Seeing the Other Side? Perspective-Taking and Reflective Political Judgements in Interpersonal Deliberation

Existing research suggests that several mutually nonexclusive mechanisms may drive reflection in citizens’ political thinking in deliberative settings. The justification rationality theory embraces the Habermasian logic of communication, and argues that any positive transformation in citizens’ judgements within public deliberative settings is the product of a ‘systematic process wherein actors tell the truth, justify their positions extensively and are willing to yield to the better argument’ (Bächtiger et al., 2010: 33, emphasis author’s). The logic underlying this pathway is that rational arguments exchanged during deliberation make citizens ‘become aware of the conflicts inherent in their own desires’ and offer them ‘new perspectives not only with respect to possible solutions, but also with respect to their own preferences’ (Manin, 1987: 350). As a result, citizens are induced to reflect on their judgements ‘by the force of the better argument’ (Habermas, 1981). Prior research offers some evidence to support this claim. For instance, Gerber et al. (2014) examine mechanisms underlying opinion transformations in the context of Europolis deliberations and find that reason-giving during deliberation affects opinion transformation among citizens (albeit this influence is context-specific). Yet, Caluwaerts and Reuchamps (2014) find that the quality of deliberation (e.g. the level of justification, respect for counter-arguments) has a very limited impact upon citizens’ political judgements (see also Himmelroos and Christensen, 2014).

The information theory, however, posits that individuals gain issue-related information and knowledge either from experts or from each other (or both) in deliberation, which leads them to reconsider their policy choices. Jason Barabas (2004) finds that information gained during deliberation leads to the formation of enlightened judgements. Still, some studies show that the effect may not be exclusive to group deliberations (e.g. Goodin and Niemeyer, 2003; Muhlberger, 2005).

Both theories capture the crucial aspects of deliberation, most importantly reason-giving and learning. However, by focusing exclusively on reasons and factual information gain as potential mechanisms, these theories either implicitly or explicitly discount the role of emotions in citizens’ reasoning processes. The research across disciplines shows that reflection is a complex process, and is not devoid of emotions. In fact, emotions precede and influence human reasoning and decision-making and interact to produce more thoughtful political judgements (Marcus et al., 2000). Therefore, the main objective of this article is to advance a theory on the role of perspective-taking – a dimension of empathy – in prompting more reflection in citizens’ political thinking in deliberative settings.

The study of perspective-taking and affective empathy is not new to deliberative democracy and deliberation research. For example, Tali Mendelberg (2002: 153) argues that ‘deliberation is expected to lead to empathy with the other and a broadened sense of people’s own interests’. Jane Mansbridge (1983: 285) posits that face-to-face discussions ‘provide the conditions in which citizens are most likely to take responsibility for and empathize with others’. Jurgen Habermas (1981), in his interpretation of discourse ethics, advocates the process of ideal role taking (i.e. perspective-taking) as being an important precondition of a good deliberation, and suggests that only by taking the perspective of others and projecting ourselves into their situation can we understand their arguments.

It is argued that perspective-taking and affective empathy lead, in turn, to a set of beneficial democratic outcomes. Jane Mansbridge suggests that empathy leads people to make each other’s good their own (Mansbridge, 1983: 27) by ‘facilitate(ing) the creation of common interest’ (Mansbridge, 1983: 285) and nurturing listening and mutual respect (Mansbridge, 1999: 225). In a similar vein, Michael Morrell (2010), in his comprehensive account of the role of empathy in deliberative democracy, Empathy and Democracy, contends that empathy is a crucial process in which we have to engage if we want deliberative democracy to succeed. He calls for reshaping deliberative theory in order to place the process of empathy at the centre of deliberation, and advocates the inclusion of empathy in the democratic education system (Morrell, 2007, 2010).

Despite this attention to perspective-taking and affective empathy, scholars have rarely theorized about why, how and under what conditions deliberation can induce the process of perspective-taking among citizens and, similarly, whether and under what conditions perspective-taking is beneficial for citizens’ reasoning processes during deliberations.

Here, I offer a theory of perspective-taking in deliberation focused on two questions. First, what institutional features of deliberative forums make them perfect for inducing the process of perspective-taking among citizen deliberators? Second, how is this process, in turn, capable of producing more reflection in citizens’ political reasoning?

