Measuring Belief Systems as Directed Networks of Attitudes

• In Event: Innovations in the Study of Constraint in Mass Belief Systems

Abstract

Ideological belief systems theoretically represent complex, directed relationships between a wide range of ideas. However, surveys that are used to test the extent to which citizens demonstrate ideological “constraint” typically ask about these ideas independently. Which is to say, taking individuals’ belief systems as networks of political attitudes, standard public opinion research allows us to observe nodes but not edges. This approach can make individuals appear incoherent or “unconstrained” when their policy preferences do not reflect party platforms.

We address this limitation with a novel survey instrument that asks respondents to explicitly articulate causal relationships between their policy positions and broader normative goals and values. We first ask respondents to report positions on a range of public policies and identify broad goals they think are worth achieving. For each position they take, we then ask them to identify whether they took that position because of each of their broader goals. We also ask respondents to report ideological identities, and follow up with an open-ended item asking respondents to elaborate on why they chose to identify that way.

Applying novel network analytic techniques, we then identify areas where members of different ideological groups see different policy positions as instrumental for achieving similar goals, and whether citizens who exhibit stronger preference constraint also articulate denser networks of political attitudes.

Authors

• Jon Green, Northeastern University

• Sarah Shugars, Rutgers University

• Nic Fishman

• Peter Levine, Tufts University