Co-exist, Colonize, or Combine?
Accounting for patterns of discourse on nonprofit evaluation

Carrie Oelberger, Achim Oberg, Karina Kloos, Valeska Korff, Woody Powell




                                                      EGOS Colloquium 2012
 Subtheme 20: Rationalization and Professionalization of the Nonprofit Sector
                                                 5 – 7 July - Helsinki, Finland
Is talk cheap?

Language is a constitutive feature of social life

  • creates cohesion and enforces social boundaries

  • reflects and reinforces common systems of norms

  • serves to enroll and exclude

   Creates communities
Crossing Boundaries


Boundary work as the establishment of local
interlanguages – pidgins and creoles – that emerge in
the interstices between social domains (Galison 1997).

Interlanguage
• facilitates local communication across social and linguistic
  boundaries.
• enables coordination of action across place, time and context.
 Connects communities
Nonprofit Evaluation

Influences from the domains of
• Civil society and associations
• Science
• Management

 Confluence of
  disparate world views
  and languages

Nonprofit evaluation discourse is an illuminating case to analyze
communication at the interface of social spheres
Research Questions


             Who is contributing what
     to the discourse on nonprofit evaluation?

1.   What kind of discourse patterns form when
     different languages come into contact?
     Does a boundary-spanning interlanguage emerge?

2.   What organizational features influence an entity’s
     proclivity to use such interlanguage?
Websites as self-representation

• Websites constitute purposeful presentations of organizations
• Information provided on websites has the capacity to travel
  across spatial and symbolic distances
• Symbolic representation of an organization whose structure
  and content reflect the features of a particular entity
  (Pollach, 2004)
• Comparable in its signaling character to national symbols:
  flags and anthems (Cerulo, 1993)
 Language used on websites reflects an organizations
   intentional portrayal of itself to diverse audiences
Sample selection by use of Webcrawler

•   Focus on websites:
    Websites as purposeful self-
    presentations
•   Crawler network:
    Websites are added based on
    number of incoming references by
    identified members of the
    relevant sample.  Snowball
    sample
•   Inclusion/exclusion decision:
    Collective analysis of website
    content to appraise extent of
    contribution to non-profit
    evaluation discourse.
•   Coding of entities:
    Websites are aggregated and core
    organizational features are coded.
Sample Characteristics

      419 highly interconnected entities involved in nonprofit evaluation
Demography
 • Age: 2 to over 200 years old
 • Size - Scale: one person blogs to 250,000 employee global organizations;
          Scope: local, regional, national and scale
Institutional Properties
 • Form: 56% nonprofits, 13% for-profits, 3% branches of state or national government, 14%
   transnational organizations, and 14% non-organizational forms
 • Activity: evaluation, funding, consulting, networking, media, advocacy, research, social services
Resources and Constituencies
 • Revenue Streams: foundation grants, government grants, corporate funding, individual
   donors, fee-for-services, membership fees, endowment, public equity market and taxes
 • Target Audiences: social service beneficiaries, donors, nonprofits, for-profits, (transnational)
   government and the public
Collecting Keywords

Keywords are “significant, indicative words in certain forms of thought” that
make up a distinctive, domain-specific vocabulary (Williams 1969: 14)
1) Iterative process of identifying keywords to develop a vocabulary of nonprofit
    evaluation:
     • Mined the discourse on the websites and consulted experts
     • Created word clusters (associational, scientific, managerial)
       Process resulted in 196 terms categorized in 3 clusters (e.g. participation, social
         change, justice = associational; data, systematic, framework = scientific;
         performance, benchmarks, outcomes = managerial)
2) Counting of occurrences of each term on all websites to calculate relative
   percentage of language used by an entity.
3) Co-occurrence analysis to affirm validity of clusters.
    Assessment of entities’ individual language use, and the collective pattern of
     language use among all entities in the sample
Models of Discourse Patterns


       Option 1: Entities co-exist, separated
                  into linguistic factions


       Option 2: Entities combine and blend
                  different languages.