Reflection involves engaging in the consideration and evaluation of diverse and conflicting perspectives before arriving at political decisions (Dewey, 1933). Research in social psychology tells us that perspective-taking – actively imagining others’ experiences, perspectives and feelings – attenuates egocentrism in human judgements, and leads to more reflective and considerate thinking in individuals (see Todd and Galinsky, 2014 for a review). Scholars suggest that there are several mechanisms underlying this relationship. First, putting oneself in someone else’s shoes may evoke feelings of increased empathic concern towards the target (whose perspective one is taking), which may, in turn, prompt the individual to overcome egocentric reasoning and reflect on his or her choices more deeply. Second, taking someone else’s perspective may lead to a ‘greater overlap in mental representations of self and other’ (Todd and Galinsky, 2014: 380). In other words, perspective-taking can make the other person appear ‘self-like’ or make the self seem more ‘other-like’ (see Todd and Galinsky, 2014 for a review). Whichever is the underlying mechanism, perspective-taking may create an increased awareness and understanding in citizen deliberators of others’ lives and perspectives. In this dawning of awareness, people may be more motivated to engage in reflective reasoning (e.g. Galinsky and Moskowitz, 2000; Todd et al., 2012).

Perspective-taking is itself a challenging process. There are several important barriers that hinder people from taking others’ perspectives accurately (see Epley and Caruso, 2009). This process requires at least two conditions. First, for a person to actively imagine the world from another person’s perspective, he or she needs to have information about that world. In the absence of information, imagining the scenario in question becomes difficult or something inaccurate is envisaged. When trying to take the perspective of someone else, people may, by default, rely on their ‘stored knowledge’, which may include ‘stereotypes or other idiosyncratic information known about the target’ (Epley and Caruso, 2009: 300–305). Potential biases in the stored knowledge may lead people to make erroneous inferences about another person’s feelings and thoughts. This could, in turn, become a potential barrier to the process of perspective-taking. In other words, for perspective-taking to work, a person needs to acquire relevant and useful information about the target and his or her world.

Second, the information about the target’s feelings and thoughts should be communicated to the perspective-taker in a way that facilitates the process of perspective-taking. In rare situations, perspective-taking is automatic. In most cases, it needs to be explicitly activated (Epley and Caruso, 2009). Communication scholars have proposed storytelling as an important communicative tool capable of inviting and encouraging the process of perspective-taking (see Gastil and Black, 2018 for a review). In Laura Black’s (2008: 99) words, ‘stories bring people’s experiences and perspectives to the conversation in a powerful way that is qualitatively different from issue-oriented discussion’, due to their ability to ‘display values and worldviews that are typically not talked about’ via other discursive modes, such as fact-based argumentation (Black, 2008: 105). In other words, personal stories should have the potential for prompting participants to ‘imagine the real[ity] of the other’ (Black, 2008: 96; Buber, 1998: 71; Gastil and Black, 2018).

The institutional features of deliberative forums make them well suited for meeting these required conditions. With respect to information, deliberative forums include learning and deliberation processes that encourage the participants to listen, learn, ask clarifying questions and get acquainted with diverse perspectives, viewpoints and lives. One of the key features of many deliberative forums is that they are usually designed to enable the participation of people with diverse and opposing points of view. After having spent some time together, participants acquire information not only about each other’s views on the issues under discussion, but also about each other’s life experiences, families and other background characteristics. The presence of diverse perspectives should, thus, create a rich informational context that should further facilitate the process of perspective-taking among participants (see Grönlund et al., 2017: 464 for a similar argument).³

Second, nowadays, deliberative forums foster not only the fact-based argumentation, but also alternative and less formal modes of communication such as storytelling, narratives, testimony and humour. These additions to the ideal type of deliberative discourse modes have mainly been inspired by the arguments put forward by feminist democrats that the neutral language of rational reason-giving is typical of middle-class white men and largely disadvantages women, minorities and the working class (e.g. Sanders, 1997; Young, 2000).

When combined, these two features of deliberation – the presence of diverse perspectives and the opportunity for expressing one’s arguments and perspectives through personal stories – should induce the process of perspective-taking in citizen deliberators. Therefore, deliberative forums should provide fertile environments for inducing in people the process of perspective-taking and, subsequently, that of reflective judgement.

Although my argument is primarily theoretical, in this section I rely on a real-world deliberative forum – the ICA – to illustrate its plausibility. The ICA was established in 2016 with the goal of carefully considering five important legal and policy issues pertaining to Irish society: a constitutional ban on abortion, an ageing population, climate change, referenda and fixed-term parliaments. It consisted of the chairperson and 99 randomly selected citizens (and 99 substitutes) from all walks of life and largely representative of the Irish voters, according to main sociodemographic features such as age, gender, social class and regional spread. During 11 weekends, assembly members met to deliberate in small groups, listened to expert presentations from both sides and produced conclusions on each of the topics discussed. These conclusions comprised the bases of reports and voted recommendations, and were later submitted to the Houses of Oireachtas (Irish parliament) for subsequent debate by the elected members. The assembly made 1 key recommendation on abortion, 15 recommendations (and 6 ancillary ones) on ageing population, 13 recommendations on climate change, 8 recommendations on referenda and 7 recommendations on fixed-term parliaments (see Farrell et al., 2019).

