through
                                                     impaired if any data is missing.
analysis of social networks
Barry Wellman, Professor of Sociology,               Studying a personal social network from an
University of Toronto                                ego-centred view may be a better means of
                                                     studying unbounded networks, and may be
Barry Wellman began the day by providing an
                                                     done using standard survey research. However,
overview of social network analysis. Wellman
                                                     this approach often tends to concentrate on
defined a social network as consisting of
                                                     strong ties at the expense of important weak
one or more nodes, connected by one or
                                                     ties, often relies on potentially inaccurate
more ties, that form distinct, analyzable pat-
                                                     reports from respondents, and can be
terns. Social network analysis is interested
                                                     hard to aggregate to a collective level.
in identifying existing relationships and
understanding the resources that are                 According to Wellman, we may increasingly
available from those relationships.                  see society as a network of networks. A
                                                                                                        7
    Synthesis Report




    networked society is characterized by a multi-         resources in networks, particularly in terms of
    plicity of specialized relations, fluid connec-        rules and standards of network operation.
    tions, and new relationships that may be less
                                                           With the support of a number of research proj-
    palpable than those of traditional societies (for
                                                           ects partaking of different theoretical traditions
    example, electronic network relations). Well-
                                                           and trends, the researcher presented different
    man noted that, particularly with new commu-
                                                           ways of approaching these issues: work that
    nications technologies, the volume and velocity
                                                           has been done on personal talents, on social
    of social relationships and interactions that
                                                                              support, on mobilization of
    individuals maintain have
                                                                              resources in the context of
    increased (think
                                             Networks are formed              key events in individuals’ lives,
    of your email networks, for
                                                                              and finally on actual evolution
    example). At the same time               and transformed over
                                                                              of the make-up of social net-
    traditional household relations           time, depending on
                                                                              works throughout the life-
    are more and more stressed                 circumstances and              course of individuals by
    and group ties are increasingly             specific contexts.            means of panel studies. These
    fluid. Finding new tools to
                                                                              projects demonstrate that net-
    navigate a networked society
                                                           works are formed and transformed over time,
    will be important
                                                           depending on circumstances and specific
    for policymakers.
                                                           contexts, and that this automatically affects
                                                           resource mobilization patterns. They also
    Measuring social capital through
                                                           demonstrate that social networks and social
    analysis of relational dynamics
                                                           capital do not always operate according to a
    Johanne Charbonneau, Research Professor,
                                                           logic of accumulation. While it is possible to
    Institut national de la recherche scientifique -
                                                           build a stock of social capital, it is also possible
    Urbanisation, culture et société (Montréal)
                                                           to see it fall apart, especially if it is not well
    Johanne Charbonneau’s presentation con-                taken care of: “You also have to make deposits
    cerned the importance of understanding the             to your social resources bank account, not just
    rules that permit networks to exist and to pro-        make withdrawals” said Charbonneau. The
    duce social capital. While recognizing the inter-      researcher sees more thorough reflection on
    est of studying network morphology for                 the rules and standards that govern social
    research on social capital, she noted that an          exchanges (e.g., reciprocity and trust) as essen-
    approach that exclusively emphasizes the char-         tial to a better understanding of why some net-
    acteristics and structure of social relations          works can be “mobilized” or “activated” in
    could limit understanding of relational dynam-         terms of social capital, and why some cannot.
    ics. She insisted that we must go beyond corre-        There are certain ingredients such as trust that
    lations, to investigate the cause-effect links         serve to “lubricate” social relations, the condi-
    between social ties and resources. She raised          tions for which are still not well known.
    the question of whether the possession of per-
                                                           Finally, the researcher pointed out the interest
    sonal resources is necessary in order to have
                                                           of adopting a life-event perspective for the
    access to a diverse, effective network, or con-
                                                           study of social networks especially to under-
    versely, whether one must start with a well-
                                                           stand the strategies adopted by individuals for
    established network in order to be able to
                                                           co-ordinating resources that they derive from
    access resources. Another important question
                                                           different sources. For example, study of the ice
    according to Charbonneau is the actual
                                                           storm emergency in Quebec made it possible
    mechanics of producing and circulating
8
Expert Workshop On The Measurement Of Social Capital For Public Policy




to understand how resource mobilization              on the methods that permit social capital to be
strategies in social networks also depend on         activated or mobilized in specific cases.
the supply of resources outside the networks,
                                                     From the public policy standpoint, Lévesque
particularly via public services.
                                                     explains that the two types of studies are com-
                                                     plementary. Programs that have the support
Social structure and relational
                                                     of pre-existing social capital must ensure that
dynamics: two complementary
                                                     account is taken of both the structural and the
approaches to the study of social
                                                                       relational sources of that capi-
capital
                                                                       tal. On this subject, Lévesque
Maurice Lévesque, Professor,
                                         Programs that make            cites the example of job train-
Department of Sociology,
                                         use of social capital         ing programs for social insur-
University of Ottawa
                                                                       ance recipients that focus on
                                     must take account of both
Maurice Lévesque made a                                                integrating people within a
detour through the paradigms
                                      structural and relational        vocational structure. This is
in order to discuss the public         sources of that capital.        certainly a quick and effective
policy interest of network                                             way of expanding contacts,
analysis.                                            but insofar as such persons are integrated as
                                                     “recipients”, relations will be developed on the
The structural analysis paradigm that is the ori-
                                                     basis of that status, and over the longer term
gin of network analysis implies that: (1) the
                                                     this may not allow for the type of integration
players’ behaviours are the result of their social
                                                     desired. This is to say that while “recipients”
position (in the network), for it is that position
                                                     may develop new structural ties to people
which determines the opportunities and con-
                                                     with resources they wish to access, the nature
straints of access to resources; and (2) the
                                                     of the ties and their interactions may not in
social structure (the network) is the result of
                                                     fact translate into an ability to access these
the players’ interaction. The studies that focus
                                                     resources. For in fact, developing ties is not
on structure will generally emphasize the
                                                     everything: those ties also have to be built on
potential of opportunities and constraints and
                                                     solid foundations. In the case of welfare recipi-
on differentiated access to resources offered
                                                     ents, the instability associated with that status
by different network structures (e.g., for job
                                                     might be transposed directly to the tie in ques-
searching, accessing information, etc.). This
                                                     tion. An approach that is more concerned with
type of research can give a good overview of
                                                     creating social capital would pay particular
the available social capital that can potentially
                                                     attention to exploiting the relational capacities
be mobilized by the players.
                                                     of the recipients so as to help them form sus-
In contrast, research emphasizing a transac-         tainable and useful ties. This example shows
tional approach will instead stress the dynam-       how interventions concerned with structure
ics at play within networks (causality,              and those concerned with the type of relations
formation and transformation of networks,            can complement each other in generating
mechanisms that activate interchange, etc.).         social capital.
This type of research can provide information




