Educating the Public on Evidence-based methods for improving inter-group civility.

Effect of Explanation on Understanding and Position Extremity

1. What They Did-Summary:

This study primarily focused on the effects that explaining one’s position on policy can have on feelings of understanding and position extremity.  198 participants were split into two groups.  One group was asked to rate their support or opposition to six different political policies such as instituting a flat tax or instituting a single-payer health care system.  Next, the participants were asked to rate their understanding of these policies.  Afterwards, they were asked to provide a mechanistic explanation for how two of these policies (chosen at random) work.  Lastly, the first group of participants were asked again to rate their position on and understanding of the issues.  The second group differed from the first in that they were only asked to rate their position on and understanding of the issues after they had provided an explanation of the policies.

The researchers predicted that people who hold extreme positions are under the illusion that they know more about policies than they really do.  Thereby, by having the participants explain how these policies work, they should realize this illusion of understanding and shift to a more moderate viewpoint.

2. What They Found-Results:

The researchers found a significant decrease in ratings of understanding following explanations.  They also found that people’s positions on the issues became significantly more moderate following explanations.  These findings confirmed the researchers’ predictions.

3. Who Was Studied-Sample:

198 U.S. residents recruited using MTurk, 52% male, 48% female.  40% Democrat, 20% Republican, 36% independent, 4% other.

4. Study Name:

Fernbach et al. 2013, Study 1.

5. Citation:

Fernbach, P.,  Rogers, T., Fox, C., and Sloman, S. “Political Extremism Is Supported by an Illusion of Understanding.” Psychological Science (2013): 1-8.

6. Link:

7. Intervention Categories:

mechanistic explanation, MTurk, judgment timing

8. Sample Size:


9. Central Reported Statistic:

Understanding: “This prediction was confirmed by a significant main effect of judgment timing: Postexplanation ratings of understanding (M = 3.45, SE = 0.12) were lower than preexplanation ratings (M = 3.82, SE = 0.11), F(1, 197) = 34.69, p < .001, ηp2 = .15.”

Position Extremity: “This prediction was confirmed, with the main effect of judgment timing significant (preexplanation-rating conditions: M = 1.41, SE = 0.07; postexplanation-rating conditions: M = 1.28, SE = 0.08), F(1, 86) = 6.10, p = .016, ηp2 = .066.”

10. Effect Size:

Understanding: t(5) = 5.74, p < .01.

Position Extremity: t(5) = 3.93, p = .011.

Read Ahead

Political Discrimination as Normative as Racial Divisions once were

Once upon a time, it was socially normative for society to divide itself along racial lines.  Thankfully, that time has passed and while racism still exists, it is generally considered to be a bad thing by most people in society.  The same trajectory is occurring with respect to attitudes toward homosexuals, with increased acceptance being not only encouraged, but mandated as the right thing to do.  However, in many circles, it remains normative for individuals to discriminate against those with the opposite political views.  Recent research indicates that this occurs amongst both parties.

Despite ample research linking conservatism to discrimination and liberalism to tolerance, both groups may discriminate. In two studies, we investigated whether conservatives and liberals support discrimination against value violators, and whether liberals’ and conservatives’ values distinctly affect discrimination. Results demonstrated that liberals and conservatives supported discrimination against ideologically dissimilar groups, an effect mediated by perceptions of value violations. Liberals were more likely than conservatives to espouse egalitarianism and universalism, which attenuated their discrimination; whereas the conservatives’ value of traditionalism predicted more discrimination, and their value of self-reliance predicted less discrimination. This suggests liberals and conservatives are equally likely to discriminate against value violators, but liberal values may ameliorate discrimination more than conservative values.

In addition, recent research out of Stanford University indicates that “hostile feelings for the opposing party are ingrained or automatic in voters’ minds, and that affective polarization based on party is just as strong as polarization based on race.”  Tackling this at the societal level is a daunting task for anyone, but there are things that one can do at the individual level.  Both research and practice indicates that positive relationships between individuals across such divides are likely to ameliorate such feelings.  Mixing group boundaries are likely to make competition less salient as well, perhaps allowing superordinate goals that we all share to come to the fore, as often happens when national emergencies strike.  Just as with discrimination based on race and sexual orientation, discrimination against opposing ideologies can be combated with similar techniques.