The selection of this particular case study is not coincidental. The ICA is a real-world and important deliberative forum, which has been influential in informing and shaping the political decisions of the broader Irish polity. Its recommendations on the topic of the Eighth Amendment to the Irish Constitution (i.e. abortion) pushed the Irish government to call for a nationwide referendum in May 2018. The assembly had a significant role in this process. First, it ‘shap[ed] the referendum question and the draft legislation’ (Field, 2018: 608). Second, public deliberations by the assembly were livestreamed and its recommendations were publicized among the wider citizenry. The commentators suggest that the assembly has contributed substantially to progressing the debate on abortion within Irish society. Early evidence from the RTÉ/Universities exit poll suggests that the assembly has influenced voting choices by improving knowledge about issues among a wider public. In other words, those participants who were more familiar with the assembly (compared to those who were unfamiliar) voted differently (Suiter, 2018). Furthermore, the debates of the ICA were also driven by broader divisive discourses outside the mini-public, especially when it came to the issues of abortion and climate change. Hence, the consequential character of this assembly for real-world policymaking, and its agenda-setting and awareness-raising roles in wider society (Suiter, 2018), makes this case study compelling with its strong ecological validity.

The ICA possessed the institutional features that should have induced the process of perspective-taking among participants according to the theory I have developed here. The assembly discussed topics that affected different parts of Irish society. For example, although abortion was clearly a feminine topic, ageing population was most relevant for the elderly and climate change mostly affected the future generation and so on. When the assembly was created, one of its main objectives was to bring together a group of citizens that closely resembled a microcosm of Irish society, not only with regard to the main sociodemographic factors, such as age, education, gender and region, but also in relation to views and perspectives. The presence of a diversity of viewpoints, or in the words of the organizers ‘allow[ing] the full spectrum of views to be heard on every issue’ (Citizens’ Assembly, 2018) constituted one of the key principles of the assembly. Ensuring a diversity of viewpoints during each round table discussion was also one of the main objectives of the organizers.

Second, the ICA created and nurtured a deliberative environment in which fact-based rational argumentation, storytelling, narratives, testimony and humour were freely expressed, whether according to expert or witness communications, round table discussions undertaken by assembly members or within the framework of communications by the wider public to the assembly. The structure of the assembly enabled the citizen deliberators to become acquainted with different perspectives through different modes of communication and, thus, facilitated the process of perspective-taking.

Thus, the ICA meets my theoretical conditions for facilitating the process of perspective-taking among participants. But did citizens feel that was the case too? To examine this question, I conducted in-depth and semi-structured interviews with 11 members of the Irish Assembly (4 female and 7 male). Data saturation was used as a criterion for discontinuing interview data collection (Saunders et al., 2018). The objective of these interviews was not to reach a quantifiable sample, but rather to obtain a deeper and more nuanced understanding about the processes underlying deliberations. The interviews were face to face and lasted on average an hour. They were audio-recorded after having obtained the informed consent of the assembly members and were later transcribed verbatim.⁴

Diversity and Inclusion: A pattern that has emerged repeatedly from interview data is that the presence of diverse voices was crucial for the processes of empathic imaginations (Goodin, 2000); it enabled participants to imagine and understand ‘where other people came from’. Consider, for example, the thoughts of an interviewee referring to the issue of ageing population in Ireland:

[A] lot of the kind of personal stuff were actually self-generated within the people who had already experienced that, you know, dealing with elderly relatives. And also . . . Let us not to forget because of the population spread within the group, there was, you know, the number of elderly, yeah, so they . . . had opinion . . ., because a lot directly affects them. . . . So for people who are little bit younger to have one to one conversation, that or round table conversation that was important.

In addition to the mere presence of these opinions, the ICA also provided the members with a safe environment in which to air these differences. This environment invited the free expression of stories and arguments and it contributed to creating bonds of trust and mutual understanding among citizens even when they disagreed with each other:

You know, everybody feels like it is a safe space for their opinion, even if they know that it [their expressed opinion] is not going to be popular with the people at the table in that forum [and] people will openly disagree with them; that it stays there and it is not being taken out to open air, where it is kind of . . . I do not like you, you know, because we do not agree and something like that.

Being encouraged to listen to and to speak with different others induced the processes of perspective-taking and affective empathy in participants:

It does not matter if you think it’s black and it is white. You know, you are listened to and spoken to. It is more important than you think it is black. Because at the end of the day, I know it is white and you think it is black. But the inclusion of it, being there. It is the empathy. I really listen to you; it is trash, but I listen to you. Yeah, you know what I mean?

