Testing the Transmission Belt

1
Testing the Transmission Belt

• Do Internal Associational Institutions Affect
  National Political Attitudes and Behaviors in the
  United States?
• Neil Carlson
• Duke University Dept. of Political Science
• nec_at_duke.edu

2
Overview

• Anecdata the Little League
• The logic of congruence
• Empirical literature brief
• 1996 NES data
• New data set on associational institutions (born
  on August 13!)
• Very preliminary results

3
Anecdata the Little League

• The Little League is profoundly democratic
• Of 1,534 respondents to the 1996 NES,
• 45 (2.9) mentioned involvement with any of the
  organizational arms of the major political
  parties
• 42 (2.7) mentioned the Little League
• The Leagues demonstration effect on democratic
  socialization could be significant

4
The Logic of Congruence

• Advocates of congruence think it imperative that
  the internal life and organization of
  associations mirror liberal democratic principles
  and practices (Rosenblum 1998, 36)
• Congruence is a diehard hypothesis
• Machiavelli (1517) on laws and customs
• Locke vs. Filmer (1690) on the analogy between
  family and state
• J. S. Mill (1848) on democratic institutions in
  detail
• Eckstein (1966) on Norwegian political culture

5
Nancy Rosenblums Critique

• 1998 book Membership and Morals
• Rebuts simplistic transmission belt model of
  democratic character, trust (vs. Putnam 1993)
• Belt claim disregards human cultural ability to
  discriminate among contexts
• Opposed to coercive legal measures forcing
  liberal democratic forms on pluralist civil
  society

6
Weak Version of Congruence

• Comprehensive democratization unnecessary and
  undesirable
• Almost any exposure to democratic practice may be
  enough to socialize citizens (Walzer 1991)
• Same cognitive ability that can distinguish
  contexts also allows abstraction between contexts
• Hypothesize higher diffuse support (Easton
  1965) due to reduction of uncertainty about
  utility of democratic procedures.

7
Empirical Literature

• Few large-N studies of authority structure in
  specific organizations Skocpol 1999, Stolle 1998
• Most recent studies use only associational
  categories (labor, veterans, etc.)
• Focus is on generalized trust, direct political
  mobilization little on diffuse support effects
  of civic engagement.

8
1996 National Election Study

• Uniquely detailed dataset
• Open-ended mentions of up to 4 associational
  names in 22 categories 4,700 total mentions
• Increased estimates of associational involvement
  over GSS (84.9 vs. 71.5)
• Access to mentions restricted, respondent IDs
  scrambled to protect confidentiality

9
Associational Institutions Leadership
Questionnaire

• Cleaned data 4,678 mentions
• Identified 161 organizations or movements
  mentioned by at least 2 respondents accounted
  for 2,030 (43.3) of all mentions
• Contacted offices of 125 groups with
  Internet-based leadership questionnaire
• Received 81 responses from 67 organizations,
  accounting for 1,455 (31) of all mentions

10
Relational Data Map
11
Institutional Data

• Attributes of membership system and volunteer
  responsibilities
• Budget sources
• Branch/chapter autonomy
• Leadership selection and terms
• Electoral competition and frequency and size of
  electorate
• Policymaking meetings, parliamentary procedure
  and agenda openness.

12
Measurement Model

• Simplifies evaluations
• Confirmatory factor analysis on data from 67
  associations see Appendix B of paper.
• Three latent constructs (alphas)
• Organizational democracy (0.78)
• Membership constitutionality (0.74)
• Organizational success (0.79)
• Result plenty of variation

13
No Sign of the Iron Cage
14
Success is Little Correlated
15
Older Associations Are A Little More Democratic
16
RespondentDistributions

• Catholic Church largest organization, below
  average
• Seniors group marks low end
• Veterans groups highest scoring, most likely to
  pull up maximum over average score
• Baptists and Methodists among largest democratic
  groups

17
Widely Null Model Results

• Organizational democracy scores appear not to
  have simple influence in a wide array of models
• Dependent variables included
• people can be trusted, latent trust
• politics too complicated, latent internal
  efficacy
• trust government, latent external efficacy
• satisfied with democracy, latent regime support
• turnout and latent campaign engagement

18
Conclusions

• Associations do not appear to be too dependent on
  the mail-order model (vs. Putnam, Skocpol?)
• Little or no evidence for congruence here strong
  congruence is fairly clearly deprecated
• Involvement (total mentions) remains
  substantively important in most models
• Feedback ideas on what to do with this data?