Citizens in modern society are faced with complex and controversial public issues. For example, Taiwanese referenda in 2018 and 2021 involved 10 and 4 proposals, respectively. These proposals covered a range of controversial topics such as whether to resume the construction of a nuclear power plant and whether to legalize same-sex marriage. Discussions of these multifaceted topics often caused heated debates and resulted in attitude polarization. People often became more opinionated and believed more firmly in their initial stances on the issues. As attitude polarization is a major cause of interpersonal and intergroup conflicts and even social unrest, this research investigated how to mitigate the polarization of attitudes toward social issues.

This study drew upon the theory of perspective taking for insights into the mitigation of attitude polarization. Perspective taking refers to one’s attempt or ability to consider something from someone else’s perspective. Perspective taking has been studied by psychologists as a means of reducing prejudice, fostering cooperation, and lowering reasoning biases that were found to be responsible for attitude polarization.

Attitude polarization is, at least partially, a result of biased information processing. For example, people may become more entrenched in their opinions after engaging in biased assimilation, which entails, for instance, evaluating information supporting their opinions as more convincing than that opposing their opinions. Per the dual process theory, the reasoning biases that cause attitude polarization often reflect the influence of Type 1 processing, which usually involves responding to information intuitively and generating outcomes consistent with existing beliefs. Such outcomes can be either reinforced or overridden by Type 2 processing, which involves deliberation and evidence-based consideration. If an intuitive response is reinforced, an individual’s attitude is likely to become entrenched, leading to polarization; if overridden, attitude polarization may be inhibited or attenuated. We proposed that perspective taking may serve as an overriding mechanism to mitigate attitude polarization.

Perspective taking enables individuals to undertake a form of Type 2 processing called decontextualized reasoning, which features a type of self-distancing that allows individuals to decouple existing beliefs from reasoning processes. Such a reasoning style can arguably be triggered by perspective taking because this requires individuals to examine information by considering their own thinking with increased psychological distance and adopting alternative perspectives. Although research has suggested a link between perspective taking and decontextualized reasoning, how decontextualized reasoning can be measured and whether it indeed serves as a mediator between perspective taking and attitude polarization have yet to be deeply investigated.

Integrating studies on perspective taking, information processing, and attitude change, this study proposed three hypotheses and tested them by conducting an experiment involving a writing task focused on social issues. First, we hypothesized that perspective taking would mitigate attitude polarization. We measured attitude polarization by calculating the difference in attitude toward an issue before and after the writing task. Second, we proposed that the mitigation of attitude polarization would be mediated by decontextualized reasoning induced by perspective taking. To investigate this hypothesis, we used a text analysis tool, namely Linguistic Inquiry and Word Count (LIWC). Research on LIWC has shown that the frequency with which people use words in their writing reflects psychological features and cognitive styles. Because decontextualized reasoning features a shift of attention from the self to others, we argued that it can be detected in the difference between the frequency of using thirdperson plural and first-person singular pronouns. The greater this difference is, the greater the extent to which decontextualized reasoning is likely to have been adopted by a writer. Finally, we drew upon the secondary transfer effect of intergroup contact theory and proposed that the mitigating effect outlined in the first hypothesis would influence people’s attitudes even when not directly instructed to engage in perspective taking. To test this transfer effect, the writing task included two sessions, and perspective taking instructions were only given in one of the sessions.

The experimental procedure was as follows. We recruited 95 participants (50 female; mean age = 23.7 years, SD = 3.4) for the writing task experiment. Participants first indicated their attitude toward two issues that are controversial in Taiwan (making Taiwan nuclear-free and abolishing the death penalty) on a 7-point Likert scale with responses ranging from absolutely disagree to absolutely agree and then wrote arguments supporting or opposing the issues in two writing sessions. When writing about one of the issues in the first session, all participants wrote arguments from their own perspective; when writing about the other issue in the second session, half of the participants continued to write from their own perspective (control condition) while the other half wrote from the perspective of someone holding the view opposite of their own (perspective-taking condition). Participants then again indicated their attitude toward both issues using the same 7-point scale. The whole experiment took approximately 30 minutes to complete.

Analysis of attitude polarization between the two conditions showed that compared with the control group (M = 0.77, SD = 1.09), for the essays written in the second session, participants in the perspective-taking group (M = −0.02, SD = 1.74) displayed weaker attitude polarization, t(93) = 2.63, p = .010, d = 0.540.

Based on LIWC output for the second essay for the decontextualized reasoning indicator (frequency of use of third-person plural pronouns minus that of first-person singular pronouns), compared with the control group (M = −1.04, SD = 1.20), the perspective-taking group (M = −0.08, SD = 0.95) engaged in decontextualized reasoning to a greater extent, t(93) = 4.34, p < .001, d = 0.890. Mediation analysis using the SPSS Process Macro further revealed that this difference mediated the mitigating effect found during analysis, and thus, participants in the perspective-taking group engaged in more decontextualized reasoning and exhibited weaker attitude polarization than those in the control group: indirect effect = −0.28, 95% confidence interval: [−0.54, −0.07]; direct effect = −0.50, p = .122.

Because all participants wrote from their own perspective in the first session, the difference in attitude polarization resulting from the first essay allowed us to examine the transfer effect. Compared with participants in the control group (M = 0.79, SD = 1.56), those in the perspective-taking group (M = 0.10, SD = 1.37) again displayed weaker attitude polarization, t(93) = 2.27, p = .026, d = 0.465.

Taken together, the results revealed that perspective taking mitigated polarization of attitudes toward social issues due to greater engagement in decontextualized reasoning, a form of Type 2 processing that involves self-distancing. This mitigating effect was evident for issues where the perspective-taking manipulation was not applied, thus indicating a transfer effect. These findings can have valuable implications for areas where controversies and conflicts abound. For example, during the COVID-19 pandemic, disputes occurred over issues such as the necessity of vaccines, legitimacy of lockdown measures, and the distribution of supplies. These issues fall into the category of socioscientific issues (SSIs), which have been defined as crucial real-world scientific issues with social significance. Other examples of SSIs include issues pertaining to climate change, genetically modified products, and genetic testing. As were the issues used in our writing tasks, SSIs are often controversial in nature, and information relating to them is often processed through an ideological lens, which can lead to attitude polarization and conflict. Helping students develop the ability to approach SSIs with scientific and fact-based reasoning has become a key objective of science and citizenship education. The manipulation used in our study can be integrated into course materials to help students analyze SSIs from multiple perspectives and avoid becoming entrenched in their opinions. Furthermore, such a practice may even affect perspectives on issues not discussed in class due to the transfer effect of perspective taking. Whether the mitigating effect of perspective taking on attitude polarization can be replicated in the classroom or other real-world settings warrants further exploration. Future research directions and other implications are also discussed.