The interplay of fact-based argumentation and personal stories: The interviewees repeatedly referred to the sequence of the ICA meetings as ‘ideal’, ‘brilliant’ and ‘perfect’ when talking about the ways in which deliberation affected their thought processes:

I think . . . it was planned brilliantly: the first few meetings were specifically about facts, and statistics, and peer-reviewed research, so that we could get understanding without any emotional discussions. . . . She [the chair] wanted to make sure that we understood the facts surrounding the 8th amendment, around the constitution, around what currently happens. . . . Further along in the process, because some of it was kind of abstract, then became more personal, when we heard personal stories of some of the women involved. We heard the recordings of them, and we heard from advocacy groups . . . which brought more emotive part from late date, once we had a grounding in proper factual information. . . . Once they understood the facts, then they heard the personal stories, and you know, they could make a decision based on that, I think that changed a lot of people.

Factual information served as a basis for citizens’ understanding of the issue, whereas personal stories and testimonies engaged their feelings with respect to the issue and made it easier for them to take the perspectives of diverse others. For some, these stories constituted a part of the factual information, especially when the story was about someone whose perspective was absent at the discussion table:

Personal stories . . . they were backed up by the factual information, by the statistics, but I think it is the part that gets you in the guts. . . . It is somebody’s story. We heard one story about an older lady who had been put in a care home, against her will, basically, and you know that was just incredibly sad, no amount of statistics could give you what actually happens underground, what actually happens in people’s lives. Because you cannot boil down the people to numbers generally speaking. But it is good to have numbers and statistics to back up the thoughts. So I think that the more impactful of the two on me anyway, was kind of personal stories and narratives from people directly. You . . . have to consider especially when somebody [is] not in the room, the personal story might be the factual personal story. When the two go hand in hand . . .

The stories were not only a part of the testimonies of the people affected by the policy issue, but also an important part of the group discussions:

[Y]ou know . . . other people’s experiences, sometimes when we sit in these kinds of sessions, we do not, maybe, fully realize that the individual experiences at the end of the day are actually big experiences, you know, you are talking about things that affect people’s lives, and quite often can change their lives, so, it is kind of, as I said, the combination of all of those processes.

In sum, these findings suggest that, consistent with the theory, citizens also found these two features of the ICA important for inviting them to take the perspectives of diverse others and imagine the world from others’ vantage points.

Interview data showed that the ICA had relevant institutional features, which according to the members, facilitated the process of perspective-taking. The next thing to consider is if this process had any effect on citizens’ engagement in reflective political reasoning during deliberations. Empirical studies examining the benefits of perspective-taking and affective empathy in deliberation are few and have mostly yielded mixed and inconclusive results. Michael Morrell (2010), for instance, studied the effect of empathy on individuals’ reciprocity and commitment to continue deliberating by designing and fielding a laboratory experiment. He found that group levels of empathic predisposition (manipulated based on previously measured individual levels of empathic predisposition) led to citizens developing heightened perceptions of the value of mutual respect and an open-minded deliberative process. Ugarriza and Nussio (2017) induced perspective-taking in citizen deliberators by asking them ‘to make references to their own personal perspective and history when justifying their proposals’ (Ugarriza and Nussio, 2017: 9), and measured their intergroup attitudes pre- and post-intervention. Their findings show that perspective-taking improved mean attitudes towards the outgroup, although this effect was limited to only one group – community members – and was not found among ex-combatants. Although these studies find some support for the claim that empathy is good for deliberation, they do not examine the role of perspective-taking in citizens’ political reasoning.

The exception is the Grönlund et al. (2017) study, in which the authors examine the question of whether the ability to consider others’ perspectives is a precondition for citizen deliberation and the extent to which deliberation enhances this consideration. Their analysis relies on an experiment in citizen deliberation on the polarizing issue of immigration, held in Finland. Their main outcome variable – consideration for an outgroup perspective – while not the same measure, seems to capture processes similar to those underlying reflection.⁵ Grönlund et al. (2017) find that deliberation increases consideration for the other side and this increase is greater among people in the con enclave (those with negative attitudes on immigration) deliberating in mixed groups. I build on this study and expand on it by studying the relationship between deliberation, perspective-taking and reflection in a real-world deliberative setting.