       Option 3: One language colonizes and
                  dominates the discourse.
Relative language use


                                                              thesroinetwork.org
                                                                    swtgroup.net
                                                          cerise-microfinance.org
                                                           mullagofoundation.org
                                                                        efqm.org
                                                             corostrandberg.com
                                                                         ladb.org
                                                               sphereproject.org
                                                                   wdi.umich.edu
                                                                        usaid.org
                                                                      fbheron.org
                                                           rainforest-alliance.org
                                           organizationalresearch.com
                                                                      eandco.org
                                                               compasspoint.org
                                                                      arnova.org
                                                                          iisd.org
                                                                   robinhood.org
                                                             broadfoundation.org
                                                                     joycefnd.org
                                                                  gistfunders.org
                                                                       gmfus.org
                                                           americanprogress.org
                                                                         hfpg.org
                                                                    alliance1.org
                                                                         aecf.org
                                                        seechangeevaluation.com
                                                                  ncvo-vol.org.uk
                                                            nonprofitquarterly.org
                                                                   unstats.un.org
                                                             onphilanthropy.com
                                                             gatesfoundation.org
                                                                 worldofgood.org

Entirety of entities involved in nonprofit evaluation
                                                                         usip.org
                                                                   cofinteract.org
                                                           africagrantmakers.org
                                                                   unwomen.org
                                                                       komen.org
                                                          sunlightfoundation.com
                                                                         350.Org
                                                                    cafonline.org
                                                                                                             Observed Distribution of Discourse
Interlanguage?

Frequency analysis:
Terms used by 85% of entities
Co-occurrence analysis:
Non-particularity of terms,               Interlanguage

but used in combination
with diverse languages


 No random blending of discourses
 Ubiquitous and multi-lateral interlanguage vocabulary
Characteristics of the Interlanguage

Interlanguage is…
                                    Accountability        Lessons learned
• Ubiquitous in being used          Quantitative        M&E/Monitoring
                                    Assessment              (& evaluation)
  by all contributors               Evaluation               Effectiveness
                                    Framework                Performance
• Composed of elements of           What works               Transparency
                                    Indicators               Best practice
  original languages                Methods          23%      Certification
                                    Survey                      Outcomes
• Distinctive to nonprofit                     Accountability Evidence
                                               Commitment           Impact
  evaluation discourse                          Participatory
                                               Social change
                                               Transparency
                                                    Trust

     Interlanguage enables communication and coordination
              across social and linguistic boundaries.
Explaining Interlanguage Use

• H1: Younger organizations use more interlanguage.

• H2: Larger entities, both in terms of a) scale (size) and
  b) (geographic) scope use more interlanguage.

• H3: Organizations engaged in activities which involve
  competition or coordination with organizations of a different
  legal form – specifically evaluation, networking and consulting
  - use more interlanguage.

• H4: Organizations with a) diverse revenue sources or b) a
  heterogeneous audiences use more interlanguage.
Interlanguage Use - Findings




• H3: Organizations engaged in activities which involve
  competition or coordination with organizations of a different
  legal form – specifically evaluation, networking and consulting
  - use more interlanguage.
Interlanguage Use - Discussion

• The youngest organizations use less interlanguage, and more managerial.
• Size does not affect interlanguage use, suggesting that structural features
  do not condition language use.
• Organizations engaged in evaluation, networking, consulting, and research
  use more interlanguage.
• Type of revenue source and audience affects interlanguage use.
    – Organizations that speak to an audience of social service beneficiaries and the
      general public use less interlanguage, while those speaking to for profits use
      more interlanguage.
    – Organizations that draw resources through grants or endowments use more
      interlanguage.
• Standard features (form, size, revenue sources) remain influential on use
  of original discourses, yet not on interlanguage use.
Observed Distribution of Interlanguage
Relative language use




                                   Entirety of entities involved in nonprofit evaluation
Conclusion

• Broad use of interlanguage across organizational forms
• Interlanguage as a channel through which ideas get
  transported across social boundaries

• Evidence of integration in combination and interlanguage use
• Evidence of colonization in composition of interlanguage
• Different discourses remain relevant: Rational – Normative

                   Interesting times ahead!
Future Plans


• Relational analysis of the application of keywords
   – Nonprofit evaluation context as a densely interconnected
     reputation network among organizations and URLs
   – Identification of clusters
Topic clusters
Some topics are
clustered in specific
regions of the
reputation network.
Positioning of clusters
Other topics are
positioned on the top
and in the center.
Positioning of clusters
… some connect center
and periphery with an
inverse distribution of
size (marginalized
term?)
Thank you
for your attention!

