From Gulf to Bridge: When Do Moral Arguments Facilitate Political Influence?

Abstract

Much of contemporary American political rhetoric is characterized by liberals and conservatives advancing arguments for the morality of their respective political positions. However, research suggests such moral rhetoric is largely ineffective for persuading those who do not already hold one's position because advocates advancing these arguments fail to account for the divergent moral commitments that undergird America's political divisions. Building on this, we hypothesize that (a) political advocates spontaneously make arguments grounded in their own moral values, not the values of those targeted for persuasion, and (b) political arguments reframed to appeal to the moral values of those holding the opposing political position are typically more effective. We find support for these claims across six studies involving diverse political issues, including same-sex marriage, universal health care, military spending, and adopting English as the nation's official langauge. Mediation and moderation analyses further indicated that reframed moral appeals were persuasive because they increased the apparent agreement between the political position and the targeted individuals' moral values.

Full text

Personality and Social
Psychology Bulletin
1 –17
© 2015 by the Society for Personality
and Social Psychology, Inc
Reprints and permissions:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/0146167215607842
pspb.sagepub.com
Article
American politics has reached historic levels of partisan divi-
sion and animosity (e.g., Pew Research Center, 2014). In
explaining political convictions and the divisions they produce,
many researchers argue that the most strongly held and con-
tested political views are those grounded in moral convictions
(Janoff-Bulman, 2009; Lakoff, 2002; Skitka & Bauman, 2008).
Consistent with this, research finds that liberals and conserva-
tives possess systematically different moral profiles, leading to
divergent moral intuitions and attitudes (Caprara, Schwartz,
Capanna, Vecchione, & Barbaranelli, 2006; Graham, Haidt, &
Nosek, 2009; Thorisdottir, Jost, Liviatan, & Shrout, 2007).
Because individuals’ moral convictions are typically experi-
enced as factual and universally applicable (Skitka, Bauman, &
Sargis, 2005), political attitudes rooted in morality are espe-
cially inflexible and resistant to influence (Skitka & Morgan,
2009; Wright, Cullum, & Schwab, 2008). Despite the inflexi-
bility of political positions rooted in morality, political advo-
cates on both sides of the ideological spectrum regularly
advance moral arguments for their political positions. Indeed
much of contemporary U.S. political rhetoric is characterized
by liberals’ and conservatives’ competing arguments for why
their respective political positions are morally correct (Lakoff,
2002). But do advocates’ messages succeed in winning over
their rivals?
Here, we explore when morally based political messages
are effective. We hypothesize that political advocates tend to
spontaneously craft arguments grounded in their own sense
of morality rather than the morality of those targeted for per-
suasion, but that such moral arguments will typically be less
persuasive than arguments reframed to appeal to the values
of those supporting the opposing political position. Because
moral convictions can shape political attitudes (Graham
et al., 2009; Lakoff, 2002; Morgan, Skitka, & Wisneski,
2010), framing a given position as consistent with individu-
als’ moral commitments should increase their willingness to
adopt that position, even if they currently oppose it. This
analysis suggests that the tight relationship between moral
convictions and political views—so often a source of divi-
sion and rigidity—can also be a basis for opinion change,
political influence, and coalition formation.
Our analysis also provides one explanation for why politi-
cal advocates likely do not employ this rhetorical strategy;
their own moral convictions reduce both their capacity and
607842PSPXXX10.1177/0146167215607842Personality and Social Psychology BulletinFeinberg and Willer
research-article2015
1University of Toronto, Ontario, Canada
2Stanford University, CA, USA
Corresponding Author:
Matthew Feinberg, Assistant Professor of Organizational Behavior and
Human Resource Management, Joseph L. Rotman School of Management,
University of Toronto, 105 St. George Street, Toronto, Ontario, Canada
M5S 3E6.
Email: Matthew.Feinberg@rotman.utoronto.ca
From Gulf to Bridge: When Do Moral
Arguments Facilitate Political Influence?
Matthew Feinberg1 and Robb Willer2
Abstract
Much of contemporary American political rhetoric is characterized by liberals and conservatives advancing arguments for
the morality of their respective political positions. However, research suggests such moral rhetoric is largely ineffective
for persuading those who do not already hold one’s position because advocates advancing these arguments fail to account
for the divergent moral commitments that undergird America’s political divisions. Building on this, we hypothesize that (a)
political advocates spontaneously make arguments grounded in their own moral values, not the values of those targeted for
persuasion, and (b) political arguments reframed to appeal to the moral values of those holding the opposing political position
are typically more effective. We find support for these claims across six studies involving diverse political issues, including
same-sex marriage, universal health care, military spending, and adopting English as the nation's official langauge. Mediation
and moderation analyses further indicated that reframed moral appeals were persuasive because they increased the apparent
agreement between the political position and the targeted individuals’ moral values.
Keywords
moral psychology, political psychology, influence, political polarization, empathy
Received January 28, 2015; revision accepted August 31, 2015
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from

2 Personality and Social Psychology Bulletin
motivation to couch arguments in terms of the morals of
those targeted for persuasion (Ditto & Koleva, 2011; Skitka
et al., 2005), leading them instead to advance arguments
grounded in a morality that the targets of persuasion do not
endorse. In this way, the most effective moral arguments may
be the least obvious ones to those with strong political
convictions.
Morality and Political Attitudes
Recently, political psychology has emphasized the role that
moral concerns play in political attitudes and behavior
(Janoff-Bulman, Sheikh, & Baldacci, 2008; Morgan et al.,
2010). Along those lines, Marietta (2008) argues that citizens
tend to reason about a political issue in either a consequen-
tialist or an absolutist way. Where consequentialist reasoning
involves cost–benefit analysis and tends to afford compro-
mise, absolutist reasoning involves unwavering stances and
the rejection of any form of compromise (see Tetlock et al.,
2000). He finds that highlighting the moral sacredness of a
political issue engenders greater absolutist reasoning than
does discussing the issue using non-sacred language.
Furthermore, he shows that such sanctification leads to
polarization by eliciting more extreme political stances.
Similarly, other research has found that individuals treat
sacred issues differently, rejecting typically rational ways of
reasoning, like cost–benefit calculation, and prioritizing the
defense of moral principles over other, non-moral consider-
ations (Skitka & Mullen, 2002; Tetlock et al., 2000).
Recent research on moral intuitions further highlights the
significance morality plays in political reasoning (Haidt,
2012). Moral intuitionists argue that moral judgments are
primarily driven by emotion-laden moral intuitions. When
presented with a morally questionable behavior, individuals’
gut reactions determine whether the behavior feels immoral
or not, forming the basis of subsequent judgments (cf.
Feinberg, Willer, Antonenko, & John, 2012). This research
finds that liberals and conservatives often hold divergent
moral intuitions; what conservatives find intuitively immoral
may be perceived as perfectly acceptable to liberals, and vice
versa (Haidt, 2012). For example, in one study conservatives
rated homosexual sex to be more offensive than liberals did
because conservatives have much stronger moral intuitions
regarding traditional sexual purity (Haidt & Hersh, 2001).
However, liberals find failing to recycle more offensive than
conservatives because they have stronger moral intuitions
about harming the environment (Feinberg & Willer, 2013).
Moral Convictions and Political
Attitudes
The above suggests that divergent moral convictions help
account for the tensions between liberals and conservatives on
many political issues. But what are the specific moral convic-
tions liberals and conservatives endorse? Moral Foundations
Theory (MFT) provides a useful framework for examining
such differences. MFT researchers surveyed thousands of
people around the world regarding the concerns that were
morally relevant to them (Graham et al., 2009; Graham
et al., 2011). Results led to the proposal of five primary moral
foundations—harm/care, fairness/reciprocity, ingroup/loyalty,
authority/respect, and purity/sanctity. Importantly, this
research finds that liberals tend to endorse foundations based
on caring and protection from harm (harm) and maintenance
of fairness and reciprocity (fairness) more strongly than con-
servatives. However, conservatives tend to endorse moral con-
cerns related to ingroup-loyalty (loyalty), respect for authority
(authority), and protection of purity and sanctity (purity) more
than liberals. Related research largely supports these findings
(Caprara et al., 2006; Jost, Glaser, Kruglanski, & Sulloway,
2003; Lakoff, 2002; Thorisdottir et al., 2007), and together,
this body of work indicates that liberals possess stronger moral
convictions related to fundamental aspects of harm and fair-
ness (e.g., benevolence, nurturance, equality, social justice),
and conservatives possess stronger convictions related to
group loyalty, authority, and purity (e.g., patriotism, tradition-
alism, strictness, religious sanctity). As a result of these differ-
ing moral commitments, liberals and conservatives tend to
view the world through different moral lenses and often have
contrasting viewpoints on morally charged issues.
Moral Convictions and Political
Arguments
The moral value differences between liberals and conserva-
tives have been shown to influence how they view and dis-
cuss morally relevant topics. For instance, when recounting
pivotal moments in their lives, liberals rely more on the harm
and fairness moral foundations than conservatives, whereas
conservatives rely more on the loyalty, authority, and purity
foundations (McAdams et al., 2008). Similarly, sermons in
liberal church congregations tended to employ more themes
around the harm and fairness moral foundations than did ser-
mons in conservative congregations. Conversely, sermons in
conservative congregations were grounded in more conser-
vative moral foundations (e.g., authority, purity) than ser-
mons in liberal congregations (Graham et al., 2009).
Liberal and conservative political arguments likely reflect
these differences in morality as well, with liberals basing
their arguments in the more liberal foundations and conser-
vatives basing their arguments in the more conservative
foundations. Some evidence exists in support of this possi-
bility. An analysis of newspaper op-ed pieces and video
advertisements arguing for environmental protection—a pri-
marily left-leaning position (e.g., Pew Research Center,
2013)—found that when moral rhetoric was employed, it
was typically grounded in the liberal moral foundations
(Feinberg & Willer, 2013), suggesting liberal advocates
argue for environmental protection in moral terms that appeal
primarily to other liberals.
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from