It is challenging (if not impossible) to measure and capture the intrapsychic processes of perspective-taking directly with interview or survey data. However, it is possible to test the plausibility of perspective-taking indirectly. This section is an attempt to gauge whether perspective-taking is positively related to citizens’ willingness to engage in a demanding task of reflective reasoning in interpersonal deliberative settings. Perspective-taking largely depends on ‘dispositions that people bring to deliberation’ (Morrell, 2018: 246). Psychologists argue that there is a wide variance in individuals’ dispositions to engage in perspective-taking (Davis, 1980; Van der Graaff et al., 2018). Some people are more inclined than others ‘to adopt the perspective, or point of view, of other people’ in everyday social interactions (Davis, 1980: 6). Thus, people who are most disposed to taking others’ perspectives may be more affected by the process of perspective-taking induced by deliberation. Conversely, individuals who have lower levels of dispositional perspective-taking may be less affected by it. Under this assumption, we can hypothesize that those who score high on dispositional perspective-taking are also more motivated to reason reflectively. Therefore, my hypothesis is that people’s dispositional perspective-taking will increase their willingness to engage in reflective political judgements.

The survey measures used for testing this expectation were part of a larger survey, fielded during the last weekend of the ICA in April 2018. I measure perspective-taking with two items from the Interpersonal Reactivity Index (IRI) questionnaire (Davis, 1980), a widely used and validated measure for dispositional perspective-taking. Participants were asked to indicate how well the following statements described them: ‘I sometimes find it difficult to see things from the other person’s point of view’ (reversed) and ‘When I am upset at someone, I usually try to put myself in his or her shoes for a while’. Each statement was rated on a 5-point scale with the greater number indicating higher levels of dispositional perspective-taking (M₁ = 4.24, SD₁ = 0.93; M₂ = 3.9, SD₂ = 0.99). Before introducing these items to assembly members, I tested them among 36 Irish citizens through an online opt-in panel designed for research, Prolific. Although the internal reliability of the measure in the pilot sample was acceptable (Cronbach’s alpha = 0.70), it was low in the ICA data. Therefore, I run my analyses on each item separately.

The proxy I use for the reflective judgement measure is adopted from a psychometric measurement called deliberation within by Weinmann (2017). The measurement builds on theoretical works by Mercier and Landemore (2012) and Gastil (2008), and taps into information processing ‘which comprises specific cognitive processes of individuals as citizens of a democratic society’ (Weinmann, 2017: 3). The original measure has five items; I use three items due to space restrictions within the survey. The validity and reliability of the measure were tested with a pilot study prior to fielding it. The items loaded well in one factor and the internal reliability of the measure was high (Cronbach’s alpha = 0.80). Respondents were asked to indicate how well each of the items below described the development of their views over the course of the ICA meetings: (a) I tended to reflect on the different views about the issues discussed, (b) I thought about arguments for and against my own and others’ opinions on the issues discussed and (c) I evaluated the arguments that speak for and against my own and others’ opinions. An index was created summing the responses to three items (α = 0.74). I rescaled the response scale from 9–15 to 1–7 for ease of interpretation (M = 5.18; SD = 0.19).

Figures 1 and 2 present the range and spread of the dataset for each variable of interest. Both independent and dependent variables seem to skew towards the high end: on average, citizen deliberators score high on the self-reported measures of dispositional perspective-taking (M_pt1 = 4.25, SD_pt1 = 0.93, N_pt1 = 69; M_pt2 = 3.9, SD_pt2 = 0.99, N_pt2 = 69; range: 1–5) and reflection (M_rj = 5.2, SD_rj = 1.60, N_rj = 69; range: 1–7).

Figure 3 presents two scatterplots visualizing the relationship between perspective-taking (in two separate items) and reflective judgements. It shows there are three observations (two in Figure 3(a) and one in Figure 3(b)) that are clear outliers. A closer look at these observations reveals that they share at least three common characteristics: all three are male, aged 65+ and retired. Dropping these outliers can substantially bias my results. Therefore, I run all my analyses twice, with and without these outliers.

For my analyses, I use a nonparametric test as my data have a very skewed distribution and have outliers that cannot be removed. I examine the relationship between perspective-taking and reflective judgements with the help of Spearman’s rho, which assesses how well the association between two variables can be described. My preference for the Spearman rank correlation is based on the fact that its correlation coefficient is not as sensitive to extreme values.

