Survey Responses vs. Actual Roll Call Votes: A Surprising Disconnect in Representation Measurement

Comparing survey-based voter preferences to roll-call voting records often suggests poor political representation. However, this common approach may be flawed because survey responses differ significantly from actual votes.

Data & Methods: Three parallel survey experiments were conducted. Respondents received either partisan information or nonpartisan details about proposed legislation and Supreme Court decisions. The study draws on survey methodologies to demonstrate how knowledge affects expressed policy positions.

The findings reveal that simply providing accessible information reshapes public opinion in ways not captured by traditional representation studies:

Key Findings:

• Information Effects: Providing partisan or nonpartisan context significantly shifted reported preferences, making them less centrist and more aligned with the respondent's preferred party.
• Knowledge Awareness: Respondents demonstrated awareness of their information gaps when asked to evaluate roll-call policy votes.
• Confidence Impact: Access to information reduced confidence judgments regarding survey questions about legislative actions.

These results mirror patterns observed in Supreme Court voting studies, showing that the "information gap" between citizens and legislators significantly influences how we interpret reported preferences. This suggests political science research should reconsider its standard methods for evaluating descriptive representation.