Feinberg and Willer 3
Arguments that appeal to one’s own moral convictions
may resonate with those who ascribe to a similar political
ideology (because they share similar moral values), but these
arguments are likely unconvincing to those with a differing
political ideology (because they do not share the same moral
convictions). Nonetheless, there is reason to believe that,
even when targeting their political opponents for persuasion,
liberals and conservatives will rely predominantly on their
own moral values as the basis of their political arguments.
Scholars have conjectured that advocates rarely recognize
the moral foundational differences between themselves and
others (e.g., Haidt, 2012; Lakoff, 2002), particularly because
they suffer from what Ditto and Koleva (2011) theorize to be
a “moral empathy gap,” or the inability to comprehend moral
worldviews different from their own. Such a gap likely
impairs the perspective-taking needed to craft an argument
that appeals to the convictions of those with different moral
values (Waytz, Gray, Epley, & Wegner, 2010). Thus, it stands
to reason that when political tensions are rooted in divergent
moral perspectives, individuals on either side of the divide
will tend to speak past one another, advocating for a given
position in the moral terms they themselves find most con-
vincing—the same moral terms that may have led them to
adopt the position in the first place.
From this reasoning, we hypothesize the following:
Hypothesis 1: When targeting their political counter-
parts, both liberals and conservatives will employ moral
arguments that reflect their own, rather than their audi-
ence’s, moral values.
We also advance a second hypothesis:
Hypothesis 2: When targeting their political counter-
parts, moral arguments reframed to appeal to the audi-
ence's moral values will be more persuasive than
arguments framed in a way that appeal to the messenger’s
own values.
To that end, we test whether messages in support of tradition-
ally liberal (or conservative) positions aimed at conservative
(or liberal) targets will be more convincing when framed in
terms of moral values endorsed more by conservatives (or
liberals).
Some existing research suggests the plausibility of moral
reframing as a strategy for political persuasion. Within the
persuasion literature, research demonstrates the value of
matching the characteristics of a message with those of its
targets (Cacioppo, Petty, & Sidera, 1982; Petty & Wegener,
1998). A variety of studies show that messages are more per-
suasive when they match traits of the target—for example,
the target’s regulatory fit (Cesario, Grant, & Higgins, 2004),
self-schemata (Wheeler, Petty, & Bizer, 2005), personal
goals (Clary, Snyder, Ridge, Miene, & Haugen, 1994), self-
monitoring (Lavine & Snyder, 1996), and need for cognition
(Haddock, Maio, Arnold, & Huskinson, 2008). In addition,
experimental studies that manipulated whether participants
formed an affect-based or cognition-based attitude toward an
object found that subsequent arguments were more persua-
sive when they matched the type of initial attitude (e.g.,
Edwards, 1990; Fabrigar & Petty, 1999). According to this
literature, matched messages engender a sense that the mes-
sage “feels right,” and this feeling in turn influences judg-
ments of the message’s position and the larger issue in
question (Avnet & Higgins, 2006; Cesario et al., 2004; Lee &
Aaker, 2004). We propose that morally reframed messages
that fit with recipients’ values should be particularly persua-
sive. The centrality of morality to an individuals’ self-
concept and worldview (Aquino & Reed, 2002; Haidt, 2001,
2007; Hofmann, Wisneski, Brandt, & Skitka, 2014) should
make moral arguments that fit with one’s values uniquely
impactful, and the tendency for moral issues to foster gut
intuitions and affect (Haidt, 2001) means moral arguments
may evade close, analytical scrutiny from their recipients.
A few studies have found evidence consistent with our
moral reframing hypothesis. Feinberg and Willer (2013,
Study 3) found that conservatives demonstrated greater sup-
port for pro-environmental legislation when advocacy state-
ments were framed in terms of the more conservative value
of purity than those presented with arguments framed in
terms of the more liberal value of harm or a control. Similarly,
Kidwell, Farmer, and Hardesty (2013) found that presenting
pro-environment arguments couched in terms of loyalty,
authority, and purity increased the likelihood that conserva-
tives would recycle, and found that these effects persisted
over a 14-week period. Together, these studies suggest moral
reframing can serve as an effective means for promoting pro-
environmental attitudes and behaviors in conservatives.
Neither of these studies, however, looks beyond environ-
mental issues, nor do they study whether moral reframing
might affect liberal targets. Another recent study found evi-
dence that moral arguments that fit a target’s political orien-
tation could be persuasive across multiple issues (Day, Fiske,
Downing, & Trail, 2014). However, this effect did not occur
consistently across studies, and was found only among con-
servatives, not liberals. It is possible that the inconsistency of
the moral framing effects in these studies owed to the spare
sentence-long, stimuli used in the studies.
The Present Research
We test our hypotheses across six studies. The first two
examine whether liberals (conservatives), when asked to
craft a persuasive argument to convince conservatives (liber-
als) to support a liberal (conservative) position craft the mes-
sage in terms of their own moral values, rather than the moral
values of the target of the message. Next, Studies 3 through
6 test the efficacy of our moral reframing hypothesis. In
each, we present a polarized political rhetoric and reframe
the moral rhetoric around it so that it is based on the moral
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from