The results show citizens’ dispositional perspective-taking and their reflective judgements are positively associated (without outliers: pt₁: r_s = 0.61, p < 0.00; pt₂: r_s = 0.30, p < 0.02; with outliers: pt₁: r_s = 0.54, p < 0.00; pt₂: r_s = 0.28, p < 0.02), although the strength of this correlation depends on the item used to measure dispositional perspective-taking. In other words, people who are most disposed to perspective-taking are more likely to engage in reflective political judgements. The excerpt that follows from my qualitative interviews illustrates the role the perspective-taking process plays in citizens’ reflections on a highly contentious issue – abortion:

[F]rom the very first day at the table I was on, there were elderly people, who would have been raised in a very catholic environment, who would have been . . . from rural environment . . . you know . . . they were totally opposed to abortion. They made it totally clear from day one that they had huge issues with this. Towards the end, they would say: ‘that is still how all stands, that is how I feel . . . But I totally understand why things need to change for other people’. . . . You know, hang on . . . This affects other people, who are younger, who have passed this. . . . ‘I will vote this way, even though this goes against my beliefs’. . . . I thought it was empathetic, brave and unselfish. I could see these people were really struggling, that was the core of their belief, that was how they grew up, that was ingrained [in them], you know. Yet, they decided that changes need to happen, even though [they] did not believe in it. . . . I think it was a very hard decision for many people. It was not just some sort of a throw-away decision, you know, there was a lot of soul-searching for a lot of people.

In sum, these findings show that citizen deliberators’ dispositional perspective-taking is associated with their willingness to engage in reflective political reasoning.

The quantitative data from the ICA, albeit having high ecological validity, have (at least) two limitations. First, dispositional perspective-taking was measured post-deliberation and, thus, it is challenging to discount the confounding effect of prior deliberations on citizens’ levels of dispositional perspective-taking. Citizen deliberators may have become better perspective-takers as a result of deliberations. Second, it is difficult to make causal claims about the relationship between deliberation, perspective-taking and reflective judgements by relying on observational and interview data.

With these limitations in mind, I subjected a part of my theoretical model to an experimental test. More specifically, I aimed to examine the relationship between dispositional perspective-taking and reflective judgement, and whether the latter is affected by small group deliberations. A laboratory experiment with UK residents was designed and fielded, in which deliberation was manipulated by inviting subjects to participate in a small group discussion on the issue of legalizing assisted dying. The study consisted of two waves. In wave 1 (on average, 7 days prior to the experiment), I fielded a survey to 600 UK residents (55% student and 45% non-student; 68% female; 55% 18–26 years old; 32% having at least a bachelor’s degree or equivalent),⁶ recruited through a behavioural laboratory of a university from the United Kingdom. In this survey, individuals’ dispositional perspective-taking was measured with a full battery of seven items from Davis’ IRI questionnaire (see the Supplemental, online appendix). For the purposes of the analyses, a sum index was created (M = 23.36; SD = 4.42; range: 9 to 35), which was standardized for ease of interpretation (M = 0.00; SD = 1; range: –3.25 to 2.63).

In wave 2 of the study, a subgroup of subjects were randomly assigned to one of three experimental conditions: deliberation, placebo or control. Those in the deliberation group were invited to take part in a small group (8–12 participants in each) discussion in the laboratory. Prior to discussions, participants read a short article about legalizing assisted dying that featured four arguments in favour and four arguments against (see the Supplemental, online appendix). Group discussions lasted 45–70 minutes. Consistent with the ideals of deliberation, I ensured that the subjects (a) received balanced information, (b) had a facilitator, (c) understood and agreed upon the rules of group deliberation and (d) were encouraged to express their thoughts and arguments through different communicative tools (such as personal stories, narratives and so on, in addition to fact-based argumentation). The only condition that the experiment did not meet was the presence of diverse perspectives in each discussion group. Most people who self-selected to participate in the experiment, having been previously notified about its objective and the possibility of being randomly assigned to a discussion group, were in favour of legalizing assisted dying.

Immediately after the discussion, respondents completed a post-intervention survey, in which outcome variables were measured. Those randomly assigned to the placebo group read the same short article with arguments and completed a post-intervention survey. Participants randomly assigned to the control condition only completed the post-intervention survey.

A total of 127 subjects participated in the study: 37 in the deliberation, 59 in the placebo and 31 in the control conditions.⁷ The unbalanced nature of the sample was due to the following factors. First, the objective was to have 60 subjects in the deliberation condition and 60 subjects in the placebo condition. There was, however, drop-outs from the experiment. Attrition was the greatest in the deliberation group, due to heavy snow on the day of discussions and, potentially, the demanding nature of deliberations. Post hoc analyses showed that the attrition was not systematic according to either the attitudinal measures or the sociodemographic characteristics. Second, I opted for a smaller sample size for the control group, due to the costly nature of the laboratory experiment and the relative lack of interest for the control condition.⁸

Consistent with the first study, reflective judgements are captured with the deliberation within measure (Weinmann, 2017). A summary index of a full battery of items was created (M = 26.03; SD = 4.41; range: 11 to 35). All items loaded in one factor well, and the Cronbach’s alpha was acceptable (0.68). The index was standardized for ease of interpretation (M = 0; SD = 1; range: –3.4 to 2.03).