                                                                                                          9
     Discussion Paper




     SESSION 2:
     Measuring the constituent elements of social capital:
     what exactly are we measuring, and how do we go about it?
     Social capital refers to the networks of social         • What is the public policy relevance of these
     relations that provide access to resources                 different elements of assessment?
     (information, co-operation, etc.) and support           • What are the most appropriate measure-
     (help, assistance, etc.). A relatively simple defi-        ment tools for this?
     nition, analytically, but one
                                                                               • What are the advantages and
     that raises a whole series of
                                                                                 limitations of the different
     empirical considerations. For             The name generator                measurement tools?
     example, certain researchers              makes it possible to
     will be interested in the poten-
                                             identify strong ties and          Measuring the social
     tial resources resulting from
                                               their specific nature;          capital produced by
     certain types of linkages that
                                               it is a useful tool for         strong ties using a
     the members of a social net-
                                            investigating issues such
                                                                               name generator
     work can deploy for an indi-
                                                                               Peter Marsden, Professor
     vidual. Others will ascribe                 as social support.
                                                                               of Sociology, Harvard
     more importance to the
                                                                               University
     resources actually mobilized by a person
     through his or her social network. Different            Measurement of social networks by means of
     analytical concerns will therefore result in            large social surveys can take various forms.
     measurement                                             Peter Marsden distinguished the general meas-
     of different values of social capital. We can           ures to be found in many major surveys from
     also explore the type of resources useful for           the more targeted measures modelled on the
     specific issues (job entry, personal problems,          “small world” approach. The first technique is
     etc.), measuring the way that these resources           to insert in general surveys a set of questions
     are distributed in the different population             related to the social networks of the respon-
     sub-groups (by age, gender, ethnic or socio-            dents, such as the number of friends they have,
     economic affiliation). Finally, we can also try         frequency of contacts, etc. There is an impres-
     to identify the type of resources linked to the         sive number of questions of this type. They
     type of relations/ties. For example, analytical         have the advantage of being easy to incorpo-
     interest is different according to whether one          rate in any social survey, but they are not
     is interested in the strong ties/weak ties typol-       always easy to answer, and not all of them are
     ogy or the bonding/bridging/linking typology.           reliable. Furthermore, they only provide a rela-
                                                             tively approximate idea of the composition of
     The second block of presentations was
                                                             a person’s social network.
     designed to answer the following questions:
     • What are the different elements to be meas-           The “small world” technique is a more precise
         ured in assessing the value of a person’s           approach to networks, where more detailed
         social capital (access to, inventories and          interest is taken in the persons with whom
         usefulness of resources; type, quality and          the respondent has close relations and in the
         quantity of relations, etc.)?                       nature of those particular relations. Marsden

     1   Ego is the person at the centre of the web of relations and with whom the analysis is concerned.
     2   The alters are the persons with whom ego relates, i.e. the members of his/her network.
10
Expert Workshop On The Measurement Of Social Capital For Public Policy




explained how the General Social Survey on            discuss the importance of looking at weak
Social Networks in the United States has              ties (or peripheral relations) for social capital
applied this investigation technique with the         research. As opposed to strong ties, weak ties
assistance of two tools known as the name             are those that are generally formed with per-
generator and name interpreter. The first tool        sons who are not members of the intimate cir-
is used to identify who is a member of an ego’s    1  cle of family and close friends. Weak ties are
network; the second uses the information thus         our connection to the outside world. They are
collected to explore the details of the relations:    important not only because of their number
characteristics of alters,2 type of relations with    (up to 400 weak ties per ego, versus a dozen
them, quality of relations, ties among alters,        strong ties), but also because of their hetero-
etc. A sample standard ques-                                            geneity. Erickson says it is a
tion from a name generator              The position generator          fact that individuals tend to
would be: “Over the last six                                            associate with persons with
                                      enables the investigation
months, with whom have you                                              whom they have the most
                                        of the weak ties; these
discussed subjects that are                                             affinities. As a result, it is the
important to you?” There                 represent our link to          persons more remote from us,
are a number of versions                 the outside world, to          with whom we have weaker
of these tools that are gener-            individuals that are          ties, who are most likely to
ally adapted to the subject              socially more distant          have different types of
being studied, as well as                 and who most likely           resources that we do not.
abundant documentation
                                       have the resources that          According to Erickson, to
on their performance.
                                            we do not have.             investigate weak ties it is not
According to Marsden, this                                              necessary to do the meticu-
method of measuring social                                              lous work of inventorying
networks has important analytical potential           every facet of all of an individual’s social rela-
for research on strong ties, for example, for         tions. One must simply have a good idea of the
studying social support issues. But it is not an      diversity of a person’s relations to be able to
approach that emphasizes resources as such,           estimate the potential resources accessible
nor the more peripheral relations that often          with sufficient certainty. The position genera-
generate social capital. In addition, the tech-       tor and resource generator are two simple sur-
nique demands considerable survey time                vey tools for measuring heterogeneity.
(about 15 minutes to study the relations
                                                      The position generator utilizes a person’s occu-
between ego and three alters).
                                                      pation as an indicator of the resources avail-
                                                      able to that person. The tool thus consists of
Measuring the social capital
                                                      a series of simple questions on the existence
produced by weak ties using
                                                      of “weak” ties to persons of diverse occupa-
position and resource generators
                                                      tional horizons who in principle have access
Bonnie Erickson, Professor of Sociology,
                                                      to similarly diverse resources. The selection
University of Toronto
                                                      of professions (from 15 to 30) is established
While acknowledging the importance of study-          on the basis of a scale of prestige reflecting
ing the role of strong ties in exploring certain      potential accessibility of various resources.
research topics, Bonnie Erickson came to              The main question is “Do you know someone


3   Bonnie Erickson kindly presented the research of Martin Van der Gaag and Tom Snijders, supported by the presentation
    material that they had prepared for the workshop.
                                                                                                                           11
     Synthesis Report




     in profession X?” In theory, the persons with        resources are nonetheless “strong” enough
     the richest and most heterogeneous networks          that the respondent knows the names of
     also have access to rich and heterogeneous           those persons and is occasionally in touch
     social capital.                                      with them, at least on a friendly basis. In thus
                                                                            specifying the level of intensity
     According to Erickson, this
                                                                            of the tie, the resource genera-
     type of capital is particularly
                                              An inventory of the           tor is sure to identify only
     useful for resolving problems
                                             number and types of            “mobilizable” ties, that is, ties
     related to job searching or
                                                                            with alters who would proba-
     information sharing. Different          social network ties is
                                                                            bly agree to offer ego the
     versions of the tool can be              a way to access the
                                                                            resources available to them.
     tailored to the research inter-        network resources: the          The resource generator
     ests. For example, to do a            more extensive a person’s requires more survey time
     gender analysis one can add
                                          network, the more diverse than the position generator,
     a question on the gender of
     the alters. Or to study changes
                                         and accessible the potential but has the advantage of
                                              supply of resources.          allowing more intensive
     over time, the tool can be
                                                                            investigation of a wider
     incorporated in a panel survey.
                                                                            variety of issues.
     The position generator is also a simple
     and quick tool (2 to 3 minutes) that can
                                                          Measuring social capital by studying
     generate a lot of information useful for
                                                          network size, diversity and density
     social capital research.
                                                          Jeff Boase, Ph.D. candidate, Department of
     The other survey technique for investigating         Sociology, University of Toronto
     social capital produced by weak ties is imple-
                                                          The three tools presented above make it pos-
     mented using a resource generator. This
                                                          sible to estimate social capital by means of
     instrument, developed by a team of Dutch
                                                          strong ties (name generator), weaker ties
     researchers (Van der Gaag and Snijders), was
                                                          (position generator) and the resources that
     also presented by Erickson3 as a simple tool
                                                          can potentially be mobilized by those ties
     that is useful for social capital research. The
                                                          (resource generator). Jeff Boase presented
     technique is to question respondents directly
                                                          a simple survey tool that estimates social capi-
     about the resources to which they have access
                                                          tal by means of more general characteristics
     within their broad network. The main question
                                                          of the social network, namely size, diversity
     is in the following form: “Do you know some-
                                                          and density. According to this approach, the
     one who can potentially allow you access to
                                                          more extensive a person’s network, the more
     resources of type X?”
                                                          diverse and accessible the potential supply
     The tool is relatively flexible since the selec-     of resources. Redundancy of supply for a spe-
     tion of resources can cover a wide spectrum          cific resource implies that demand is better
     of issues: prestige and knowledge resources,         distributed among persons who can potentially
     information resources, skills and competencies       supply the resource.
     resources, social support resources. Further-
                                                          The technique presented by the researcher
     more, it is supported by a concept of probable
                                                          consists in a set of simple questions developed
     accessibility of resources, assuming that the
                                                          by McCarty and his collaborators, which allow
     weak ties to the persons who have those