4 Personality and Social Psychology Bulletin
foundations of the side traditionally opposed to it. We also
test our reasoning for why such moral reframing might be
influential, examining the extent to which reframed moral
appeals lead the target audience to consider the policy stance
to be more in line with their own moral principles.
Study 1
In Studies 1 and 2, we instruct participants to write persua-
sive arguments specifically aimed at influencing those sup-
porting a rival position. In Study 1, we recruited liberal
participants to write arguments supporting same-sex mar-
riage that would appeal to conservative targets. Then, in
Study 2, we recruited conservative participants to write argu-
ments in favor of making English the official language of the
United States that would appeal to liberal targets.1 We
hypothesize that both liberal and conservative advocates will
employ moral arguments, but will couch these arguments in
terms of moral values distinctive of their own ideological
group, not their rival’s.
Participants
Ninety-three (54 male, 39 female) participants were recruited
from the Amazon Mechanical Turk website for a small pay-
ment (please see online materials for more information about
determining sample sizes for our studies). We used a two-step
recruitment procedure. First, we invited participants to com-
plete a short prescreening demographic questionnaire that
included a question measuring their political ideology on a
scale from 1 (strongly liberal) to 4 (moderate) to 7 (strongly
conservative), and a second question that asked, “What is your
position on same-sex marriage? Are you in favor of legalizing
same-sex marriage in the U.S. or are you opposed to it?”
Participants indicated either in favor or opposed. Liberal par-
ticipants (“3” or below on the political ideology scale) in favor
of same-sex marriage were filtered into the study.
Procedure
All participants took part online. Participants read,
Now, we would like you to write a persuasive argument (4-5
sentences) aimed at convincing conservative Americans who
oppose same-sex marriage of why they should be in favor of
same-sex marriage.
Note: In a follow-up study, we will actually present your argument
to conservative Americans who oppose same-sex marriage to see
if they are persuaded by your argument. Participants who are
able to effectively persuade these future study participants will be
entered into a drawing for a $50 bonus.
Once participants finished writing their arguments, the
computer interface informed them that the study was
complete.
Argument Coding
Four research assistants unaware of our hypotheses, first
read background information about MFT and the five foun-
dations of morality (Haidt & Joseph, 2004). Then these
research assistants coded each participant’s argument on
three variables: Fit With Conservative Morality, Contradicted
Conservative Morality, and Fit With Liberal Morality. In
each case, coders indicated either “yes” or “no.” For Fit With
Conservative Morality, coders indicated “yes” for an argu-
ment if the argument’s rhetoric reflected at least one of the
three moral foundations endorsed at higher levels by conser-
vatives (loyalty, authority, and purity; e.g., “Our fellow citi-
zens of the United States of America deserve to stand
alongside us, deserve to be able to make the same choices as
everyone else can . . . Our goal as Americans should be to
strive for that ideal. We should lift our fellow citizens up, not
bring them down.”) For Contradicted Conservative Morality,
coders indicated “yes” for arguments that negated or chal-
lenged at least one of the more conservative moral founda-
tions (e.g., “Although you may personally believe your faith
should be against such a thing . . . your religion should play
no part in the laws of the United States.”). And, for Fit With
Liberal Morality, coders indicated “yes” for arguments in
line with the two more liberal moral foundations (e.g., “Why
would we punish these people for being born a certain way?
They deserve the same equal rights as other Americans.”).
For instances when arguments appealed to both liberal and
conservative morality, or appealed to and contradicted one
type of morality, coders indicated “yes” for each of these
instances—that is, arguments could be coded with more than
one “yes.” In addition, if participants received a “yes” for
any of the coded variables, they were also coded as having
made a moral argument.
Reliability across coders was moderate (average α = .70).
Instances where two coders selected “yes” and two selected
“no” were resolved by a fifth coder. Arguments that received
3 or 4 “yes” ratings were scored as fitting with the coded
variable, whereas arguments receiving fewer than 3 “yes”
ratings were scored as not fitting with the variable.
Results
Table 1 presents how often participants made any type of moral
argument, appealed to conservative morality, contradicted con-
servative morality, or appealed to liberal morality. In all, 86%
of participants employed moral rhetoric of some type, and, in
line with our hypothesis, only a small minority of liberals (9%)
made arguments that reflected more conservative moral prin-
ciples. Instead, as predicted, a majority of participants made
arguments reflecting liberal moral foundations in their same-
sex marriage arguments (74%). Furthermore, we found that a
portion of the participants wrote arguments that actually con-
tradicted conservative morality (34%). Comparisons of these
percentages using a McNemar’s test found that participants’
arguments were significantly more likely to reflect their own
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from

Feinberg and Willer 5
moral foundations than those of their conservative targets,
χ2(1) = 49.32, p < .001, and they were actually more likely to
contradict conservative morality than they were to appeal to it,
χ2(1) = 14.69, p < .001.
Study 2
Participants
Eighty-four (46 male, 38 female) participants recruited from
Amazon Mechanical Turk took part for a small payment.
Recruitment involved a similar procedure as Study 1, how-
ever instead of asking about same-sex marriage attitudes, the
prescreening questionnaire asked participants whether they
were in favor or opposed to making English the official lan-
guage of the United States (see online materials for complete
text). Only conservative participants (5 or higher on the
political ideology scale) in favor of making English the offi-
cial language were filtered into the study.
Procedure
The instructions were identical as those in Study 1 except
that Study 2’s instructions asked the participants to write a
persuasive argument aimed at convincing liberal Americans
who oppose making English the official language of the
United States why they should support making English the
official language.
Argument Coding
The coding procedure was the same as in Study 1 except that
in Study 2, we used five coders. Coders indicated whether or
not each argument Fit With Liberal Morality (e.g., “By mak-
ing English our official language, there will be less racism
and discrimination . . . upon the very groups that are being
persecuted.”), Contradicted Liberal Morality (e.g., “So those
of you preaching diversity and equality, who think everyone
should take advantage of us, should think real hard.”), and
Fit With Conservative Morality (e.g., “Every country needs
to be unified. Making English the official language will help
unify the country as we all can communicate with each other
and speak the same national language.”). In addition, as in
Study 1, if participants received a “yes” for any of the coded
variables, they were also coded as having made a moral
argument.
Reliability across coders was moderate (average α = .76).
In instances where three coders selected “yes,” a sixth coder
read the argument and provided an additional “yes” or “no”
rating. Arguments earning 4 or more “yes” ratings for a given
variable were then scored as being in line with that variable,
whereas arguments with 3 or fewer “yes” ratings for a given
variable were scored as not being in line with the variable.
Results
Table 1 presents how often participants made a moral argu-
ment of any type, appealed to liberal morality, contradicted
liberal morality, and appealed to conservative morality. A
total of 79% of participants’ arguments used some type of
moral rhetoric. Only 8% of participants wrote arguments
that fit with the more liberal moral foundations, whereas
70% wrote arguments that fit with their own more conser-
vative moral foundations. Furthermore, 14% of participants
wrote arguments that contradicted liberal moral values. As
in Study 1, McNemar’s test revealed the participants were
significantly more likely to write arguments reflecting their
own moral values than the values of their intended liberal
targets, χ2(1) = 39.41, p < .001. However, there was no sig-
nificant difference between the percent of participants writ-
ing arguments reflecting liberal moral values and the
percent writing arguments contradicting liberal moral val-
ues, χ2(1) = .84, p = .359.
Discussion
Studies 1 and 2 support our first hypothesis that political
advocates, in attempts to persuade those with rival political
Table 1. Frequencies in Which Participants in Studies 1 and 2 Wrote Moral Arguments That Appealed to Their Own Moral Values,
Appealed to The Targeted Group’s Moral Values, and Attacked The Targeted Group's Moral Values.
Study 1. Liberals instructed to write arguments in favor of same-sex marriage aimed at convincing conservatives who oppose same-sex
marriage.
Moral argument Fit With Conservative Morality Fit With Liberal Morality Attacked Conservative Morality
Yes 80 8 69 32
No 13 85 24 61
Study 2. Conservatives instructed to write arguments in favor of official English aimed at convincing liberals who oppose official English.
Moral argument Fit With Liberal Morality Fit With Conservative Morality Attacked Liberal Morality
Yes 66 7 59 12
No 18 77 25 72
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from