Figures 4 and 5 visualize the range and spread of dispositional perspective-taking and reflective judgements among participants (n = 126).

Table 1 shows the means and standard deviations of reflective judgements across the three experimental conditions.

As can be seen, those in the deliberation group report having reflected about their choices more than those in the other two conditions. However, the results of difference in means t-tests show that this difference is not statistically significant at p < 0.05 significance level.

To estimate the relationship between deliberation, dispositional perspective-taking and reflective judgements, I use linear regression analyses. Table 2 presents the results of these analyses.

Model 1 estimates the bivariate relationship between dispositional perspective-taking and reflective judgements. It shows that a one unit increase in dispositional perspective-taking leads to a 0.32 SD increase in reflective judgements. Model 2 examines the effect of experimental interventions (information and deliberation) on the reflectiveness of citizen deliberators’ political judgements, controlling for their dispositional perspective-taking. While the direction of the coefficients for both information and deliberation is consistent with theoretical expectations that they would have positive effects on reflective judgements, neither effect is statistically significant at a conventional significance level (p < 0.05). Furthermore, the data do not allow us to directly test whether deliberation induced the perspective-taking process in citizen deliberators. Furthermore, care must be taken in interpreting this finding. One of the key conditions for the process of perspective-taking that is identified in the theory developed in this article – the presence of a diversity of viewpoints on the issue – could not be guaranteed in this experiment. Most of the discussion groups in this experiment consisted of like-minded individuals. This was different in the Irish case in which deliberators were exposed to opposing viewpoints during deliberations.

A healthy democracy needs citizens to make reflective and well-thought-out political judgements, as opposed to unreflective and impulsive ones. Proponents of deliberative democracy suggest that citizen deliberation is capable of inducing more reflective political thinking among people. Yet, little is known about the mechanisms underlying this relationship. This article offers a bridge between psychology and political science and proposes a theory of perspective-taking in deliberation. It argues that under the right conditions, deliberation produces more reflection when citizens make judgements by inducing the process of perspective-taking – actively imagining others’ experiences, perspectives and feelings – in citizen deliberators. Two institutional features of citizen deliberation are particularly important in this respect: the presence of a diversity of viewpoints and the interplay of fact-based rational argumentation and storytelling.

Although my contribution is primarily theoretical, I illustrate my case by studying the real-world and influential deliberative forum, the Irish Citizens’ Assembly. To examine the first part of the theory – that citizen deliberation induces the process of perspective-taking among citizens – I rely on in-depth interview data obtained from the members of the ICA. The patterns identified in the qualitative data provide evidence for the proposed theory.

To indirectly study the second part of the theory – the beneficial effect of perspective-taking on reflective reasoning – I rely on unique survey data obtained from members of the ICA and a laboratory experiment. The results across two studies (the ICA and a laboratory experiment) show that citizens’ dispositional perspective-taking influence the effectiveness of deliberation in activating reflective political reasoning among citizen deliberators. In other words, while deliberation may provide the forum in which perspective-taking and reflection are activated, dispositional perspective-taking is a pre-requisite for these processes.

Of course, this research is not without its caveats. The quantitative studies used in this article have several limitations that are worth discussing. First, neither the survey nor the experimental data are able to directly capture the complex processes of perspective-taking and reflection. Instead, they rely on self-reported measures of dispositional perspective-taking and reflective judgements. Thus, these studies are only indirect tests for the proposed theory. Future research could endeavour to improve on these measures. For example, in future research, reflection could be captured through thought-listing techniques (Weinmann, 2017) or the cognitive complexity of reasoning measure (Colombo, 2018).

Second, perspective-taking may be challenging for people. A person may be more likely to take the perspective of someone with whom he or she shares some common characteristics. For example, women may be more likely to take the perspective of a woman, and vice versa. As a result, the process of perspective-taking may be biased. Scudder (2016) posits that the success of perspective-taking is contingent upon unrealizable assumptions, such as the commonality between the observer and the target, and that empathy in general has the potential for distracting the citizen deliberators from listening to each other. The data in this research do not allow me to study these limitations. Elsewhere with different data, I find no empirical evidence for this claim (Muradova, 2019). However, future research could design experiments that could manipulate the common ground between the target and the perspective-taker and study its moderating role in the relationship between perspective-taking and reflection in deliberative settings.

Third, future research could endeavour to experimentally manipulate perspective-taking and test its effect on reflective judgements in order to contribute to our understanding as to whether reflective judgement truly causally follows perspective-taking (similar to Muradova, 2019).