Co-exist, Colonize, or Combine?

  • 1.
    Co-exist, Colonize, orCombine? Accounting for patterns of discourse on nonprofit evaluation Carrie Oelberger, Achim Oberg, Karina Kloos, Valeska Korff, Woody Powell EGOS Colloquium 2012 Subtheme 20: Rationalization and Professionalization of the Nonprofit Sector 5 – 7 July - Helsinki, Finland
  • 2.
    Is talk cheap? Languageis a constitutive feature of social life • creates cohesion and enforces social boundaries • reflects and reinforces common systems of norms • serves to enroll and exclude  Creates communities
  • 3.
    Crossing Boundaries Boundary workas the establishment of local interlanguages – pidgins and creoles – that emerge in the interstices between social domains (Galison 1997). Interlanguage • facilitates local communication across social and linguistic boundaries. • enables coordination of action across place, time and context.  Connects communities
  • 4.
    Nonprofit Evaluation Influences fromthe domains of • Civil society and associations • Science • Management  Confluence of disparate world views and languages Nonprofit evaluation discourse is an illuminating case to analyze communication at the interface of social spheres
  • 5.
    Research Questions Who is contributing what to the discourse on nonprofit evaluation? 1. What kind of discourse patterns form when different languages come into contact? Does a boundary-spanning interlanguage emerge? 2. What organizational features influence an entity’s proclivity to use such interlanguage?
  • 6.
    Websites as self-representation •Websites constitute purposeful presentations of organizations • Information provided on websites has the capacity to travel across spatial and symbolic distances • Symbolic representation of an organization whose structure and content reflect the features of a particular entity (Pollach, 2004) • Comparable in its signaling character to national symbols: flags and anthems (Cerulo, 1993)  Language used on websites reflects an organizations intentional portrayal of itself to diverse audiences
  • 7.
    Sample selection byuse of Webcrawler • Focus on websites: Websites as purposeful self- presentations • Crawler network: Websites are added based on number of incoming references by identified members of the relevant sample.  Snowball sample • Inclusion/exclusion decision: Collective analysis of website content to appraise extent of contribution to non-profit evaluation discourse. • Coding of entities: Websites are aggregated and core organizational features are coded.
  • 8.
    Sample Characteristics 419 highly interconnected entities involved in nonprofit evaluation Demography • Age: 2 to over 200 years old • Size - Scale: one person blogs to 250,000 employee global organizations; Scope: local, regional, national and scale Institutional Properties • Form: 56% nonprofits, 13% for-profits, 3% branches of state or national government, 14% transnational organizations, and 14% non-organizational forms • Activity: evaluation, funding, consulting, networking, media, advocacy, research, social services Resources and Constituencies • Revenue Streams: foundation grants, government grants, corporate funding, individual donors, fee-for-services, membership fees, endowment, public equity market and taxes • Target Audiences: social service beneficiaries, donors, nonprofits, for-profits, (transnational) government and the public
  • 9.
    Collecting Keywords Keywords are“significant, indicative words in certain forms of thought” that make up a distinctive, domain-specific vocabulary (Williams 1969: 14) 1) Iterative process of identifying keywords to develop a vocabulary of nonprofit evaluation: • Mined the discourse on the websites and consulted experts • Created word clusters (associational, scientific, managerial)  Process resulted in 196 terms categorized in 3 clusters (e.g. participation, social change, justice = associational; data, systematic, framework = scientific; performance, benchmarks, outcomes = managerial) 2) Counting of occurrences of each term on all websites to calculate relative percentage of language used by an entity. 3) Co-occurrence analysis to affirm validity of clusters.  Assessment of entities’ individual language use, and the collective pattern of language use among all entities in the sample
  • 10.