6 Personality and Social Psychology Bulletin
positions, typically compose arguments based on their own
moral values rather than those of the individuals targeted for
persuasion. Our results, however, do not distinguish whether
participants’ egocentric bias was due to them being incapable
of recognizing that their audience possesses different values,
their inability to recognize that appealing to those different
values would be persuasive, or potentially, their inability to
successfully write arguments that appeal to those different
values. Furthermore, our results also do not explore the pos-
sibility that participants were motivated to avoid using, and
tacitly endorsing, moral arguments they might themselves
disagree with—an issue that we discuss further in the general
discussion. Regardless of the underlying mechanism, such a
failure to tailor their arguments to their audience’s moral val-
ues, we believe, will result in these arguments being ineffec-
tive, when compared with arguments reframed to appeal to
the values of those supporting the opposing political posi-
tion. We test this reasoning in Studies 3 to 6.
Study 3
In Study 3, we present participants with arguments in favor
of universal health care framed in terms of fairness, the sort
of moral rhetoric typically invoked in support of national
health care (i.e., access to health care is a right), or in terms
of purity (i.e., sick people are disgusting). We then measure
participants’ attitudes toward national health care and the
Affordable Care Act (ObamaCare). If our more reframing
hypothesis is correct, we would expect that the purity-based
arguments will be more persuasive to conservative partici-
pants than the fairness-based argument.
Method
Participants. In all, 288 participants (150 male, 138 female)
were recruited from across the United States using Amazon
Mechanical Turk. Participants received a small payment for
their participation. Because Mechanical Turk participants
may take participation in studies less seriously than more
volunteer-based participants, we took precautions to ensure
that we excluded individuals who demonstrated that they
did not take participation seriously. We measured how long
participants spent on the webpage that presented the
persuasive moral message (see below). Based on speed-
reading trials by research assistants, we determined that
participants who spent less than 20 s on this page could not
have fully read the argument and did not truly take part in
the study (Johnson, 2005; Mason & Suri, 2012). We
removed these participants, leaving a total of 258 partici-
pants (139 male, 119 female).
Procedure. Participants completed a demographic question-
naire including a two-item composite measuring political
ideology (“When it comes to social policy, do you consider
yourself liberal or conservative?” and “When it comes to
economic policy, do you consider yourself liberal or conser-
vative?” (α = .79). Participants answered both items on a
scale ranging from 1 (very liberal) to 4 (moderate) to 7 (very
conservative; M = 3.21, SD = 1.51). They then learned that
they would read a short opinion article and answer some
questions about their impressions of it. Participants were ran-
domly assigned to read a short article in support of universal
health care framed either in terms of fairness (health cover-
age is a basic human right) or purity (uninsured people means
more unclean, infected, and diseased Americans). Both arti-
cles were modeled after newspaper opinion pieces and were
of similar length (approximately 350 words) and style. In
addition, to ensure that we successfully grounded the argu-
ments in each foundation, we utilized words from the Moral
Foundations Dictionary (Graham et al., 2009) that corre-
sponded with the argument’s specific moral foundation (e.g.,
Fairness: unfair, unjust; Purity: unclean, disgusting). Pilot
testing further confirmed that each article strongly repre-
sented the intended moral foundation (see online materials
for full text).
After reading their assigned article, participants com-
pleted a three-item questionnaire measuring their support
for universal health care (e.g., “I am in favor of universal
health care”), followed by a four-item questionnaire gaug-
ing endorsement of the Affordable Care Act (i.e.,
ObamaCare; for example, “The Affordable Care Act
(ObamaCare) was a mistake that should never have become
a law” [reverse-scored]). Participants answered these items
on a scale ranging from 1 (strongly disagree) to 5 (strongly
agree). The reliabilities for these two measures were high
(αs = .83 and .94, respectively), so we averaged each
together to form universal health care support and
ObamaCare support composites.
Results and Discussion
We ran multiple regression analyses entering messaging con-
dition, political ideology, and their interaction as predictors
of both universal health care and ObamaCare support sepa-
rately. For both dependent variables, there was no significant
effect of condition, bs < .09, ps > .47; a significant effect of
ideology, bs < −.37, ps < .001; and a significant interaction,
bs > .15, ps < .025. Simple slope analyses indicated that the
conservative participants (+1 SD) in the purity condition
demonstrated significantly greater support for ObamaCare,
b = .38, SE = .18, p = .032, 95% confidence interval (CI) =
[0.03, 0.72], and marginally more support for universal
health care, b = .26, SE = .14 p = .066, 95% CI = [−0.02,
0.54], than conservative participants in the fairness condi-
tion. There was no effect of condition for the liberal partici-
pants (−1 SD), bs < −.19, ps > .17 (Figure 1). Overall, these
results show that arguments in support of universal health
care framed in terms of purity concerns led conservatives to
be more supportive of universal health care, and specifically,
ObamaCare.
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from

Feinberg and Willer 7
Study 4
Study 4 extends our research in several ways. First, instead
of presenting reframed arguments in favor of a typically lib-
eral position that fit with conservative moral foundations, we
presented arguments in favor of a typically conservative
position that fit with a moral value endorsed more by liber-
als. Specifically, we presented participants with arguments in
favor of maintaining high levels of military spending framed
in terms of either fairness or authority and loyalty. In addi-
tion, we included questions that allowed us to capture
whether any movement in liberal attitudes was due to agree-
ment between the message and the targeted participants’
moral values—specifically, whether the message convinced
liberals that the military as an institution serves to increase
fairness in society. We also asked participants how persua-
sive and convincing, as well as how novel, they found the
argument to be. Including these questions allowed us to fur-
ther verify that the reframed arguments swayed participants
because they found them persuasive and not due to other
possibilities such as the article’s novelty.
Method
Participants. In all, 306 volunteers (88 male, 217 female, 1
did not indicate) were recruited from across the United States
via Craigslist.org. For participating, participants were
entered into a drawing for a US$50 gift certificate.
Procedure. As before, participants completed a demographic
questionnaire including the two-item measure of political
ideology (α = .72), with higher scores indicating greater con-
servatism (M = 3.20, SD = 1.32). Participants learned that
they would read an opinion article and be asked about their
impressions of it. Participants were then randomly assigned
to read a short article in favor of maintaining high levels of
military spending grounded in terms of fairness (through the
military, the disadvantaged can achieve equal standing and
overcome the challenges of poverty and inequality) or a
combination of loyalty and authority (the military unifies us
and ensures that the United States is the greatest nation in the
world). The articles were similar in length (approximately
260 words) and style, and employed words from the Moral
Foundations Dictionary (e.g., Fairness: equal, inequality;
Loyalty and Authority: nation, loyal, leader, oppose). Pilot
testing confirmed that each article best fit with the intended
moral foundation (see online materials for full text).
After reading their assigned argument, participants
answered two questions gauging how persuasive they found
the article to be: “How convincing was this article?” and
“How persuasive was this article?” (α = .91), and an item
measuring the novelty of the article (“How original was the
article’s argument?”). Participants responded to these items
on scales ranging from 1 (not at all) to 5 (extremely). Next,
participants completed a three-item questionnaire measuring
how much they viewed the military as an institution that pro-
motes fairness in society (e.g., “The military is an institution
that helps reduce inequality in the United States.”).
Participants responded to each item on a scale from 1
(strongly disagree) to 5 (strongly agree). Reliability among
these items was high (α = .81), so we combined them into a
military promotes fairness composite. Finally, participants
completed a three-item questionnaire that measured partici-
pants’ attitudes toward maintaining high levels of military
spending (e.g., “Cutting funding to the military would be a
mistake.”). Participants indicated how strongly they agreed
with each statement on a scale from 1 (strongly disagree) to
5 (strongly agree). Reliability across these items was high
(α = .91), so we averaged them to form a military spending
support composite.
Results and Discussion
Military spending support. A multiple regression analysis
entering experimental condition, political ideology, and their
interaction as predictors of military spending attitudes yielded
a significant effect of condition, b = .25, SE = .12, p = .038; a
significant effect of ideology, b = .34, SE = .05, p < .001; and
a significant interaction, b = −.21, SE = .09, p = .026. Simple
slope analyses found that liberal participants (−1 SD) in the
fairness condition scored significantly higher than the liberals
in the loyalty/authority condition, b = .52, SE = .17, p = .003,
95% CI = [0.18, 0.86]. Conservative participants (+1 SD) did
not differ due to experimental condition, b = −.02, SE = .17, p
= .910 (Figure 2).
Persuasiveness. We conducted a parallel multiple regression
analysis entering the persuasiveness composite as the depen-
dent variable. This analysis yielded a significant effect of
condition, b = .60, SE = .12, p < .001; a significant effect of
political ideology, b = .15, SE = .05, p < .001; and a signifi-
cant interaction, b = −.25, SE = .09, p < .01. Simple slope
analyses confirmed that liberal participants (−1 SD) in the
Figure 1. The interaction between political ideology and
message framing on support for the Affordable Care Act (i.e.,
ObamaCare; Study 3).
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from