- Ravi Iyer


Read Ahead

When Morality Threatens Civility

Politics is fundamentally a group phenomenon that should be examined in relation to person’s identification with a particular group label (see Allport, 1954). When a person identifies as a “conservative” or “liberal” it means that he or she not only assumes a particular set of political positions but also identifies with other partisans as well as the shared sense of reality implied therein (Devine, 2014; see also Jost, Ledgerwood, & Hardin, 2008). To define oneself in this way often implies the identification of opponents, rivals or even enemies (cf. Edelman, 1988) that, at increasing levels, becomes a means of further clarifying what it means to be a “conservative” or a “liberal.”

“Civility,” which is the concern of this group, should therefore be examined as an intergroup, rather than interpersonal matter. What, then, does research on intergroup relations have to say about increasing civility among political partisans? I would argue that the greatest threat to political civility within a stable democracy is the “moralizing” of political discourse as it pertains to political groups. Namely, it is one thing to view your political opponent as misguided or simply wrong; it is another thing entirely to view him/her and the group that he/she represents as immoral, transgressive, or just plain evil. When placed in the latter realm, political discourse becomes decidedly less civil as the concept of “loyal opposition” becomes not only oxymoronic, but tantamount to treason.

When Morality Threatens Civility

Why must we as social scientists, practitioners, and scholars be wary when political discourse becomes moralized? The simple answer is that morality is powerful. The capacity for moral judgment and moral action may be encoded into our very fiber as a species and may have been the building blocks upon which human civilization was founded (Haidt, 2007, 2012; Greene, 2013). Putting aside the ontology of human morality, a much more parsimonious reason, at least for my purposes, for why morality is powerful is that people tend to view moral values as being objectively true (Goodwin & Darley, 2008). That is, a moral statement (e.g., “It is wrong to kill.”) is perceived to be more like an empirically verifiable fact (e.g., “The Earth revolves around the Sun.”) and less like a statement of social convention (e.g., “An appropriate tip for your server is 15 to 20 percent of the bill.”).

And there’s the rub, so to speak. A person’s morality is rooted in beliefs that are perceived to be as true as the Earth revolves around the Sun and which also imply a proscriptive element: not only is it wrong to kill, but one ought not to kill. A person’s moral worldview not only describes social reality but also guides future behavior as well as how future behavior is to be evaluated. One only needs to consult the work of Linda Skitka on moral conviction (see Skitka, 2010) or Jeremy Ginges’ work on sacred values (e.g., Atran & Ginges, 2012; Ginges & Atran, 2011) to see how these aspects of morality function in politics. What their work demonstrates is that we often judge the actions of others in relation to whether they reflect or confirm our moral values, even if violates considerations of procedural justice (Skitka & Houston, 2001) or our own utilitarian benefit (Ginges & Atran, 2011).

The “moralizing” of intergroup relations is often reflected in the attribution of moral or immoral qualities to other groups. I am currently examining the consequences of this process as part of my dissertation. In my preliminary findings (see Pilecki et al., 2013), I have found that when people perceive that members of another group (e.g., liberals, conservatives, feminists, evangelicals, etc.) as being typically less moral than most other people they are more likely to view violence or acts of political repression towards that group as being appropriate. These findings reinforce previous empirical and theoretical work by Susan Opotow (1990, 1993, 1994) and others (e.g., Bar-Tal, 1990) on the “scope of justice,” which refers to the distinction that people make between those considered worthy of moral treatment and those considered unworthy of moral treatment. When a social group is imbued with immoral qualities by political leaders, pundits, or other “entrepreneurs of identity” (Reicher & Hopkins, 2001) that group is effectively set apart from others and, in effect, becomes a legitimate and morally justifiable target of harm rather than civil discussion.