12
Expert Workshop On The Measurement Of Social Capital For Public Policy




respondents to inventory their social network        asking questions about the proportion of per-
based on the number of ties maintained by            sons in the network who know one other,
type (very close, somewhat close). To facilitate     maintaining the distinction between strong
the inventory by the respondent, the question-       ties and weak(er) ties.
naire suggests some categories of ties (immedi-
ate family, other relatives, neighbours, work,       Strengths and limitations of instru-
social group, etc.). For very close ties, the main   ments of analysis of social networks
question takes this form: “Let’s start with the      for measuring social capital
people you feel VERY close to, which might           Group discussion
include those you discuss important matters
                                                     The interest of measuring social capital using
with, regularly keep in touch with, or are there
                                                     the survey techniques presented lies in their
for you when you need help… how many are:
                                                     relative simplicity. All the approaches estimate
a) members of your immediate family; b) other
                                                     the resources potentially mobilizable by mem-
relatives; c) neighbours…” etc. To make the
                                                     bers of a social network from a general per-
inventory easier, respondents can be asked to
                                                     spective, i.e., by associating the characteristics
write down the names of the persons they list.
                                                     of the networks with overall “inventories” of
They can refer to their list of names in subse-
                                                     social capital. However, most researchers
quent questions. The questions on diversity and
                                                     agree that such instruments are less useful for
density that complete the tool are also organ-
                                                     illuminating specific problem in specific situa-
ized according to type of ties. For the diversity
                                                     tions, that is particular issues that require more
of very close ties, for example, questions can
                                                     intensive exploration not only of the quantity
be asked about the gender or ethnic origin of
                                                     of social capital available but also of its nature
alters. The question will take this form: “Think-
                                                     and conditions of use. It is possible, however,
ing of the people you feel very close to, how
                                                     to adapt and/or combine the generators to or
many of them are: a) of the same ethnicity as
                                                     with other sets of questions that get at more
you; b) the same gender as you; c) etc.” As for
                                                     specific interactions.
density of the network, this is measured by




                                                                                                          13
     Synthesis Report




     SESSION 3:
     Measuring social capital at the community level
     There is considerable exploration of whether           Measuring the relation between
     it is possible to measure social capital at the        community resilience and access to
     community level. For instance, instead of              social capital both within and outside
     mobilizing the social capital derived from indi-       the community
     viduals’ social networks, one can explore the          Ralph Matthews, Professor of Sociology,
     possibilities/capacities of these individuals to       University of British Columbia and McMaster
     exploit resources produced within groups or            University.
     organizations for purposes of
     dealing with adversity. Con-                                            As part of the Resilient Com-
     versely, one can investigate             A community’s level            munities Project he is direct-
                                                                             ing on the west coast of
     the extent to which communi-          of “resilience” depends
     ties can rely on the social                                             British Columbia, Professor
                                            on the residents’ ability
     capital of their members to                                             Matthews has become inter-
                                         to mobilize social networks ested in the relationship
     revitalize themselves or ensure
                                          to access certain resources. between costal communities’
     their prosperity.
                                                                             social capital and their socio-
     The third block of presentations was designed          economic development. More specifically, his
     to answer the following questions:                     research is an attempt to determine whether
     • How is this “collective” share of social capi-       networks of social relations operating within
         tal to be measured?                                and between the communities can provide a
     • What are the most appropriate levels of              buffer to economic crisis and help them adjust
         analysis for evaluating collective social capi-    to change. The underlying hypothesis is that a
         tal (region, community, neighbourhood)?            community’s level of “resilience” depends on
                                                                              the residents’ ability to mobi-
     • Can communities be ana-
                                                                              lize social networks to access
         lyzed as “networks of net-
                                             The social capital of a          certain resources.
         works”? If so, are
                                               community may be
         measurement tools based                                              Methodologically, the task
         on network analysis rele-         associated with networks of the Matthews team was to
         vant at this scale?                   that exist inside the          measure access to social cap-
     • What are the limits of aggre-           community or with              ital by individual members of
         gation of individual social         networks that bridge             the communities observed.
         capital for estimating collec-       to members of other             The research is supported by
         tive social capital?                                                 a questionnaire survey utiliz-
                                                  communities.
                                                                              ing a position generator and
     • Does context have specific
                                                                              by in-depth interviews of a
         effects on production and
                                                            sample of households and stakeholders in the
         use of the social capital generated by social
                                                            communities studied. The qualitative portion
         networks? If so, how are those effects to be
                                                            of the research was designed to complement
         measured?




14
Expert Workshop On The Measurement Of Social Capital For Public Policy




the information obtained from the position           Measuring the level of social support
generator so as to document the processes            in rural communities
whereby networks were mobilized and                  Janice Keefe, Associate Professor and Canada
resources exchanged.                                 Research Chair in Aging and Caregiving
To measure access to social capital, Matthews        Policy, Mount Saint Vincent University
explained how he had adapted the position             Janice Keefe came to discuss a method of
generator to include certain employment cate-         measuring a specific form of social capital:
gories characteristic of the region studied,          the social support for seniors in rural commu-
related in particular to the strong Aboriginal        nities. In the context of a vast research project
presence in these communities. The instrument         on the aging of the population, the concerns
was also modified to take account of weak ties                          of the research team associ-
and strong ties, and to distin-                                         ated with Professor Keefe
guish whether these ties                  An appreciation of            included that of understanding
existed within the communi-                                             and identifying the character-
                                      the community as a unit
ties or served as a bridge to                                           istics of the communities that
                                        of analysis allows for
members of other communi-                                               could be described as “sup-
ties. According to Matthews,            study of its distinctive
                                                                        portive”, i.e., communities
employing the position genera- characteristics, particularly characterized by a strong
tor in this way makes it possi-            in terms of social           stock of social support.
ble to differentiate the               resources and support.
intensity of relations affording                                        Methodologically, the commu-
access to “rich” versus “poor”                                          nity as a unit of analysis was
social capital, and according to whether that         studied by aggregating the individual character-
capital came from inside or outside the com-          istics of its members. For example, to create a
munity. The analysis is the richer for studying       typology of communities based on the level of
these outcomes in conjunction with other sur-         social support for seniors, the researchers used
vey data related to social capital, such as           the number of unpaid hours that the residents
participation in associative activities or level of   of each community devoted to caring for a sen-
trust. By aggregating the individual results thus     ior during a reference year (this question
obtained, the researcher hopes to identify the        appeared in the 2001 census). The communi-
social capital structure in each of the commu-        ties studied were differentiated by three levels
nities, to which he will then associate the quali-    of social support (high, moderate and low),
tative data on processes of producing and             which were associated with certain geo-
utilizing resources in those communities. Once        graphic, socio-economic and demographic
this multi-phase analysis has been completed,         characteristics (the independent variables)
he plans to draw conclusions on the relation-         derived from proportions (% of women, % of
ship between the level of resilience of these         longstanding community residents, % of sen-
west coast communities and their access to            iors, etc.)
social capital.                                       According to Keefe, this method of using statis-
                                                     tics can give one a true appreciation of the
                                                     community as a unit of analysis and allow for




                                                                                                          15
     Synthesis Report




     analysis of its distinctive characteristics, partic-  The analytical distinction between potentially
     ularly in terms of what they say about social         available social capital and used or mobilized
     support. Like other social networks experts,          social capital is another interesting aspect
     however, she cautions against the limitations of      of the researcher’s approach. According to
     this sort of approach, which does not take into       Reimer, the collective share of the supply of
     account the specific dynamics of creation and         social capital can be measured by studying
     mobilization of social resources. Keefe’s team        the presence in a community of collective
     will investigate whether the social support a         institutions – organizations, associations,
     person receives can be a form                                            social groups or networks.
     of social credit that is built up                                        The type of social capital on
                                               From an analytical
     over time. For a better under-                                           offer will therefore vary with
     standing of these specific                  perspective, it is           the relational systems that
     dynamics, further stages of           important to distinguish           predominate in each type of
     analysis will complement the          the potentially available          institution. For example, a
     study of the supportiveness               social capital from            government organization is a
     of the communities. Phase 2                the social capital            manifestation of bureaucratic
     of the study calls for a survey                                          social capital. This typology
                                               actually mobilized.
     of a sample of seniors to                                                makes it possible to document
     examine the role of specific                                             the way that different commu-
     factors related to social support, particularly       nities co-ordinate the supply of social capital in
     social networks and types of social support.          different economic situations, e.g., in periods
     A third phase consisting of case studies will         of decline or revitalization. Supply could be
     conclude the project, providing a clearer             dominated by certain institutions that are pri-
     understanding of how the socio-spatial context        marily based on one type of relation (e.g.,
     affects the health and independence of seniors        bureaucratic), something that automatically
     living in different types of rural communities.       entails certain considerations as to the rules
                                                           whereby this capital is to be mobilized by
     Measuring the capacity of                             the stakeholders.
     communities to produce social capital
                                                           As for the collective use of social capital,
     Bill Reimer, Professor of Sociology, Concordia
                                                           Reimer recommends that it be measured at
     University
                                                           the household level and then aggregated to
     As part of a vast research project on the new         the community level. In his research project,
     rural economy, Bill Reimer became involved            Reimer measures use of social capital by
     in exploring the role of social capital in the        means of a household survey, distinguishing
     revitalization of communities in decline.             the source of the resources mobilized by indi-
     Instead of employing the “strong ties/weak            viduals according to the typology of the rela-
     ties” typology, Reimer distinguishes four over-       tional systems. Separate measurement of the
     lapping systems of social relations that can          availability and use of social capital in a given
     potentially produce social capital: market            community makes it possible to identify and
     relations, bureaucratic relations, associative        study situations where certain forms of social
     relations and communal relations. The under-          capital are available but not used, thus shed-
     lying hypothesis is that access to social capi-       ding light on certain research questions con-
     tal depends on the ability of individuals to          cerned with the more complex dynamics that
     co-ordinate the resources deriving from these         underlie community development.
     different systems of relations.
16
Expert Workshop On The Measurement Of Social Capital For Public Policy