8 Personality and Social Psychology Bulletin
fairness condition found the argument significantly more
persuasive, b = .93, SE = .17, p < .001, 95% CI = [0.60,
1.26]; whereas conservative participants (+1 SD) did not dif-
fer due to condition, b = .27, SE = .17, p = .108.
Moral-fit mediation. We next explored the possibility that
scores on the military promotes fairness composite would
explain why liberal participants were more supportive of
military spending when presented with the fairness message.
A multiple regression analysis entering experimental condi-
tion, political ideology, and their interaction as independent
variables and scores on the military promotes fairness com-
posite as the dependent variable yielded a main effect of con-
dition, b = .37, SE = .11, p < .001; a main effect of ideology,
b = .10, SE = .04, p = .012; and a significant interaction,
b = −.19, SE = .08, p = .016, finding that liberal participants
(−1 SD) in the fairness condition scored higher on the mili-
tary promotes fairness composite than liberals in the loyalty/
authority condition, b = .63, SE = .15, p < .001, 95% CI =
[0.33, 0.92]. There was no effect of condition for the conser-
vative participants (+1 SD), b = .12, SE = .15, p = .436.
We next conducted mediated moderation analyses
(Preacher & Hayes, 2008) to explore whether the military
promotes fairness composite explained the effect of condition
on the liberal participants’ attitudes toward military spending.
The 95% CI for the indirect effect did not include zero [−0.21,
−0.02] (Figure 3), suggesting that the fairness message per-
suaded liberals to be more supportive of military spending by
increasing their perception that the military promotes fair-
ness, thus increasing the apparent fit between the position and
their moral values. In addition, we conducted parallel medi-
ated moderation analyses to examine whether perceptions
that the fairness message was novel might also explain why
liberal participants in the fairness condition reported greater
support for military spending. We found that perceptions of
novelty was a significant mediator, but entering both novelty
and the military promotes fairness composite as simultaneous
mediators revealed that only the military promotes fairness
composite remained significant (for details, see supplemen-
tary materials).
These results complement and extend the findings of
Study 3. In line with our moral reframing hypothesis, when
arguments in favor of maintaining high levels of military
spending were presented as consistent with the moral value
of fairness, liberals found them to be more persuasive and
reported greater support for military spending than if they
read a more typical loyalty/authority-based message. These
results, along with those of Study 3, suggest that morally
reframing political messages can be effective in persuading
both liberals and conservatives. In addition, we found evi-
dence that the fairness message was persuasive to liberal par-
ticipants because it led them to view the military as promoting
fairness, a moral value generally held at high levels among
liberal Americans. These results provide support for our
hypothesis that moral reframing shifts individuals’ attitudes
by leading them to view an issue stance they would normally
oppose as more in line with their moral values.
Study 5
In Study 5, we test whether moral reframing can influence
conservatives’ attitudes regarding same-sex marriage. As
Study 1 suggests, typical arguments in favor of same-sex
marriage tend to be grounded in the more liberal moral foun-
dations, particularly fairness concerns that emphasize the
immorality of denying equal rights to individuals because of
their sexual orientation. In line with our moral reframing
hypothesis, we predict that grounding same-sex marriage
support in loyalty-based moral concerns will be more per-
suasive to conservatives, leading them to be more supportive
of legalizing same-sex marriage.
In addition, in Study 5 we include a neutral condition
where participants read about an unrelated, amoral topic. We
have claimed that the morally reframed messages influenced
the attitudes of conservatives (Study 3) and liberals (Study 4)
such that they became more supportive of a stance they
would traditionally oppose. Yet, possibly, rather than the
morally reframed message shifting attitudes, the more typi-
cal moral arguments prompted reactance in participants,
Figure 2. The interaction between political ideology and
message framing on support for military spending (Study 4).
Figure 3. Depiction of the mediating role participants’
perceptions of military spending as an issue of fairness has on the
relationship between the interaction and support for maintaining
high levels of military spending (Study 4).
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from

Feinberg and Willer 9
pushing them to be more opposed to the stance than they
would normally be. By comparing same-sex marriage atti-
tudes of conservatives in the morally reframed (loyalty-
based) condition with the attitudes of conservatives in the
neutral condition, we can assess whether the morally
reframed message actually increases conservative support
for same-sex marriage.
We also more closely explore the mechanism underlying
the effectiveness of moral reframing. Specifically, to test that
moral fit between the reframed argument and the targeted
audience’s values drives the persuasive impact of moral
reframing, we directly ask participants to indicate how much
the arguments we present to them resonate with their moral
values.
Participants
A total of 261(89 male, 172 female) participants were
recruited from a volunteer-based university alumni pool. For
their participation, participants were entered into a raffle for
a US$50 gift certificate.
Procedure
As before, participants completed a demographic question-
naire including the two-item measure of political ideology
(M = 3.49, SD = 1.45, α = .79). Participants then learned that
they would be presented with an opinion article to read and
afterward answer questions about their impressions of it.
Participants were randomly assigned to read a short article
(approximately 280 words) in support of same-sex marriage
framed in terms of either fairness (all citizens should be
treated equally) or loyalty (same-sex couples are proud and
patriotic Americans), or to read a neutral article describing the
history of skiing. To help ensure that we successfully grounded
the fairness and loyalty arguments in their respective founda-
tion, we utilized words from the Moral Foundations Dictionary
(Graham et al., 2009) that corresponded with the argument’s
specific moral foundation (e.g., Fairness: fair, equality;
Loyalty: patriotic, community). Pilot testing confirmed that
each article strongly represented the intended moral founda-
tion (see online materials for full text).
After reading their assigned argument, participants
answered the following measure of moral fit: “To what
extent do you feel the article’s message resonates with your
values?” on a scale ranging from 1 (not at all) to 5 (a whole
lot). Then participants completed a six-item measure of
same-sex marriage attitudes (e.g., “The government should
allow same-sex couples to marry legally”), answering each
item on a scale from 1 (strongly disagree) to 5 (strongly
agree; α = .93).
Results and Discussion
Same-sex marriage attitudes. We first dummy-coded whether
participants were in the neutral condition or the fairness-
based argument condition. Then we entered these two
dummy-coded variables, political ideology (z scored), and
the interaction of each dummy-coded variable and political
ideology as predictors of participants’ same-sex marriage
attitudes. This analysis yielded a non-significant effect of the
neutral condition, b = −.16, SE = .16, p = .338; a non-signif-
icant effect of the fairness condition, b = −.07, SE = .15, p =
.631; a significant effect of political ideology, b = −.29, SE =
.11, p = .011; a significant Neutral-Dummy × Political Ideol-
ogy interaction, b = −.36, SE = .70, p = .032; and a significant
Fairness-Dummy × Political Ideology interaction, b = −.51,
SE = .15, p = .001. We next conducted simple slope analyses
to compare the effect of each experimental condition on con-
servative participants’ same-sex marriage attitudes. We
found that conservative participants (+1 SD) indicated more
same-sex marriage support in the loyalty condition than in
the fairness condition, b = −.58, SE = .20, p = .004, 95% CI
= [−0.98, −0.18]. This result indicates that morally reframing
the issue of same-sex marriage as an issue of loyalty was
more persuasive to conservatives than the more typical
fairness-based argument, a result that conceptually replicates
the findings of both, Studies 3 and 4.
We also found that the conservatives (+1 SD) in the loy-
alty condition demonstrated greater same-sex marriage sup-
port than conservatives in the neutral condition, b = −.52,
SE = .25, p = .039, 95% CI = [−1.02, −0.02]. This result helps
verify that our moral reframing findings are not solely due to
the targeted group (in this case, conservatives) becoming
more opposed in response to the more typical moral argu-
ment (in this case, a fairness-based argument). In addition, a
third simple slope analysis demonstrated that there was no
significant difference between the conservatives (+1 SD) in
the neutral and fairness conditions, b = .07, SE = .24, p = .786
(Figure 4).
A parallel set of analyses looking at the effect of experi-
mental condition on liberal participants (−1 SD) found no
differences between the neutral and loyalty conditions,
b = .21, SE = .22, p = .343, or the neutral and fairness
Figure 4. The interaction between political ideology and
message framing on participants’ support for same-sex marriage
(Study 5).
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from