When people moralize intergroup relations they limit the potential for civil discourse to emerge as they frame political issues within the realm of sacred values, thereby making trade-offs and compromises less likely (Tetlock, 2003; Tetlock, Kristel, Elson, Green & Lerner, 2000). The words and labels we use to describe the social groups with which we identify and those that we oppose shape how we think about us, them and how we relate to one another (see Hammack & Pilecki, 2012). In other words, language matters and it is therefore incumbent for social scientists and practitioners to hold political leaders, media figures and other influential people accountable for their use of moralizing rhetoric to mobilize support, gain more viewers, and/or delegitimize criticism.

Read Ahead

The First New York City Asteroids Club Dinner

The Asteroids Club is any group of people  with diverging political views who gather not to debate, but to listen to the other side explain why it is concerned about certain threats. The metaphor is meant to capture the fact that there are many threats coming at the United States–like asteroids scheduled for direct hits—yet each side of the political spectrum focuses on a few of them and ignores or discounts the asteroids that most worry the other side. As John Stuart Mill said in 1840: “in almost every one of the leading controversies… both sides were in the right in what they affirmed, though in the wrong in what they denied.” The Asteroids Club is a novel format for bringing people together, over a meal, for a discussion in which each side helps the other to see more clearly. The format has been developed by The Village Square, in Tallahassee Florida.

The Nathan Cummings Foundation hosted the first New York City Asteroids Club dinner on February 26, 2014, in the home of its president, Simon Greer. The foundation is in the process of updating its research and funding portfolios, and its leaders were interested in learning more about the issues of income inequality and the causes of poverty. They were also interested in developing relationships with experts from diverse perspectives who could help them understand these complex topics. Simon and I therefore chose the two asteroids of rising income inequality (a topic of greater concern on the left), and declining rates of marriage and family stability (a topic of greater concern on the right).

This report is intended to be useful for anyone interested in hosting future dinners, or in bridging the political divide more generally. However, we note that we decided to invite experts from right and left to the first dinner, rather than ordinary people, to maximize the degree to which we could learn about the best thinking and research on these topics. This dinner may therefore not be representative of what will happen if the Asteroids Club format is rolled out and used widely by civic groups across the nation.

Preparing For the Dinner

We ran the dinner using the Chatham House rule, which says that participants are free to write about the event, but cannot reveal the identities or affiliations of the speakers. So we will simply say that we invited two very prominent and well-respected journalists, one clearly identified with liberalism, one with conservatism. We asked these two co-hosts to suggest additional people they wanted on their “team.” (We had hoped to avoid the terms “team” or “side,” but it was hard to do so.) These two co-hosts then nominated several other experts, and took a few suggestions from Simon and me about experts we know on both sides. All 10 people who were invited accepted the invitation. The final dinner was attended by 18 people: 6 liberals, 6 conservatives, Simon and me (who served as co-hosts), and four observers affiliated with the foundation. The expertise in the room was extraordinary, including journalists, economists, activists, and people with experience in government and policy-making.

Two weeks before the dinner, we assembled a list of readings nominated by the participants and posted them all in a google document that we used to share information and coordinate the evening. Because these prominent participants were all extremely busy, we did not push for them to do much preparation beforehand, although we did encourage them to read one or two articles from the other side, and to watch my TED talk that introduced the idea of the asteroids club.

The Dinner

The evening began at 7:30 with drinks, hors d’oeuvres, and unstructured socializing, which was quite cross-partisan. At 8:00 the two teams convened separately to go over final plans for their presentation. (In retrospect we should have encouraged the two teams to do more of this online, before arriving at the dinner). Around 8:30 everyone moved into the dining room and sat in seats that had been assigned to ensure a good mix of people at each of the two long tables. Simon welcomed everyone to his home and presented the Hebrew conception of two kinds of argument: machloket l’shem shamayim – argument that is for the sake of heaven which is intended to better discern truth and move the world forward,  and 2) machloket she’lo l’shem shamayim – an argument that is not for the sake of heaven, but is just for the sake of itself or for the sake of being controversial. Argument in pursuit of truth is considered sacred in the Jewish tradition.