SESSION 4:
Measuring the cause-effect link between social capital and
socio-economic outcomes
We now have many important data sources               and quality of life of the participants) and col-
that demonstrate correlations between forms           lective outcomes (capacity of local communi-
of social capital and economic outcomes. How-         ties). The CEIP is a long-term research and
ever, we often know very little about the causal      demonstration project designed to evaluate the
directions of these relationships. For example,       performance of a new economic integration
we may find that there are many diverse social        program that consists in offering employment
networks flourishing in prosperous communi-           insurance and income assistance recipients a
ties. Is it the dynamics of the social networks       “community-based” wage guaranteed for three
that allows the communities to prosper, or is         years (instead of benefits) in exchange for
it the community prosperity                                             their participation in various
that allow the diverse net-          The prospective nature of local projects. The projects
works to flourish? Or, indeed,                                          are set up by the communities
                                      a study makes it possible
are there other mediating fac-                                          themselves (13 communities
tors at work?                            to define the causal           took part in the study). In
                                        direction between an            addition to offering the chance
The fourth block of presenta-
                                     intervention, the creation to acquire work experience,
tions was designed to answer
                                         of social capital and          the CEIP allows the partici-
the following questions:
                                          the individual and            pants to enrich their social
• How do we assess the spe-                                             capital by broadening and
    cific impact of social capital       collective outcomes.
                                                                        strengthening their network
    on individual and collective                                        of social relations. To the local
    outcomes?                                         communities it offers a work force, financial
• And how do we evaluate the specific impact          resources and professional support to ensure
    of individual outcomes on social capital?         the activities are viable.
• What are the most effective empirical               Darrell Kyte explained that, methodologically,
  approaches and tools for measuring these            the strategy was deployed differently at the
  complex dynamics?                                   individual and community levels. Individual
                                                      outcomes were evaluated by means of quali-
Measuring the influence of social                     tative interviews of participants and focus
capital on individual and collective                  groups. A resource generator and other ques-
outcomes through experimental                         tions suited to network analysis were used to
research: example of the CEIP                         conduct the study of the social capital of CEIP
David Gyarmati & Darrell Kyte, Social                 participants. The resource generator was
Research and Demonstration Corporation                adapted so as to identify the presence of
The Community Employment Innovation Pro-              resources relevant to the desired labour mar-
ject (CEIP) offered the researchers a unique          ket outcomes, namely: help finding a job, spe-
opportunity to study the effect of public inter-      cialized advice, emotional support and help
vention on the creation of social capital, as         with household activities. The questionnaire
well as the influence of the social capital thus      also contained a number of questions on the
created on individual outcomes (employability         three key dimensions of social networks: size,
                                                                                                            17
     Discussion Paper




     density and heterogeneity. Community out-             associations, social networks, families or the
     comes were assessed by means of audits of             workplace, it is important either to sample the
     the vitality of community organizations and           respondents on the basis of these particular
     interviews with community informants.                 aspects or to at least collect data documenting
                                                           them, so that they can be analyzed as inde-
     David Gyarmati explained that analysis of
                                                           pendent variables. In some cases, it would
     the dynamic between the CEIP, formation
                                                           even be helpful to consider associations or
     of social capital and outcomes was made
                                                           groups as units of analysis, in order to comple-
     possible thanks to the prospective nature of
                                                           ment the analysis of social capital with the
     the research. In studying the configuration of
                                                           organizational viewpoint.
     the networks of experimental groups before
     and after the Project and com-                                            Another simple way of
     paring with the control groups,                                           addressing the causality issue
     the researchers managed to                  A simple way of               is to introduce dynamic vari-
     isolate the mediating effect of        addressing the causality           ables in the survey instru-
     the networks on the employa-             issue is to introduce            ments, that is, variables
     bility of the participants. The        dynamic variables in the           capable of offering an indi-
     same approach was applied to              survey instruments.             cation as to the significance
     evaluate the influence of social                                          of the relationship between
     capital on the vitality of the                                            social capital and the out-
     local organizations in each community.                comes measured. The introduction of ques-
                                                           tions for tracing the sequence of past events
     The role of dynamic variables,                        (retrospective questions) or questions that pro-
     retrospective studies and adapted                     vide an idea of the duration of the events can
     research plans in measuring the                       be an alternative to more complex and expen-
     influence of social capital                           sive longitudinal surveys. For example, to bet-
     Dietlind Stolle, Assistant Professor,                 ter understand the genesis of a person’s social
     Department of political science,                      abilities, Stolle cites the interest of questions
     McGill University                                     that investigate the duration of membership in
                                                           a given group, length of residence in a given
     According to Dietlind Stolle, the advancement
                                                           neighbourhood, or the circumstances that
     of research on social capital must be supported
                                                           allowed a person to create ties with given
     by the development of a number of tools that
                                                           members of his or her network. Stolle said that
     are able to take better account of: 1) causality
                                                           panel research is obviously the most rigorous
     (better differentiation of the causes from the
                                                           method of studying causality, but combining
     effects of social capital); 2) comparisons
                                                           other techniques can also prove very effective.
     between different social groups (creation of
                                                           Finally, she mentioned the importance of quasi-
     indicators of social capital); and 3) changes
                                                           experimental designs where the research activ-
     over time (creation of time series).
                                                           ity commences right from the initial phase. To
     On the issue of causality, Stolle stressed the        conclude, Stolle emphasized the importance
     importance of better documenting a number             of incorporating data on attitudes, and specifi-
     of contextual factors, for it is on the basis of      cally on trust and reciprocity, for these are
     context that the conditions for deployment            dimensions of social capital that are just as
     of social exchange are created. Whether at            fundamental as the networks themselves.
     the level of the dynamics of neighbourhoods,


18
Expert Workshop On The Measurement Of Social Capital For Public Policy




The interest of an integrated                          constructed based on a version of the name
longitudinal survey for measuring                      generator that inventories ties to significant
the influence of social capital:                       persons. This instrument is chosen in order
example of the Socioeconomic and                       to explore in greater depth such problems as
Health Integrated Longitudinal                         social support, sociability and isolation. For
Survey                                                 certain survey cycles, complementary ques-
Paul Bernard, Professor of Sociology,                  tions similar to those used with a resource
Université de Montréal, and Johanne                    generator have also been added, to explore
Charbonneau, Research Professor, INRS                  certain themes such as emotional support,
urbanisation-culture-société (Montreal)                financial assistance and advisory and infor-
                                                                         mation resources. Other
The proposed Socioeconomic                                               complementary questions,
and Health Integrated Longitu-
                                        A life-course approach           for example on the frequency
dinal Survey (SHILS), as pre-                                            and quality of contacts or rela-
                                        allows the study of the
sented by Paul Bernard and                                               tional skills, offer means of
Johanne Charbonneau, shares            conditions under which
                                                                         better understanding certain
certain points in common with             social capital can be
                                                                         strategies for mobilizing rela-
the British Panel Survey. It is      accessed, based on specific
                                                                         tional networks.
a longitudinal survey vehicle          situations and contexts.
that covers a range of issues in                                         In terms of analytical poten-
the fields of health, education,                                         tial, the longitudinal nature
family life and the labour market, thus making         of the survey will allow us to track the chang-
it possible to study the interrelations between        ing composition of social networks over time
the different facets of individuals’ lives and         and understand how they are transformed
how they change over time. The SHILS con-              with successive events in the life cycle of
cept is based on a life-course approach, which         individuals. For research on social capital,
can situate issues within very specific life con-      the main interest of such a survey lies in the
texts (major life transitions, disruptive events,      possibility of enhancing knowledge of the
new personal or interpersonal situations, etc.).       conditions of access to the resources produced
The survey is constructed to be able to explore        by networks based on specific situations and
certain details by means of “factual modules”          contexts. It also affords a better understanding
that are applied when respondents are report-          of how, when certain circumstances arise, indi-
ing specific experiences.                              viduals combine social capital with other types
                                                       of resources (personal, community, institu-
For the study of social capital, the SHILS             tional) to achieve different socio-economic
planned to include a module exclusively                and health outcomes.
devoted to social networks, which is