10 Personality and Social Psychology Bulletin
conditions, b = −.23, SE = .22, p = .293. However, we did
find that the liberal participants in the fairness condition
reported greater same-sex marriage support than liberal par-
ticipants in the loyalty condition, b = .44, SE = .22, p = .049,
95% CI = [0.00, 0.88].
Moral fit. Finally, we explored whether scores on the mea-
sure of moral fit mediated the effect of condition (comparing
the fairness and loyalty conditions, and comparing the neu-
tral and loyalty conditions) that we found among conserva-
tive participants. First, we conducted the same multiple
regression analysis as before, but entered scores on the
moral-fit measure as the dependent variable. This analysis
yielded a significant effect of the neutral condition,
b = −1.78, SE = .17, p < .001; a non-significant effect of the
fairness condition, b = .03, SE = .16, p = .838; and a signifi-
cant effect of political ideology, b = −.31, SE = .12, p = .008.
We also found a significant Neutral-Dummy × Political Ide-
ology interaction, b = .39, SE = .17, p = .025, and a signifi-
cant Fairness-Dummy × Political Ideology interaction, b =
−.51, SE = .15, p = .001. Simple slope analyses demonstrated
that conservatives (+1 SD) in the loyalty condition indicated
that the argument fit with their moral values more than the
conservatives in either the neutral, b = −1.39, SE = .26, p <
.001, 95% CI = [−1.91, −0.87], or fairness conditions, b =
−.48, SE = .21, p = .024, 95% CI = [−0.90, −0.06]. Next, we
conducted mediated moderation analyses (Preacher & Hayes,
2008) to test whether participants’ moral-fit scores explained
the difference between conservatives’ same-sex marriage
attitudes when in the fairness versus loyalty conditions or
neutral versus loyalty conditions. Bootstrap analyses
revealed that for both comparisons the 95% CI did not
include 0 (loyalty vs. fairness: [0.11, 0.55]; loyalty vs. neu-
tral: [0.01, 0.21]; Figure 5), suggesting that conservatives in
the loyalty condition were more favorable toward same-sex
marriage than their counterparts in the other two conditions
because the loyalty message resonated more with their
values.
Overall, Study 5 provides further support for our moral
reframing hypothesis within the context of same-sex marriage.
Importantly, in Study 5, we found that conservatives in the
loyalty moral argument condition were more supportive of
same-sex marriage than conservatives in the neutral condition,
and that conservative participants in the neutral and fairness
condition demonstrated no significant differences in their
same-sex marriage support. Together, these results suggest
that the moral reframing effects we have found throughout our
studies are not due to the targeted group taking a stronger
oppositional stance when they are presented with a typical
moral argument. In addition, Study 5’s moral-fit mediation
finding more directly establishes that the fit between the
reframed argument and the targeted audience’s values drives
the persuasive impact of moral reframing
Study 6
In Study 6, we present participants with an argument in favor
of making English the official language of the United States,
or a neutral message. As Study 2 suggests, supporters of
“official English” typically frame their support in terms of
group loyalty, arguing that the English language is some-
thing that unifies Americans and is a fundamental part of a
larger cultural assimilation process. We reframed this posi-
tion in terms of the more liberal moral value of fairness,
emphasizing how making English the official language of
the United States would compel immigrants to learn the lan-
guage, leading them in turn to face better job prospects and
less discrimination.
In Study 6, we also sought to address some alternative
explanations for our earlier findings. In Studies 3 to 5, we
used participants’ reported political ideology as a proxy for
the moral values they endorse, assuming that liberals endorse
harm and fairness at higher levels than conservatives and
conservatives endorse loyalty, authority, and purity at higher
levels than liberals. Past research has consistently estab-
lished these moral value differences between liberals and
conservatives (Graham et al., 2009; Jost et al., 2003; Lakoff,
2002). However, to ensure that our results were indeed due to
such differences, in Study 5 we directly measure partici-
pants’ endorsement of the moral value of fairness. By doing
so, we can more precisely test whether a political argument
based on fairness moral principles is more persuasive to
those who hold this moral value most strongly.
Method
Participants. A total of 290 participants recruited from Ama-
zon Mechanical Turk completed the online survey (167
male, 122 female, and 1 did not indicate). To ensure we gath-
ered high quality data from our online source, we included
two attention checks (Downs, Holbrook, Sheng, & Cranor,
2010; Moral Foundations Questionnaire [MFQ], 2012).
Thirteen participants did not successfully complete these
items (described below), leaving a total of 278 participants
(163 male, 114 female, 1 did not indicate). Participants
Figure 5. Depiction of the mediating role of message fit on the
relationship between the experimental Condition (fairness vs.
loyalty) × Political Ideology interaction and participants’ support
for same-sex marriage (Study 4).
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from

