Disaggregating the Effects of Deliberation on Policy Attitudes in the United States

Cynthia Farrar
James S. Fishkin
Christian List
Robert Luskin
Fieldwork by:
Connecticut, United States of America
133 participants from an interview sample of 1,032
2002 - 2002
Intervention type:
  • Information
AEA RCT registration number:

“Deliberative Polling” aims to gauge the opinions people would hold if they knew more about the issues. A typical deliberative poll engages a random sample of citizens in small group discussions and conversations with policy experts to create more informed and reflective public opinion. Researchers evaluated the impact of deliberation on voters’ attitudes about two policy issues: airport expansion and revenue sharing. The results indicate that formal on-site deliberations produced changes in attitudes and increased the likelihood that participants would have ordered, consistent preferences, particularly for the less salient policy issue.

Policy issue

“Deliberative Polling” aims to gauge the opinions people would hold if they knew more about the issues. In a typical deliberative poll, a random, representative sample is first polled on the targeted issues. After this baseline poll, members of the sample are invited to gather at a single location for a weekend in order to discuss the issues. Participants are given the opportunity to engage with competing experts and political leaders based on questions they develop in small group discussions with trained moderators. After the deliberations, the sample is again asked the original questions. The resulting changes in opinion are supposed to represent the conclusions the public would reach, if people had opportunity to become more informed and more engaged with the issues. Since 1994, dozens of Deliberative Polls have been run in the US and other countries. The evidence from these polls suggests that deliberation altered individual and group policy attitudes, and that people who engaged in deliberation were more likely to have clear and ordered policy preferences. This is the first study to rigorously evaluate the impact of Deliberative Polling using a randomized control trial and tried to breakdown what aspects of the treatment accounted for the change in policy attitudes and ranking of preferences.

Context of the evaluation

This study involved a Deliberative Poll in the area around the US city of New Haven, Connecticut. Participants deliberated about two local issues—the level of service at a local airport and whether there should be property tax revenue sharing for new commercial developments. These issues differed in prominence. Over the preceding year, the region’s most popular daily newspaper ran 74 articles on the airport, along with editorials and letters from citizens, but only mentioned revenue sharing seven times. The airport coverage also spanned the full range of commonly held views, meaning that participants would have thought, learned and talked far more about the airport than about revenue sharing prior to the deliberative sessions.

Details of the intervention

Researchers used a randomized control trial to rigorously measure the impact of Deliberative Polling on participants’ attitudes, and breakdown what aspects of the treatment accounted for the change. As with normal Deliberative Polling, a random sample was drawn, interviewed, and then invited to attend a weekend of deliberations in one location. Before attending, participants were given balanced briefing materials about two local issues—airport expansion and revenue sharing—and were randomly assigned to small discussion groups led by trained moderators. The formal on-site deliberations consisted of three deliberative sessions, each involving both small-group discussions and plenary questions-and-answer opportunities with panels of policy experts and advocates. The first two sessions, occupying the whole of Saturday, concentrated on one issue apiece. Groups were randomized as to whether they discussed the airport or revenue sharing in the first session. Participants were surveyed on three occasions—an initial telephone interview, after the first deliberative session, and after the conclusion of the weekend. Participant attitudes when interviewed after the first deliberative session gave researchers the impact of deliberation since at that point some had only discussed the airport and others had only discussed revenue sharing. After completing the final questionnaire, participants were paid US$200 each.

Results and policy lessons

The results show that on-site deliberation led to a net attitude change, and produced opinions that were more likely to be single-peaked . When researchers measured participant attitudes half way through the sessions, one group had only discussed the airport while the other had only discussed revenue sharing. Participants who had discussed revenue sharing dramatically shifted their opinions on the subject, while the group that had not yet discussed revenue sharing had no change in opinion on the issue. This shift suggests that Deliberative Polling created a net attitude change on revenue sharing, and that the deliberations themselves drove most of the change, rather than other aspects of the Deliberative Polling treatment such as the invitation, briefing materials, or casual conversations with other attendees. Researchers suggest that there was some effect from deliberating the airport expansion, but the effect was smaller. This was probably due to the fact that the airport was a far more prominent issue that participants had already discussed and heard information about. Other Deliberative Polls have shown similar results—that there are larger changes on less salient issues.

Farrar, Cynthia, James S. Fishkin, Donald P. Green, Christian List, Robert C. Luskin, and Elizabeth Levy Paluck. 2010. "Disaggregating Deliberation’s Effects: An Experiment within a Deliberative Poll." British Journal of Political Science 40(2): 333-347.