                                                                                                            19
     Discussion Paper




     CONCLUSION:
     Advantages and limitations of the main tools of measuring social
     capital for the development and evaluation of federal policies
     and programs
     Comments and reflections from Doug                    2. Norris also questioned the relevance of
     Norris, Director General, Census and                     investigating social capital by means of
     Demographic Statistics Branch,                           major surveys such as those produced by
     Statistics Canada                                        Statistics Canada. It may be that, at this
                                                              stage, social capital research can benefit
     Doug Norris noted that interest in research              more from the flexibility of smaller-scale
     on social capital continues to grow, even in             experimental studies.
     the face of persisting ambigui-
                                                                              3. The conceptualization and
     ties about the concept. He
                                            Analysts are invited to              measurement of social capi-
     mentioned the contribution
                                             identify the analytical             tal based on a network
     that can be made by analysis
                                                                                 approach represent a fairly
     of recent databases closely           potential of the available
                                                                                 new way of exploring this
     associated with the concept             data in light of recent             field of research. Norris
     that have been produced by             developments in social               admitted that Statistics
     Statistics Canada, notably the
                                                 capital research.               Canada has not had the
     Ethnic Diversity Survey and
                                                                                 opportunity to consider
     the General Social Survey on
                                                                                 social capital from this
     Social Engagement. Keeping in mind that these
                                                               perspective, and at the moment does not
     major surveys were developed while social
                                                               have at its disposal measures expressly
     capital research was still in its infancy, analysts
                                                               designed to document social networks.
     were invited to identify the analytical potential
                                                               This is a subject of definite interest, but
     of the available data in light of recent develop-
                                                               one whose sensitive nature also raises
     ments in research on social capital.
                                                               considerations that will require strict
     Norris briefly summarized the lessons for                 evaluation of the performance of the
     Statistics Canada that could be drawn from                measurement instruments.
     the presentations heard during the day:                4. Mr. Norris also recognized the interest of
     1. Social capital can be either approached              better documenting social capital at the
        generally or associated with specific issues.        community level. It would be advisable to
        In all cases, context appears to be a key            include in the household surveys some ques-
        to understanding the way that social capital         tions that can explore the collective dimen-
        is deployed. In terms of measurement, Nor-           sion. The introduction of dynamic variables
        ris finds that this raises the question of the       or retrospective questions, as suggested by
        limitations of the very general surveys,             Stolle, is certainly one avenue that should be
        which cannot satisfactorily document con-            further explored.
        textual elements so as to meet the require-        5. Social capital is not simply a dependent vari-
        ments of social capital research. It would            able whose interest lies in its determinants.
        probably be more promising to consider                It is also a variable that explains many
        incorporating specialized modules within              socio-economic and health outcomes. It is
        thematic surveys.                                     essential to include in the measurement
20
Expert Workshop On The Measurement Of Social Capital For Public Policy




  instruments variables that document out-               of questions on social capital. Certainly it
  comes in a more satisfactory manner. Here              would be helpful to replace a few existing
  Norris called upon the policy designers, who           questions on social ties, assistance or sup-
  are in the best position to clarify specific           port, for example, with “revisited” questions
  needs in this area.                                    on social networks that might afford more
6. Quantitative measurement instruments are              rigorous measurement of social capital. On
   of limited benefit to research on social capi-        the other hand, however, this would entail
   tal. Qualitative research                                            the interruption of time series,
   is essential and must be                                             something that is not always
   maintained in order to                lt will be important           desirable from the standpoint
   complement the knowl-                 to better document             of historical comparability.
   edge afforded by statistics.                                         Finally, the in-depth revision
                                         social capital at the
   Statistics Canada must                                               of the General Social Survey
                                           community level.             that Statistics Canada is
   make greater use of alterna-
   tive approaches to measure-                                          preparing to undertake,
   ment of social capital in the interest                based on a life-course approach, might
   of improving its own measurement                      also be an interesting opportunity to
   instruments.                                          introduce a few questions on social
                                                         networks and social capital.
7. Also of interest is comparative analysis of
   social capital at the country level. Norris
                                                      Comments and reflections from
   mentioned that the proposal to include a
                                                      Derek Hum, Professor, Faculty
   specialized module on the subject in the
                                                      of Economics, University
   international social survey program is still
                                                      of Manitoba
   being reviewed, and that it might be appro-
   priate to draw inspiration from it.                Derek Hum’s thoughts related to certain ques-
8. In closing, Norris reviewed some of the chal-      tions raised in the workshop regarding the
   lenges posed by the introduction of new            strengths and weaknesses of the suggested
   subjects within the major Statistics Canada        methods of measuring social capital, and in
   surveys. Since survey space is scarce and          particular their relevance to public policy. The
   expensive, there will have to be solid sup-        complete version of Professor Hum’s commen-
   port for the more permanent incorporation          tary follows.




                                                                                                           21
     Synthesis Report




     MEASURING SOCIAL CAPITAL FOR PUBLIC
     POLICY: COMMENTS AND REFLECTIONS
     Derek Hum                                            social networks, most of it having to do with
                                                          Canadian data and applications. The summary,
     Introduction                                         then, is based upon my understanding of their
                                                          spoken comments rather than any subsequent
     The Policy Research Initiative (PRI) has been
                                                          written text produced. As well, my remarks
     engaged in explorations of the notion of social
                                                          attempts to reflect the understanding reached
     capital and its potential to assist public policy.
                                                          through subsequent discussion among the par-
     Following a series of consultation activities
                                                          ticipants at the workshop.
     with government as well as the academic
     research community, PRI has narrowed its             The synthesis comments reflect my personal
     focus on an analytic framework that empha-           emphasis, presented without attribution to
     sizes a network approach to understanding            individual speakers, in my charge as a work-
     social capital, deeming this strategy to be most     shop discussant and commentator. Finally, I
     amenable and helpful to policy development           attempt to relate the approaches and difficul-
     and delivery organizations. At the same time,        ties of measuring social networks to what I
     the directions pursued by researchers in their       perceive to be possibly fruitful avenues for
     many different emphasis and interests in social      public policy design, delivery and program
     networks do not often transfer to immediate          evaluation – issues that were less drawn
     usefulness to specific policy requirements. The      together due to time pressures on the parti-
     measures developed by academic researchers           cipants at the workshop.
     from time to time to study particular questions,
                                                          In sum, what follows represent discussant
     and in different countries and institutional con-
                                                          remarks in response to spoken presentations
     texts, and employing various data sets and
                                                          by experts with experience measuring social
     investigative techniques are often not readily
                                                          networks. The social network approach is the
     applicable to policy purposes. This state of
                                                          one selected by PRI for considering the role
     affairs gives rise to the question of how best to
                                                          and potential of the social capital paradigm
     measure social networks for particular policy
                                                          for public policy concerns. The framework
     purposes. To that end, a workshop of policy
                                                          and emphasis is upon measurement of social
     professionals and academic experts on the
                                                          networks, and I will offer some additional
     measurement of social networks was convened
                                                          comments expanding on the potential for
     (June, 2004) in which researchers studying
                                                          public policy purposes. To this end, the next
     social networks reported on their different
                                                          section characterizes the social network and
     measurement approaches, all the while empha-
                                                          some of the salient attributes discussed by
     sizing the particular focus of their investigative
                                                          the workshop presenters. This is followed
     interest and, perhaps, more important, the data
                                                          by a discussion of measurement approaches
     at hand or available in Canada.
                                                          employed to investigate social networks.
     The remarks that follow represent a selective        The data requirements of different measure-
     summary and synthesis of the oral presenta-          ment approaches is mentioned in general
     tions made (or at least heard by me) by the          categories, and related to the type of policy
     research experts as they described their differ-     questions that can be addressed. Finally,
     ent approaches and experience with measuring         some suggestions for policy needs are
22
Expert Workshop On The Measurement Of Social Capital For Public Policy