Feinberg and Willer 11
received modest compensation for participation.
Procedure. Participants completed the same demographic
questionnaire as before, including the political ideology
composite (M = 3.47, SD = 1.51, α = .79). Participants then
completed the 30-item MFQ (Graham et al., 2011; αHarm =
.74, αFairness = .66, αLoyalty = .74, αAuthority = .74, αPurity = .86).
Participants indicating that being good at math was relevant
to determining whether someone was moral or not (n = 8), a
filter item from the MFQ, were removed from all analyses.
Furthermore, instructions explained that a recent theory pro-
poses that there are five moral foundations, listing each foun-
dation along with a brief example of reasoning based on
each—for example, “Fairness (example: it is immoral to not
treat others fairly/justly)”—to ensure participants understood
the distinction between domains. Participants indicated how
they would rank the importance of the five foundations (from
1 to 5) when determining whether an act is moral or immoral.
Participants then learned that they would read a brief
opinion article and answer questions about their impressions
of it afterward. Participants were randomly assigned to either
the neutral or the fairness condition. Participants in the neu-
tral condition read a short article about the history of wearing
ties containing no moral content. Participants in the fairness
condition read a persuasive message in favor of making
English the official language of the United States (both mes-
sages were approximately 350 words). Using fairness-based
words from the Moral Foundations Dictionary (e.g., fair,
equal), this message argued that making English the official
language would lead to more fair outcomes for immigrants
and help them avoid discrimination. Pilot testing confirmed
that the article strongly represented the fairness moral foun-
dation (see online materials for full text).
After reading their assigned article, participants com-
pleted a four-item support for official English questionnaire
measuring support for making English the official language
of the United States (e.g., “I am in favor of making English
the official language of the United States”; α = .95).
Participants indicated the extent to which they agreed with
each of these items on a scale from 1 (strongly disagree) to 5
(strongly agree). In addition, embedded in this scale was a
“check” item that directed participants to select the “agree”
option for that item to verify they were paying attention and
taking the task seriously (i.e., not randomly clicking
responses). Six participants failed this attention check, two
of whom also failed the “math” check item described above,
and were eliminated from analysis.
Results and Discussion
Political ideology. We examined whether the fairness message
impacted individuals differently depending on their political
ideology. A multiple regression analysis entering condition,
political ideology, and their interaction as predictors of sup-
port for official English yielded a non-significant effect of
condition, b = .20, SE = .12, p = .113; a significant effect of
political ideology, b = .32, SE = .04, p < .001; and a margin-
ally significant interaction b = −.14, SE = .08, p = .082. Sim-
ple slope analyses showed that liberal participants (−1 SD) in
the fairness condition were more supportive of official Eng-
lish than liberal participants in the control, b = .42, SE = .18,
p = .020, 95% CI = [0.07, 0.77]. There was no significant
effect of condition for the conservative participants (+1 SD),
b = −.02, SE = .18, p = .909 (Figure 6). Thus, these results
conceptually replicate the results of Studies 3 to 5.
Fairness endorsement. We measured endorsement of the fair-
ness foundation in two ways. First, we calculated each par-
ticipant’s average score on the fairness scale and subtracted
out the average of the four remaining moral domains. Doing
so allowed us to capture each participant’s specific endorse-
ment of the fairness foundation, while accounting for the ten-
dency for some participants to report high levels of
endorsement across all the foundations, labeling this mea-
sure fairness score—continuous. Second, we used partici-
pants’ ranking of the fairness foundation, relative to the other
foundations, labeling this measure fairness score—rank.
Next, we examined the correlations between fairness
score—continuous, fairness score—rank, and political ideol-
ogy. Fairness score—continuous correlated −.53, p < .001,
and fairness score—rank correlated −.24, p = .001, with
political ideology. Thus, in line with past research (Graham
et al., 2009), the more conservative participants were, the
less strongly they endorsed the fairness moral foundation
(see also Jost et al., 2003).
Figure 6. The interaction between political ideology and
message framing (top panel) and between fairness morality and
message framing (bottom panel) on support for making English
the official language of the United States (Study 6).
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from

12 Personality and Social Psychology Bulletin
Next, we tested Study 5’s central hypothesis that partici-
pants who strongly endorse the fairness moral foundation
will be most influenced by the fairness-based argument. A
multiple regression analysis entering experimental condi-
tion, fairness score—continuous, and the interaction of the
two as predictors of support for official English yielded a
non-significant effect of experimental condition, b = .14, SE
= .13, p = .262, and a significant effect of fairness score—
continuous, b = −.53, SE = .07, p < .001, revealing that the
more individuals endorsed the fairness foundation, the less
likely they were to support making English the official lan-
guage. The analysis also yielded a significant interaction, b =
.39, SE = .15, p = .009. Simple slope analyses revealed that
among participants who endorsed the fairness foundation at
high levels (+1 SD), there was a significant effect of experi-
mental condition, b = .48, SE = .18, p = .008, 95% CI = [0.12,
0.84], such that those in the fairness message condition
endorsed official English more than those in the control con-
dition. We found no effect of condition among those who
endorsed the fairness foundation at low levels (−1 SD), b =
−.20, SE = .18, p = .276 (Figure 6).2
As a second test, we conducted a parallel regression anal-
ysis entering fairness score—rank as the measure of endorse-
ment of the fairness moral foundation. This analysis yielded
a non-significant effect of experimental condition, b = .12,
SE = .14, p = .391; a significant effect of fairness score—
rank, b = −.26, SE = .08, p = .002; and a significant interac-
tion, b = .33, SE = .17, p = .049. Simple slope analyses
revealed that for participants who ranked the fairness moral
foundation high (+1 SD), there was a significant difference
due to experimental condition, b = .40, SE = .20, p = .040,
95% CI = [0.02, 0.79], such that those in the fairness condi-
tion endorsed making English the official language more
than their counterparts in the control condition. Among par-
ticipants who ranked the fairness moral foundation low
(−1 SD), there was no difference due to experimental condi-
tion, b = −.17, SE = .20, p = .406.
Results of this study demonstrate that a fairness-based
political argument was uniquely persuasive not only to lib-
eral participants but also, in parallel analyses, to participants
who endorsed the value of fairness at high levels. Such
results offer discerning evidence for our theoretical reason-
ing that reframed messages influence recipients through
agreement with their deeply held moral values.
General Discussion
Liberal and conservative advocates often use moral rhetoric
in attempts to persuade their rivals. We hypothesized that
because liberals and conservatives face a moral empathy
gap, even when targeting the other side for persuasion, both
groups will typically frame their arguments in a way that fits
with their own sense of morality rather than the morality of
the intended audience. As a result, their moral arguments will
be less effective than arguments reframed to appeal to the
values of those holding the opposing political position.
Specifically, we reasoned that people would support an
opposing political stance more when messages illustrated
consistency between the stance and their most cherished
moral values.
Results of six studies supported our hypotheses. We
found that both liberals and conservatives composed persua-
sive messages that reflected their own moral values, not val-
ues unique to those who typically would oppose the political
stance (Studies 1 and 2). Furthermore, these moral messages
framed in a manner consistent with the moral values of
those already supporting the political stance were less per-
suasive than moral arguments reframed to appeal to the val-
ues of the intended audience—those who typically oppose
the political position that the messenger is arguing in favor
of (Studies 3-6).
These findings offer important theoretical insights. For
instance, our results support the claim that moral values can
causally influence political attitudes. Although abundant cor-
relational research links morality and politics, far less work
has established this causal link, leading critics to suggest that
values are only tools for self-justification and sense-making
that have few causal effects in the political domain. This is
not to say, however, that morality is necessarily the primary
cause of the correlation between morality and politics, and it
is likely that the relationship between these concepts is
bidirectional.
In addition, our results support the efficacy of moral rhet-
oric and framing. Where political strategists and theorists
have emphasized the importance of moral content in politi-
cal advocacy (Atran & Axelrod, 2008; Lakoff, 2002; Luntz,
2007), little systematic, academic research has tested their
claims. Indeed, evidence from the field is ambivalent, as
both successful (e.g., the civil rights movement) and unsuc-
cessful (e.g., universal background checks for guns) politi-
cal efforts often feature moral rhetoric prominently. Our
results provide nuanced support for claims that political
arguments couched in moral terms can be more persuasive,
demonstrating that the effectiveness of moral framing
depends critically on fitting the values underlying a message
to those held by the targeted audience.
Our results align well with existing persuasion research
conducted outside the moral domain. As with this past
research, we found that matching messages were particularly
effective in persuading the targeted group, and that perceived
fit drove these persuasion effects (e.g., Cacioppo et al., 1982;
Cesario et al., 2004; Petty & Wegener, 1998). That said, in
the present research we did not explore the cognitive and
affective processes that underlie the moral-fit effects we
observe. Previous research on message matching suggests
that the fit between the message and an individual’s charac-
teristics engenders a sense that the argument feels right, and
these feelings transfer to judgments of the message’s posi-
tion and the larger issue in question, and facilitate cognitive
processing regarding the topic (Avnet & Higgins, 2006;
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from