I then explained the history of the Asteroids Club format, and went over the groundrules, including the Chatham house rule, the exact schedule, and the role that I would play as timekeeper and moderator. Each of the participants then offered a 1-minute introduction of him- or herself. We then served ourselves dinner from the buffet, and began the structured discussion. The schedule that we intended to follow was this:

9:00 begin Asteroid #1: rising inequality, liberal side presents
–15 minutes to describe the threat; why this matters, why it’s urgent….During this time, the listening side gets to ask a few brief elaborative questions, but hold argumentative questions for the next step.
–5 minutes for “telescope” time — conservatives ask critical questions, request more info, challenge assumptions.

9:20 begin Asteroid #2 family breakdown, conservative side presents
–15  minutes to describe the threat; why this matters, why it’s urgent…. During this time, the listening side gets to ask a few brief elaborative questions, but hold argumentative questions for the next step.
–5 minutes for “telescope” time — liberals ask critical questions, request more info, challenge assumptions.

9:40: dessert served, short break
9:45 Integrative discussion, 20 min.
–10 min: Return to Asteroid #1: Inequality. Go deeper: what does the liberal side most want to change, address, or at least have understood (taking into account family breakdown)
–10 min: Return to Asteroid #2: Family Breakdown: Go deeper: what does the conservative side most want to change, address, or at least have understood? (taking into account rising inequality)

10:05 Concluding discussion and resolutions.
10:30: End of formal discussion, take closing survey.

The schedule we actually followed was close to this, but because we were running behind and because there was clearly a surprising amount of agreement on what aspects of poverty were of great concern, I decided to merge the two parts of the integrative discussion into a single discussion of what really matters – what issues either side had raised that seemed to elicit at least some assent from the other side.

In general there was bipartisan agreement that income inequality has been rising in recent decades, although there was NOT bipartisan agreement that rising inequality itself was a problem that needed to be addressed. There was, however, bipartisan (though not necessarily unanimous) agreement that the following issues are concerns or problems, and that we would be a better country if we could address them:

  • Dignity, and the indignities disproportionately suffered by the poor

  • Distrust of government and democracy

  • Waste: the financial squeeze and inefficiency caused by “arms races” as people spend more money to attain “positional goods” such as a home in a good school district.

  • Lack of opportunity for large segments of society; The waste of human potential among the poor.

  • Declining motivation and economic dynamism – which can be stimulated by moderate degrees of inequality (this point was noted by a liberal)

  • Abuses of power, which becomes easier for those with a lot of money

  • Separateness – having communities that are cut off from the mainstream of society; having low social trust and cohesion.

In general there WAS bipartisan agreement that marriage rates and family stability have declined in recent decades. There WAS bipartisan agreement that these trends are bad for society, and that it would be good if we could find ways of reversing the trends. There was a very high degree of bipartisan concern about poverty, especially for its pernicious effects on children. There was also a consensus that the people who suffer most from these trends are single mothers (who bear the main costs and stresses of raising children under often adverse circumstances) and their sons. Daughters suffer too, of course, but there was consensus that boys’ outcomes are more adversely affected by the absence of a father, and these adverse outcomes then set up a feedback loop for the next generation in which there are way too few stable, employed, and marriageable young men who could break the cycle of father-absence.

The Post-Dinner Assessment

We are working with my colleagues at to develop assessment tools that can be used by any organization running any kind of civility-enhancing event. For future events, we will send all participants a link to a web-based survey  a few days before the event, then pass out a paper questionnaire at the conclusion of the event, then send out a web-based followup survey a week after the event, to assess the effects of the event on attitudes about the asteroids, and about the people on the other side. For this first dinner, with 12 super-busy people, we only obtained four responses to the first web survey. (You can see that first survey here. Future versions will be much more extensive.) We therefore decided to focus our efforts on obtaining measures of the key variables at the conclusion of the evening. Did people come to see the asteroid presented by the other side as a clearer or more pressing threat than they had before the dinner? That is the central goal of an Asteroids Club dinner.