mentioned with respect to measuring social           relationships between individuals and groups,
networks, and why this might be useful for           or between groups), with the network con-
program assessment and monitoring.                   ceived as the collection of all such intercon-
                                                     nected links. Put just slightly more formally,
Characterizing the Social Network                    consider two individuals, Ian (represented by i)
and its attributes                                   and Jane (represented by j). The relationship
                                                     (R) defining the link or tie between Ian and
The notion of social capital as a network-based
                                                     Jane can be conveniently represented as “iRj”.
approach is clearly set out in a PRI draft dis-
                                                     The set or collection of all such relationships
cussion paper (PRI: Social Capital: Building
                                                     between any pair of individuals i and j will then
on a Network-based Approach, October 2003).
                                                     comprise the network to be measured, charac-
Additionally, the rationale and advantages
                                                     terized and studied. The point of this formal
for adopting a social network approach are
                                                     statement is to allow a clearer distinction of
discussed and need not repeated here. Con-
                                                     measurement issues that pertain to the rela-
trasting the network approach with the func-
                                                     tionship (R) itself, and the sampling frame (all
tional viewpoints associated with Coleman
                                                     the individual i’s and j’s) that might be surveyed.
or Putnam, the network approach to social
capital “refers to the network of social rela-       What might we wish to measure in a particular
tions that provide access to resources and           social network? What sort of relationships
support” (ibid., 2). It is worthwhile quoting in     should we highlight? How is the social net-
full the elaboration by PRI of a social capital      work to be characterized? These questions
framework deemed appropriate for examining           are rhetorical and the answers obviously
policy issues:                                       depend upon context, sample selected, and
                                                     most important, the issue under investigation.
  A social capital framework to support
                                                     In the present context, the formulation has to
  research and policy analysis uses the
                                                     do with particular public policies and govern-
  core network concept but is multidimen-
                                                     ment programs. Yet, it is possible to catalogue
  sional. Depending on the particular
                                                     some of the general characteristics of net-
  research and policy application, social
                                                     works that characterize the relationships
  capital studies should encompass, for
                                                     studied by the experts that reported, and
  example, the investments that individuals
                                                     to extrapolate their applicability to specific
  and collectivities make in the establish-
                                                     policy issues. We do this primarily to establish
  ment and maintenance of social net-
                                                     a background template for discussion.
  works, the various characteristics of the
  networks and transactions, the norms               The academic studies presented in the work-
  and institutional frameworks in which              shop suggest that social networks may be
  such networks operate, the resources               expected to vary in a number of attributes that
  that can be potentially accessed through           must be borne in mind when employing the
  participation in the networks, and the             social network approach for policy planning
  returns to those investments in the form           and assessment. We list some selected attrib-
  of economic , social, and health out-              utes of social ties or networks, in no particular
  comes for individuals, communities and             order and, we repeat, without attribution.
  societies. (ibid.).
                                                     Strength of relationship in Network
The primary or primitive element to define           Many authors alluded to the strength or weak-
or measure is the “tie” or “link” between two        ness of individual relationships (iRj) and, a for-
individuals (other possibilities would include       tiori, the strength or weakness of the network,
                                                                                                           23
     Synthesis Report




     when conceived as the collection of all              and the like, the relevant domain of interest
     such relationships for an individual. Other          will vary, and accordingly, the extent and
     metaphors such as “depth” or “breadth” or            boundaries of the social network of relations
     “thinness” could also be used to characterize        to be studied. The relationship (represented
     social networks. The importance for policy           hereafter as “ R = ”) of interest may be, for
     studies is to recognize the appropriateness          example,
     and efficacy of “strong ties” or alternatively,
                                                          R = “Do you know a doctor specializing in –?”
     “weak ties” in specific contexts. For example,
                                                          or
     a policy perspective focused upon social sup-
     port networks for seniors, or those with health      R = “ Do you know anyone who came to
     impairments might wish to measure “strong            Canada in the last (five, ten …) years from – ?”
     ties” in a social network that delivers resources
                                                          R = “Do you know someone who could
     and support to this group. On the other hand, it
                                                          translate a government form from French
     is often suggested that job search and employ-
                                                          to Chinese for you?”
     ment opportunities are best achieved through
     a social network characterized by “weak” ties
                                                          Intensity of the relationship in the
     that are numerous and diversified. In plain lan-
                                                          Network
     guage: the determination and measurement of
     a relation R through asking a respondent:            A relationship is also characterized by its
                                                          “intensity”, or some such phrase. That is, the
     R = “Do you know anyone who would look
                                                          relationship may be one of long standing with
     after you at home for a month if you were to
                                                          frequent contacts, such as that among close
     suffer a stroke?”
                                                          family members. On the other hand, a relation-
     implies interest in a strong tie. This is decid-     ship may be a casual acquaintance in which
     edly different from a relationship (say, R*)         support and transfer of sought resources is ten-
     asking about a weak tie, such as:                    uous and unreliable; e.g., persons who share
                                                          only an interest in supporting the same sports
     R = “Do you know anyone who might give
                                                          team. Merely detecting a relationship’s exis-
     you a letter of reference or assist you to find
                                                          tence is often insufficient for some policy
     a job?”
                                                          applications, and the degree of intimacy must
     Specialized (or specific) nature of                  be gauged as to the amount of support or
     relationship in Network                              resources that might be forthcoming. Addition-
     The above example highlights the relevance           ally, the qualification attached to the expected
     of determining the appropriate set of relation-      support must be measured.
     ships to measure, and that relationships are
                                                          R = “ Do you know anyone who would lend
     often specific or specialized. In short, people
                                                          you $500 if you need it?”
     access different social networks for different
     purposes, and measurement of a social net-           R = “ Do you know anyone who would let
     work presupposes some clearly articulated            you live with them for three months if you
     focus of the issue to be addressed. For exam-        lost your home?”
     ple, depending whether the policy focus or
                                                          R = “Do you know someone who would donate
     program under discussion concerns seniors,
                                                          an organ to you if you need it?”
     labour market participants, children, immi-
     grants, visible minorities, persons with disabili-   Volition of relationship in the Network
     tie, university students, specific occupation        Individual relationships may be voluntary
     (professionals) or a particular sector (farming)     or “compulsory” and the distinction can be
24
Expert Workshop On The Measurement Of Social Capital For Public Policy