    Models of DiscoursePatterns  Option 1: Entities co-exist, separated into linguistic factions  Option 2: Entities combine and blend different languages.  Option 3: One language colonizes and dominates the discourse.
  • 11.
    Relative language use thesroinetwork.org swtgroup.net cerise-microfinance.org mullagofoundation.org efqm.org corostrandberg.com ladb.org sphereproject.org wdi.umich.edu usaid.org fbheron.org rainforest-alliance.org organizationalresearch.com eandco.org compasspoint.org arnova.org iisd.org robinhood.org broadfoundation.org joycefnd.org gistfunders.org gmfus.org americanprogress.org hfpg.org alliance1.org aecf.org seechangeevaluation.com ncvo-vol.org.uk nonprofitquarterly.org unstats.un.org onphilanthropy.com gatesfoundation.org worldofgood.org Entirety of entities involved in nonprofit evaluation usip.org cofinteract.org africagrantmakers.org unwomen.org komen.org sunlightfoundation.com 350.Org cafonline.org Observed Distribution of Discourse
  • 12.
    Interlanguage? Frequency analysis: Terms usedby 85% of entities Co-occurrence analysis: Non-particularity of terms, Interlanguage but used in combination with diverse languages  No random blending of discourses  Ubiquitous and multi-lateral interlanguage vocabulary
  • 13.
    Characteristics of theInterlanguage Interlanguage is… Accountability Lessons learned • Ubiquitous in being used Quantitative M&E/Monitoring Assessment (& evaluation) by all contributors Evaluation Effectiveness Framework Performance • Composed of elements of What works Transparency Indicators Best practice original languages Methods 23% Certification Survey Outcomes • Distinctive to nonprofit Accountability Evidence Commitment Impact evaluation discourse Participatory Social change Transparency Trust Interlanguage enables communication and coordination across social and linguistic boundaries.
  • 14.
    Explaining Interlanguage Use •H1: Younger organizations use more interlanguage. • H2: Larger entities, both in terms of a) scale (size) and b) (geographic) scope use more interlanguage. • H3: Organizations engaged in activities which involve competition or coordination with organizations of a different legal form – specifically evaluation, networking and consulting - use more interlanguage. • H4: Organizations with a) diverse revenue sources or b) a heterogeneous audiences use more interlanguage.
  • 15.
    Interlanguage Use -Findings • H3: Organizations engaged in activities which involve competition or coordination with organizations of a different legal form – specifically evaluation, networking and consulting - use more interlanguage.
  • 16.
    Interlanguage Use -Discussion • The youngest organizations use less interlanguage, and more managerial. • Size does not affect interlanguage use, suggesting that structural features do not condition language use. • Organizations engaged in evaluation, networking, consulting, and research use more interlanguage. • Type of revenue source and audience affects interlanguage use. – Organizations that speak to an audience of social service beneficiaries and the general public use less interlanguage, while those speaking to for profits use more interlanguage. – Organizations that draw resources through grants or endowments use more interlanguage. • Standard features (form, size, revenue sources) remain influential on use of original discourses, yet not on interlanguage use.
  • 17.
    Observed Distribution ofInterlanguage Relative language use Entirety of entities involved in nonprofit evaluation
  • 18.
    Conclusion • Broad useof interlanguage across organizational forms • Interlanguage as a channel through which ideas get transported across social boundaries • Evidence of integration in combination and interlanguage use • Evidence of colonization in composition of interlanguage • Different discourses remain relevant: Rational – Normative Interesting times ahead!
  • 19.
    Future Plans • Relationalanalysis of the application of keywords – Nonprofit evaluation context as a densely interconnected reputation network among organizations and URLs – Identification of clusters
  • 20.
    Topic clusters Some topicsare clustered in specific regions of the reputation network.
  • 21.
    Positioning of clusters Othertopics are positioned on the top and in the center.
  • 22.
    Positioning of clusters …some connect center and periphery with an inverse distribution of size (marginalized term?)
  • 24.