We collected responses from all 6 liberals and from 5 of the conservatives.  For each asteroid, we asked participants whether their views had changed regarding both the PACE of the threat and the SEVERITY of the threat.  For example, here is the exact text of one of the four main questions:

How has tonight’s discussion influenced your beliefs about the severity of the problem of income inequality for the USA?
___I now see it as a much less severe problem than I did a week ago
___I now see it as a slightly less severe problem than I did a week ago
___The discussion did not influence my beliefs in either direction
___I now see it as a slightly more severe problem than I did a week ago
___I now see it as a much more severe problem than I did a week ago

The graph below shows the results. Let’s start with family decline. The liberals (shown on the left half) DID move in the desired direction. They now see it as a more rapidly approaching threat (gold bar, 4 moved), and a more severe threat (purple bar, 5 moved). The conservatives did not really move on family decline, nor did we expect them to. (The gold bar doesn’t even show  because nobody moved.)

On inequality, the story was quite different. No conservative moved at all on the speed of the change (which is why the blue bar does not show), and the only movement on severity came from a single conservative who said that he/she now sees the problem as “slightly less severe” than before. (The liberals also showed no movement on inequality, except for a single participant who said “slightly more severe” than before.)

What can we make of this pattern?  It appears that the asteroids club format worked for the family decline asteroid. The conservatives presented the asteroid, the liberals listened, and then came to see it as a greater threat. This is very encouraging.

But it is harder to know how to interpret the results on the inequality asteroid. The lack of movement could have resulted because A) the liberal team made a weaker case about inequality than the conservatives had for family decline, or B) the conservatives were less open-minded and willing to listen, or C) the facts about inequality and its harms are truly more ambiguous and contestable than they are for the family decline asteroid. My conclusion from the background reading, and from the conversation during the dinner, is that C is true and is at least part of the explanation.

We also asked: “Compared to other discussions about policy and politics you’ve had in politically ‘mixed company,’ how enjoyable was this ‘asteroids club’ format?”  We offered 5 choices, ranging from “much less enjoyable” (scored as -2) to “much more enjoyable” (scored at +2). Six respondents said “much more enjoyable,” one said “about the same,” and four said “slightly more enjoyable.” There was no significant difference between the liberals and conservatives, which is quite encouraging


Conclusions and advice for future Asteroids Club dinners:

Based on the data above, and on comments made by participants after the dinner, I draw the following conclusions and lessons:

1) The format is enjoyable and promotes civil interaction. Sharing a meal in a private home seems to have made people particularly polite and open.

2) It is difficult to do two asteroids in a single 2 hour discussion. There are advantages to doing two – it led to a sense of balance and fairness. But at least 3 hours of discussion would have been needed to cover both.

3) Relationships matter, and it takes time to get to know each other and develop trust. Ideally, asteroids clubs will be true clubs, with a stable membership that meets every month or two.

4) The exact instructions for each part of the evening matter. More work is needed, drawing from experts in negotiation and facilitated discussion, on how to optimize the integrative discussion and conclusion.

5) More “warmup” exercises could be tried. We kept this evening fairly direct and cerebral – focused on the ideas and research. But future events could begin with more activities to build trust and cohesion, such as singing the National Anthem, or doing introductions in a much more personal way, or pairing off in bipartisan teams for short initial discussions, and then having participants report to the group by introducing their partner and his/her main concerns.

The bottom line is that we created a novel social situation which called for openness, trust, and collaborative thinking, and the participants rose to the occasion. There was no partisan sniping and not a shred of hostility. People sometimes made points that supported the other side. Many on both sides expressed a desire to continue the conversation. Which we will do, in a way soon to be announced.

If you are interested in hosting an Asteroids Club dinner yourself, please visit:

Jon Haidt

Read Ahead
Our goal is to educate the public about social science research on improving inter-group relations across moral divides.