Skip to main content
Log in

The potential of deliberative reasoning: patterns of attitude change and consistency in cross-cutting and like-minded deliberation

  • Original Article
  • Published:
Acta Politica Aims and scope Submit manuscript

Abstract

Previous studies have found that deliberative practices such as mini-publics produce opinion changes among participants. Nevertheless, the underlying mechanisms and whether these conform to deliberative ideals have received much less attention. This is problematic since research on public opinion and political psychology suggests that political opinions often are unstable or driven by prior notions. For this reason, we examine the underlying mechanisms of change in opinions and attitude consistency. We do so with data from an experiment with two deliberative treatments—cross-cutting and like-minded discussions—as well as a control group, where no deliberation took place to be able to determine whether deliberation actually cause the observed changes. The results suggest that participants in cross-cutting deliberation are more willing to change opinions, even when they have prior experiences with discussing the topic at hand, which is in line with deliberative theory, but attitude consistency is largely unaffected by the deliberations.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2

Similar content being viewed by others

Notes

  1. For access to data, see https://services.fsd.uta.fi/catalogue/FSD2958?lang=en&study_language=.

  2. An exploratory factor analysis of the 4681 respondents indicated that the fourteen items formed a single dimension measuring attitudes for or against immigration. Although the data material we use here excludes respondents with intermediate values, we corroborated this uni-dimensionality with both Mokkan scale analysis and confirmatory factor analysis at both T1 and T4. Although the Mokkan scale analysis suggested that some items could be left out, the correlations between the indexes with and without these items were extremely high at both T1 and T4 (about 0.99), meaning the inclusion is unlikely to affect the substantive results. We therefore proceed with considering migration attitudes as a one-dimensional phenomenon since the basic outline of the experiment is based on this premise.

  3. See Karjalainen and Rapeli (2015) for an analysis of the attrition in connection to this experiment.

  4. The hypothetical maximum value of this index is 14, but this is unlikely to be observed since it entails that a respondent on each question moved from one extreme to the other, which is highly unlikely.

  5. We refrain from using multilevel modeling since the control group is one large group including 369 of the 576 respondents, meaning it would not be possible to estimate the group effect adequately. This is also warranted empirically since empty models suggest that a small share of the variance is at the group level (ICC = 0.02). The robust standard errors correct biases from heteroscedasticity and therefore alleviates the problem.

  6. We fitted a number of models that includes age gender and education as control variables, but these show substantially similar results.

  7. A closer look at the mean scores shows that the like-minded treatment has a higher mean than the cross-cutting treatment and the control group. This is not due to a major flaw in the randomization process, but rather a consequence of more con-immigration respondents abstaining in the final stage of the experiment, which means that this treatment ended up with a couple of more pro like-minded groups than con like-minded groups. The cross-cutting treatment and the control group, where no such imbalances exist, have similar means.

  8. We also tried to include the interaction effects in separate models and mean-centering variables, but since the effects were substantially similar, we do not report these models in the tables.

References

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Staffan Himmelroos.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Himmelroos, S., Christensen, H.S. The potential of deliberative reasoning: patterns of attitude change and consistency in cross-cutting and like-minded deliberation. Acta Polit 55, 135–155 (2020). https://doi.org/10.1057/s41269-018-0103-3

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1057/s41269-018-0103-3

Keywords

Navigation