Fourth, people who agree to attend deliberative forums may be better perspective-takers and reflective thinkers than non-attenders. In other words, despite the recruitment process being random, deliberative events may suffer from a self-selection problem. Although the Irish data does not allow me to test this contention, I can examine it with my experimental data. The findings show that there is no significant difference in mean (standardized) dispositional perspective-taking between those who attended the experiment and those who did not (M = –0.02; SD = 1, n = 472; M = 0.06; SD = 1, n = 126, p < 0.41). However, real-world deliberative events are more demanding, require more commitment from citizens and usually last longer. Therefore, future research could test this assumption with data from real-world deliberative forums.

Finally, this study does not allow for a direct and causal test of the effect of different levels of deliberation on perspective-taking and reflective judgements. Future research could address this shortcoming. One way to go, for example, would be to transcribe and code group deliberations for their deliberativeness and study the relationship between different levels of deliberation and perspective-taking. Another strategy would be to experimentally vary the different elements of deliberation and study their potentially differing implications for the processes of perspective-taking and reflection. In other words, future research could endeavour to test the proposed theory with data that would permit a mediation analysis, in which the effect of the levels of deliberation on the mediator (perspective-taking) and the outcome variable (reflective judgements) could be examined.

These shortcomings notwithstanding, this study is a first of its kind to offer a systematic study of the role of perspective-taking in deliberation and reflection. The findings contribute to several different strands of literature.

First, citizen deliberation forums seem to be the most favoured democratic innovation nowadays. Although their increasing usage across the globe is laudable (see www.participedia.net), the lack of a systematic understanding of the processes underlying the so-called deliberative transformations observed as a result of these forums is worrisome. This study endeavoured to contribute to this gap by systematically examining the processes underlying deliberation and reflective reasoning among citizens.

Second, this research speaks to the body of scholarship within deliberative democracy that is examining the microprocesses underlying deliberation (see Bächtiger and Parkinson, 2019 for a review). It also brings a renewed focus on the arguments by difference and feminist democrats about the importance of embracing different kinds of communication such as storytelling, narratives and testimony in deliberation (Black, 2008; Sanders, 1997; Young, 2000). It suggests that in addition to making deliberation more democratic and egalitarian, they have the potential for facilitating greater understanding and perspective-taking among individuals. Although evidence-based reasons help individuals to gain new information and knowledge about the issue under discussion, personal stories and testimonies enable them to get acquainted with the true identities, values and worldviews of the storytellers (Black, 2008). However, my study shows that stories and testimonies told in isolation clearly cannot convey the full picture. Only if accompanied by factual information and argumentation can they inform the citizens of the nuances and complexities of the issue under discussion and facilitate the process of perspective-taking.

Third, these findings contribute to the scholarship in political psychology that is studying the correlates of reflective reasoning (e.g. Arceneaux and Vander Wielen, 2017). It shows that perspective-taking is beneficial for people’s willingness to engage in more reflective reasoning.

Fourth, the research contributes to the strand of literature in social psychology that is looking at the inter- and intrapersonal effects of perspective-taking, such as decreased intergroup prejudice, and altruism (see Todd and Galinsky, 2014 for a review) and expands on it with empirical evidence on the role of perspective-taking in people’s political reasoning.

Fifth, these findings have implications for systemic approaches to deliberative democracy (Mansbridge et al., 2012). The deliberative systemic approach ‘has emphasized multiple sites of communication, each of which can host various forms of speech that can enrich the inclusive character of a deliberative system’ (Curato et al., 2017: 30). To the extent that perspective-taking induces more reflection when citizens make judgements, it can be evoked and nurtured in different areas of democratic politics, for instance, through empathy training at schools (Morrell, 2007), choose-your-own story games among adolescents (e.g. Simonovitz et al., 2018) and empathy-inducing ads in the media.

The author is grateful to David Farrell and Jane Suiter for the opportunity to observe the meetings of the Irish Citizens’ Assembly and field her survey questions among its members. She thanks the Secretariat of the Assembly for facilitating her in-depth interviews with the assembly members and the interviewees for their time and contribution. She extends her sincere gratitude to André Bächtiger, Nicole Curato, David Farrell, Andrea Felicetti, Sofie Marien, Simon Niemeyer and Jane Suiter for reading and commenting on previous drafts of this article. Parts of this article was presented at the European Consortium for Political Research (ECPR) General Conference, Hamburg, 2018 and Politicologenetmaal – Dutch-Belgian Political Science Annual Conference, 2018 . The laboratory experiment was conducted at the Nuffield College Centre for Experimental Social Sciences (CESS), University of Oxford, while the author was a visiting scholar at the Nuffield College. She gratefully acknowledges the support of the CESS in fielding this experiment. She is also thankful to Wouter Vanbroekhoven for his research assistance. Finally, the author thanks the editor and the anonymous reviewers for their insightful comments and suggestions.