important in defining social network patterns.       relationship is one in which iRj and jRk implies
A person’s ethnic affiliation and membership in      that iRk. Transitivity need not obtain. Ian may
a social network may not be entirely voluntary,      enjoy a relationship with Jane (be willing and
nor is one’s relationship to parents or children.    able to lend money to, or work with, or help
Indeed, others often prescribe social networks.      with job search), and Jane may also enjoy a
Parents or guardians determine a child’s             similar relationship with her friend Keith, but
schooling, religion, certain sports and recre-       Ian may not have the same or any relationship
ation activities, friends, and even acceptable       with Keith. The key concept is the nature by
marriage partners. Hence a child’s social net-       which individual relationships extend through-
work is quasi determined, and this network           out an inclusive social network through some
influences and conditions the networks that          “chaining” property, and this feature ultimately
the parents experience as well. More impor-          defines the efficacy and utility of the social
tant, these social networks affect the set of        network itself. It is not necessarily so that
potential future networks that these children        any friend (or enemy) of yours is a friend (or
confront in youth and possibly adult life.           enemy) of mine and vice versa.Clearly, meas-
Whether relationships and the associated             urements must be carried out at the individual
social networks result from voluntary associa-       level as well as the collective (or entire net-
tion can be a key ingredient in understanding        work) level for complete understand-ing of
the operations of the social network. A per-         social networks. (Relationships that satisfy
son’s social network at one time may well be         symmetry and transitivity are known as equiv-
an important determinant of their future net-        alence relationships, and their mathematical
work relations.                                      properties well known. On the other hand,
                                                     almost all relationships of interest in a social
R = “ Are you a member of (a youth gang)
                                                     network framework are of the non-equivalent
(a Church) (a visible minority group)
                                                     variety, and it is this very heterogeneity that
(a union) …?”
                                                     is of interest.)
Symmetry and equality of relationship
                                                     Relative position in Social Network
in Networks
                                                     The need to encompass relationships in a
There are many features of a relationship that
                                                     social network beyond simple pairs is appar-
could be ascertained in a measurement, such
                                                     ent. This requirement is in additional to any
as symmetry, equality and transitivity. A rela-
                                                     exploration of the conditions or contingencies
tionship is symmetrical if iRj implies simultane-
                                                     under which the relationship itself holds. Some
ously jRi. It may be the case that Ian claims a
                                                     relationships are valid only under specific cir-
relationship with Jane (such that j would lend
                                                     cumstances, such as emergencies, or will strain
i a large sum of money) but the reciprocal rela-
                                                     under repeated demands, and the like. This is
tionship does not exist (Jane does not expect
                                                     as expected. The “relative” position in a net-
to borrow money from Ian). Or the “strength”
                                                     work refers to the existence of a relationship
of the tie in iRj is not the same as the jRi rela-
                                                     iRj that is interpretable only with respect to
tionship. Jane may be willing to shop for Ian
                                                     some other individual, say k. For example, sup-
when Ian is sick, but Ian is not willing or able
                                                     pose iRj is a relationship that denotes that i can
to do the same for Jane in like circumstances.
                                                     access the resources of j (say, borrow $1000)
Finally, a social network involves more than
                                                     but the relationship is contingent (or under-
simple relationship between two individuals,
                                                     stood) that k has priority over i. For example,
and the pattern of relationship “chains” are
                                                     iRj might be: (illustrative only)
worthy of measurement and study. Indeed,
they are often of most interest. A transitive
                                                                                                          25
     Synthesis Report




     R = Son (i) is allowed to borrow father’s (j)         For example, as suggested by one researcher
     car (but only if mother (k) does not need it).        at the workshop, a question might ask:

     R = Individual (out of province) can have             R = “Over the past six months, can you name
     access to medical care (but only if some              individuals with whom you discussed issues
     provincial resident does not require the              of personal importance to you?”
     service).
                                                           Note that the question is open-ended with
     R = Individual can be hired (but only if there        respect to the number of names asked, and the
     exist no other qualified visible minority candi-      subject matter deemed important. The instru-
     date).                                                ment may also specify qualifiers, such as “out-
                                                           side the home” or “good friends you’re close
     The upshot of the above is simply that the
                                                           to” and the like. The aim is to generate a frame
     entire social network is relevant for some
                                                           of names (and to follow up on the specific
     applications, and not merely the set of relation-
                                                           names, if desired). Researchers at the work-
     ships or links between individuals examined in
                                                           shop have pointed out that the domain of the
     isolation. However, measuring an entire social
                                                           social network is egocentric and limited by
     network requires, as we shall see, much more
                                                           the knowledge and recall of the respondent (as
     effort and resources.
                                                           well as the usual limitations of all surveys), and
     The above list of attributes of social networks       tends to nominate “middle strength” ties, rather
     is neither exhaustive nor do they represent a         than the really weak or really strong ones.
     consensus among researchers. To repeat, the
                                                           With the above as background and an example,
     list is a composite of elements discussed by
                                                           one could easily see how this instrument can,
     the experts present at the workshop as well
                                                           and has been, varied. The name generator
     as my own interpretative interjections. Though
                                                           could be employed as a “position generator
     incomplete, the list should give some hint of
                                                           instrument” in which individuals are asked
     the difficulties to be encountered in measuring
                                                           about their relationships with persons who
     social networks. This is considered next.
                                                           satisfy a “positional” criterion:

     Measuring the relationships and                       R = “Do you know (a medical specialist, a legal
     Social Network                                        expert, a mason)? The position generator can
                                                           give an idea of the diversity of the network
     We now turn attention to the methods and
                                                           of a person.
     techniques employed by many researchers to
     measure social networks, confining attention          Yet a third possibility is the “resource genera-
     to those approaches that employ some sort of          tor”.
     survey based approach in contrast with eth-
                                                           R = “Do you know anyone who (would lend
     nomethological studies or narratives from field
                                                           you $5000, owns a cottage, maintains a private
     observations. With this limitation, it is perhaps
                                                           fishing lodge, has box seats to a sports event,
     useful to describe a common technique, such
                                                           has a cordless drill…)?
     as the “name generator instrument”. Recogniz-
     ing the inadequacy of simply asking an individ-       This approach is “resource driven” as opposed
     ual to give a list of names of their “friends”, the   to “name driven”.
     name generator instrument asks questions of
                                                           All three approaches share a common meas-
     specific relationships and specific contacts.
                                                           urement approach, and it might be helpful


26
Expert Workshop On The Measurement Of Social Capital For Public Policy




to restate the procedure in survey research          the (policy) question at hand, and to define the
jargon. Essentially, the “generator” survey          domain of the social network so as to charac-
method can be likened to the first stage a           terize it properly.
“snowball” sampling technique which is typi-
                                                     For emphasis, it is useful to restate the prac-
cally used when there is no predetermined or
                                                     tice from a policy or program delivery per-
available sampling frame to begin with. A par-
                                                     spective. From a sampling survey perspective,
ticular program may have a list of its partici-
                                                     the kernel or “generator” (whether name, posi-
pants from administrative files to serve as a
                                                     tion, resource etc.) is used to construct the
sampling frame. On the other hand, a third
                                                     domain of the social network of interest. The
party investigator without access to adminis-
                                                     generator requires a screening question to
trative files may have to determine the relevant
                                                     determine the respondent’s eligibility for fur-
study group by surveying the general popula-
                                                     ther query, and the screen must obviously be
tion with some screener question. The set of
                                                     framed with an underlying purpose of inquiry.
respondents (and thus, social network to be
                                                     We are assuming that there does not exist a
studied) is obtained by specifying a “kernel”
                                                     predetermined sampling frame such as an
(generator) that leads to a list of respondents
                                                     administrative file. For example, if one is inter-
from whom further information (and/or
                                                     ested in the social network of elderly immi-
“names”) is gathered. Accordingly, the screen-
                                                     grants, a relevant screener question must be
ing question of the generator (along with any
                                                     first administered to determine if the respon-
qualifiers) serves the pivotal function of defin-
                                                     dent is, in fact, elderly and an immigrant.
ing the domain boundaries of the social net-
                                                     Reversing this description would start the
work as well as the eligibility threshold of
                                                     practice with determining the policy question
participation. For example, to study social net-
                                                     of interest or the program to be assessed to
works constituting strong (rather than weak)
                                                     determine the set of respondents whose net-
ties that involve access to resources the screen
                                                     work relationships are of prime interest.
question generator might be:
                                                     A concrete example might be a social service
R = “Do you know someone who would lend
                                                     program concerned with home care services
you money for a down payment for a house
                                                     for the elderly or persons with functional
(rather than $10.00)?”
                                                     limitations. The nature of the network of
The implicit assumption in the above example         interest might not concern access to financial
is that the amount of money sought for a loan        resources so much as certain “in kind” volun-
is an indicator of the strength of the tie, with     tary resources such as assistance with day to
a sum of $10 representing a weak tie. The sum        day shopping, transport assistance to medical
typically involved in a down payment for a           care, occasional conversation and concern,
house is large, and would be obtained only           and simple monitoring of daily coping. Since
from a person with a strong tie.                     these resources must be rendered by persons
                                                     in close geographical proximity to the respon-
In sum, the study and measurement of social
                                                     dent, the geographical scope of the social net-
networks involving relationship between and
                                                     work is also prescribed. If so, then the screen
among individuals is beset with a number of
                                                     question of the generator should target these
measurement difficulties. For example, the
                                                     types of relationships of the social network,
requirement to define the nature of the rela-
                                                     and the survey designed accordingly. Questions
tionships of interest, its salient attributes for



