WHEN DELIBERATION PRODUCES EXTREMISM

Abstract

What are the effects of deliberation about political issues by likeminded people? An experimental investigation involving two deliberative exercises, one among self-identified liberals and another among self-identified conservatives, showed that participants' views became more extreme after deliberation. Deliberation also increased consensus and significantly reduced diversity of opinion within the two groups. Even anonymous statements of personal opinion became more extreme and homogeneous after deliberation.

Notes

1. Consistent with the general demographics of the two areas, 90 percent of respondents were Caucasian. In both cities three of the five groups contained one non-Caucasian. There were no significant differences between groups with and without a non-Caucasian in any group or individual responses related to the affirmative-action question. There was no significant difference in age between the samples (the median was 46 years in both counties). Age did not have a significant effect on the willingness to change one's opinion; the correlation between age and the extent to which a person changed their opinion in the direction of the group was .12, which is not statistically significant in this sample.

2. Screening questions included the following. (a) “In general, would you describe your political views as very conservative, conservative, moderate, liberal, or very liberal?” (b) “Suppose you were in the voting booth and you came across an office for which two candidates. .. were running and you had never heard of either one. Which candidate would you choose— the Democrat or the Republican—or would you just not vote for that office?” Participants were also asked to assign grades to various public figures, predicting how they would perform as president of the United States. The conservative names included Dick Cheney, Wayne Allard (U.S. senator from Colorado), Rush Limbaugh, and Pat Robertson. The liberal names included Edward Kennedy, Hillary Rodham Clinton, Jesse Jackson, and John Kerry.

3. In the group sessions the designated monitor was given five numbered envelopes to be opened in order as the previous envelope's task was completed. The first three envelopes contained instructions for the group to discuss and reach a consensus (if possible) on one of the three focal issues. A fourth contained individual forms identical to those the participants had completed before the groups were convened, which asked for their private opinions about all three topics after the group discussions were completed. The other envelope asked the group to discuss an unrelated issue.

4. A repeated-measures anova (Analysis of Variance) showed that there were highly significant differences between the two samples in their pre-deliberation opinions about the issues to be discussed: F(1,61)=234.3, (p<.001). This difference was separately significant for each of the three issues (each issue p<.001).

5. F(1,61)=56.1, p<.001.

6. Global warming, p<.001; affirmative action, p<.001; civil unions, p<.02.

7. p<.01.

8. p<.001.

9. z=4.7, p<.001, by a sign test.

10. A good outline can be found in Ross and Nisbett 1991.

11. See Kuran 1997, discussing general phenomenon of “preference falsification,” in which people's public statements are systematically inconsistent with their actual private views.

12. On the complex relations between public statements and private views, see Kuran 1997.

13. See Schkade et al. 2000, which found that the pre-deliberation median predicts opinion movement even when there is considerable group diversity.

14. Cohen also notes that the policy favored by the relevant party affected participants' views even in the absence of supporting arguments. See also http://americans-world.org/digest/global_issues/global_warming/gw2.cfm. In 2005, about 70 percent of Americans favored the Kyoto Protocol, designed to control global warming—but when people were informed that President Bush rejects the Kyoto Protocol, the percentage in favor dropped to 43 percent.

15. http://cdd.stanford.edu/polls/docs/summary; see also Fishkin 1995, 22 (showing an increase, on a scale of 1 to 3, from 1.40 to 1.59 in commitment to spending on foreign aid; also showing a decrease, on a scale of 1 to 3, from 2.38 to 2.27 in commitment to spending on Social Security).

16. http://cdd.stanford.edu/polls/docs/summary/

17. See Sunstein 2007 for discussion. Of course the Internet also makes it very easy to encounter new and different positions, and to the extent that people use the Internet to find such positions, ideological amplification will be less likely, not more so.

Additional information

Notes on contributors

David Schkade

is the author, interalia, of Are Judges Political? (Brookings, 2006) and Punitive Damages: How Juries Decide (Chicago, 2002)

Cass R. Sunstein

is the author, inter alia, of Going to Extremes: How Like Minds Unite and Divide (Oxford, 2009)

Reid Hastie

Robert S. Hamada Professor of Behavioral Science, Booth School of Business, is the author, inter alia, of Rational Choice in an Uncertain World (Sage, 2009) and Inside the Jury (Harvard, 1983). This article is adapted from and reproduces portions of Schkade, Hastie, and Sunstein 2007, ‘‘What Happened on Deliberation Day?’’ California Law Review, 95: 915–40, which was completed prior to Sunstein's appointment as Administrator of the Office of Information and Regulatory Affairs. Nothing said here represents an official government position in any way