                                                                                                          27
     Synthesis Report




     pertaining to other types of social networks        special surveys. The result, then, is survey data
     (such as pen pals in distant countries) are not     that is characterized by “small” sample size
     of direct interest. This is an example only, but    and “cross section” in nature. The limitations
     one can readily foresee that each program           of this type of data are well known and need
     delivery could have a different social network      not be catalogued again.
     of primary interest, thus making the possibility
                                                         It is appropriate, however, to point out the limi-
     of measuring a general purpose all-encompass-
                                                         tations or range of issues that are precluded
     ing social network less appealing for specific
                                                         from study with small cross sections. Many of
     policy questions.
                                                         the workshop experts mentioned that relation-
     We make one further distinction without much        ships (R(t)), and a fortiori, the social network,
     discussion here. If a sample is only asked          would change over time and altered circum-
     about their contacts without further follow up      stances. These cannot be captured with cross
     of the nominated contacts, this would consti-       sectional data that can only portray the social
     tute merely the “contacts list” of the set of       network at a single instant. More specifically,
     respondents. We might, as suggested, consider       iR(t)j may alter simply with the passage of time
     this the first stage of a snowball sample. With-    t, as either i(t) or j(t) mature, or circumstances
     out further investigation of the contacts of the    change. For example, the resources of j (who
     contacts (the second round), we cannot truly        was willing to lend them to individual i at time
     characterize a “network” since the multi-linked     t) may have since dwindled so that i’s expected
     relationships cannot be known. Therefore it         access is no longer realistic. Or person k is no
     would not be possible to characterize the           longer is a position to offer employment assis-
     entirety of the social network itself. Perhaps      tance to j since k’s set of social contacts have
     an example will help. Suppose we ask a small        been transformed for some reason. Take
     number (n) of respondents to “name” a small         another whimsical (?) example. The social net-
     number of “contacts” (say, two contacts). If n      work and “strong” ties among a set of bache-
     is small relative to the study population of        lors may undergo radical transformation upon
     interest, and we do not contact the contacts        marriage. Which relationships are maintained?
     named by the first round, we may not be able        Abandoned? Or transformed or added to? The
     to detect anything of significance of the entire    list of examples goes on, but the salient point
     social network, such as its total size, density,    is simply that cross sectional surveys cannot
     diversity, spatial distribution etc. In sum, the    “track” changing or transforming social net-
     distinction between merely gathering a list of      works. But for some policy issues (such as pro-
     peoples’ contacts, and characterizing the social    gram monitoring of outcomes of interventions,
     network of a collective group depends upon          etc.), the nature of the network over time, suc-
     whether the snowball sampling is truncated at       cess and failure of relationships, and the like
     the first stage or not.                             are of fundamental interest. For example,
                                                         how are the relationships of new immigrants
     The curse of data requirements                      changed over time with lengthing duration in
                                                         Canada? These issues of dynamic network
     Any attempt to gauge and characterize the
                                                         relationships can only be addressed with longi-
     social network of a set of individuals is a large
                                                         tudinal data. For many public programs, their
     task. Unlike studies that can employ compre-
                                                         administrative files or case data may prove a
     hensive data sets (such as the census), policy
                                                         more useful starting point for measuring net-
     specific questions that pertain to target popu-
                                                         work issues than an omnibus survey con-
     lations will typically have to rely on smaller
                                                         ducted by some third party for other purposes.
28
Expert Workshop On The Measurement Of Social Capital For Public Policy




The application to public policy                     surveys such as the GSS conducted infre-
concerns                                             quently on time schedules of Statistics
                                                     Canada’s choosing are less useful for parti-
The concerns of the policy research and pro-         cular programs and policies than would be
gram delivery community are often different          internally designed data gathering exercises,
from that of the general academic community.         often exploiting the program’s administrative
In particular, the program delivery community        data bases for a good sampling frame. How-
is accountable for monitoring their program          ever, it is still an “art” to design the right set
outcomes (and budgets) and assessing results         of questions and to interpret the findings. And
of their interventions. The unit of analysis is      it is highly unlikely that survey data can pro-
often a program’s target group rather than the       vide all the insights. The role of the intensive
general population. As well, the program has         case study, with open-ended focus group dis-
specific purposes, and their progress towards        cussion, is also important to achieve richly
these ends are assisted (or hindered) by spe-        detailed information.
cific social networks of the affected group.
This means that the study of social networks         Many researchers at the workshop alluded
in this context is less driven by idle curiosity     to the value of case studies in understanding
than the need for a specialized, specific pur-       social net-work patterns, particularly collectivi-
pose inquiry, whose results must be produced         ties. While these remarks have not emphasized
by a given deadline and within a given budget.       the measure-ment of social capital at the group
                                                     or community level, their study and measure-
This characterization (with much violence            ment remains fundamentally important.
to subtlety) suggest that large scale omnibus




                                                                                                          29
     Synthesis Report




     APPENDIX 1: WORKSHOP PROGRAM
     Expert Workshop On The Measurement Of Social Capital
     For Public Policy

     Session 1:         Approaches to analyzing social networks and their usefulness
                        in a public policy context
     Barry Wellman                  Professor of Sociology, University of Toronto
     Johanne Charbonneau Research Professor, INRS-Urbanisation, Culture et Société
     Maurice Lévesque               Professor of Sociology, University of Ottawa

     Session 2:         Measuring the constituent elements of social capital: What
                        exactly are we measuring, and how do we go about it?
     Peter Marsden                  Professor of Sociology, Harvard University
     Tom Snijders*                  Professor of methodology and statistics, University of Groningen
     Bonnie Erickson                Professor of Sociology, University of Toronto
     Jeff Boase                     Doctoral candidate, University of Toronto

     Session 3:         Measuring social capital at the community level
     Ralph Matthews                 Professor of Sociology, University of British Columbia
     Janice Keefe                   Research Chair on Aging and Caregiving Policy, Maritime Data Centre
                                    for Aging Research & Policy Analysis, Mount Saint Vincent University
     Bill Reimer                    Professor of Sociology, Concordia University
     Barry Wellman                  Professor of Sociology, University of Toronto

     Session 4:         The usefulness of social capital: How do we measure its real
                        effects?
     David Gyarmati                 Researchers, Social Research and Demonstration Corporation, Ottawa
     Darrell Kyte                   Researchers, Social Research and Demonstration Corporation, Maritimes
     Dietlind Stolle                Professor of Political Science, McGill University
     Paul Bernard                   Sociology professor, Université de Montréal
     Johanne Charbonneau Professor-researcher at INRS-UCS

     Roundtable: Elements of a toolbox for government action
     Derek Hum                      Professor of Sociology, University of Manitoba
     Doug Norris                    Director General, Statistics Canada

     * Bonnie Erickson replaced Tom Snijders using the presentation material that he had prepared for the workshop.


30
Expert Workshop On The Measurement Of Social Capital For Public Policy




APPENDIX 2: WORKSHOP INFORMATION
    organized by     Policy Research Initiative in partnership with the Social Sciences and
                     Humanities Research Council

 project director    Catherine Demers, Senior Policy Research Officer

   workshop lead     Sandra Franke, Policy Research Officer

PRI coordinators     Lori Brooks, Analyst
                     Louis-Phillippe Gascon, Technical Services Officer
                     Marissa Smyth, Administrative Assistant
                     Carole Tremblay, Operations Administrator
                     Élisabeth Vu, Linguistic Assistant

                     We gratefully acknowledge the support of Statistics Canada.



                      For more information please visit our website at
                      <www.policyresearch.gc.ca> and refer to the conference program.
                      Most of the presentations are available, along with a few reference
                      documents that were provided by the experts. They can be accessed
                      directly from the conference program.




                                                                                              31

								
To top