Feinberg and Willer 13
Cesario et al., 2004; Schwarz & Clore, 1983). Given that the
influence of morality on individuals’ judgments occurs
largely at an intuitive and emotional level (Haidt, 2001),
morally reframed messages should be particularly likely to
trigger positive intuitions about the message—so much so
that moral reframing can lead liberals and conservatives to
increase their support for positions they typically would not
endorse. Examining further the cognitive and affective pro-
cesses underlying the moral reframing effects we document
would be a worthwhile avenue for future research.
The results of our studies suggest that arguments appeal-
ing to the messenger’s values rather than the audience’s val-
ues were not only unpersuasive to the target audience but
also did not impact the attitudes of those already in favor of
the political position being argued for. For instance, conser-
vatives presented with a military spending argument
grounded in loyalty and authority concerns did not show any
greater support for military spending than conservatives pre-
sented with an argument grounded in fairness concerns. One
reason for these non-effects could be that the two arguments
were equally persuasive for a group already supporting the
position. However, in Study 5, we included a neutral condi-
tion and found no difference in liberal support for same-sex
marriage between the neutral, the argument grounded in fair-
ness concerns, and the reframed argument grounded in loy-
alty concerns, suggesting that neither moral argument in
favor of same-sex marriage had an effect on liberals’ atti-
tudes. These null results could be the product of a ceiling
effect because liberal participants’ support for same-sex mar-
riage was close to the top of the 5-point scale (M = 4.54,
SD = 0.67). Still another possibility is that this type of fair-
ness-based argument was not persuasive because it was not
novel. However, this would imply that a highly novel mes-
sage that appealed to liberal moral values would be persua-
sive to liberals, which may seem at odds with Study 4’s
results showing that increased liberal support for military
spending after a morally reframed argument was not driven
by the novelty of this argument. Still another possibility is
that novelty impacts the persuasiveness of pro-attitudinal
messages grounded in relevant moral concerns, but not of
counter-attitudinal messages grounded in relevant moral
concerns. Overall, we believe, the most parsimonious
account for these non-effects is that there was a ceiling effect,
but hope future research might explore this issue further.
Why did so few participants in Studies 1 and 2 tailor their
arguments to appeal to the moral values of their target audi-
ence? It could be that people are unable to recognize that
ideologically different others endorse moral values that are
different from their own, and thus find different moral argu-
ments convincing. Or, it could be people understand this, but
they do not want to advance moral arguments they disagree
with. As an initial investigation of this question, we had 171
liberal participants read both the same-sex marriage argu-
ments used in Study 5, and 105 conservative participants
read both of the military spending arguments used in Study
4. We asked these participants “Which article do you think
would be more effective at persuading political conserva-
tives [liberals] who oppose same-sex marriage [military
spending] to be more supportive of same-sex marriage [mili-
tary spending]?” and “Which article would you select to
present to conservatives [liberals] who oppose same-sex
marriage [military spending] if you wanted to persuade them
to be more in favor of same-sex marriage [military spend-
ing]?”. We found that a majority of participants (64% of lib-
erals, 85% of conservatives) identified the morally reframed
argument as the more persuasive one. Considering a large
majority of participants in Studies 1 and 2 did not write mor-
ally reframed arguments, it appears most individuals do not
spontaneously make morally reframed arguments, but can
recognize the persuasive value of such arguments. In addi-
tion, we found that among those who reported that the mor-
ally reframed argument would be more persuasive, a small
minority (20% of liberals, 6% of conservatives) indicated
they would not present the reframed argument. Thus, a small
portion of participants identified the argument that was more
persuasive, but indicated they would not send it, perhaps
because it was articulated in terms of moral values they did
not endorse.
These results, coupled with those found in Studies 1 and
2, suggest that although few individuals spontaneously con-
structed morally reframed arguments, most were able to rec-
ognize the persuasive value of morally reframed messages
when presented with them. This disjuncture is likely attribut-
able to both capacity- and motivated-based factors. Regarding
the former, it is more difficult to craft a highly persuasive
argument than it is to recognize one. Regarding the latter, a
minority of participants were not motivated to make reframed
arguments, though they viewed them as likely to be more
persuasive. We hope future research might delve deeper into
the reasons why advocates typically employ egocentrically
biased arguments when trying to win over those holding an
opposing stance.
Future research could also investigate whether these
dynamics occur in non-moral contexts. For instance, will
highly affect-oriented managers fail to recognize the useful-
ness of tailoring directives to employees in cognitively ori-
ented ways? And, do they do so because they are unable or
unwilling to communicate more persuasively? We believe
the failure of individuals to spontaneously craft arguments
that resonate with a very different target audience will likely
occur in many other contexts as well, though the effects may
be weaker in non-moral settings where individuals may have
greater capacity to imagine another’s perspective and are less
motivated to be true to the reasons for their own positions.
While we found that the morally reframed arguments we
tested in Studies 3 to 6 were persuasive to the target audi-
ence, in crafting those arguments, we relied upon insights
from research on the foundations of liberal and conserva-
tive morality—research lay advocates would be largely
unaware of. Thus, it remains unknown if lay people can
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from

14 Personality and Social Psychology Bulletin
craft persuasive, morally reframed arguments. As an initial
investigation of this question, we randomly presented a set
of liberal participants (n = 364) with one of the morally
reframed arguments in favor of official English that conser-
vative participants wrote in Study 2 or one of the non-
reframed moral arguments participants wrote in that study.
Likewise, we presented a set of conservative participants
(n = 339) with one of the morally reframed arguments in
favor of same-sex marriage that liberal participants wrote
in Study 1 or one of the non-reframed arguments that par-
ticipants wrote. We found that the morally reframed argu-
ments conservatives wrote in favor of official English were
more persuasive to the liberal targets than the other mes-
sages, and this difference was mediated by the extent to
which liberal participants reported that the message reso-
nated with their moral values more. However, we found
that the two types of messages liberals composed did not
differentially affect conservative attitudes toward same-sex
marriage (see supplementary materials for a complete
description of study methods and results). These results
offer some evidence that lay people, in this case conserva-
tive participants, can construct effective, morally reframed
arguments. Future research should explore whether direct-
ing individuals to strategically reframe their moral argu-
ments might increase the number of people able to compose
such reframed arguments.
It should be noted that although our research primarily
utilized MFT’s five moral foundations as a framework for
testing our hypotheses, moral reframing is not necessarily
tied to that framework. It is likely that other moral typologies
(e.g., Lakoff, 2002; Schwartz & Bilsky, 1987; Shweder,
Much, Mahapatra, & Park, 1997) could be used to morally
reframe political, and other, messages in terms that would
also be persuasive. In fact, a more comprehensive integration
of the different moral frameworks could result in more con-
sistently effective, morally reframed messages.
Although we found significant effects across our studies,
it is important also to scrutinize the size of those effects.
Figure 7 gives mean levels of support among the targeted
ideological groups in the treatment and control conditions of
Studies 3 to 6. Although many of our effect sizes appear
moderate, we would argue that these differences between
conditions are large in the context of political attitudes
research, which has found that views on morally charged
political issues are uniquely strong and resistant to influence.
A critic could reasonably note that we did not study the
effects of the reframed messages on whether attitude change
endured over time, or the efficacy of reframing effects in the
face of counter-arguments. These are all important avenues
for future research, though it would likely be necessary to
strengthen the reframed messages (perhaps through repeated
presentation across time), as our treatments in this article
were quite minimal.
Although our results demonstrate the efficacy of framing
moral arguments in terms that appeal to the values of those
on the other side of the political spectrum, the method could
have a variety of unintended consequences. Appealing to the
other side’s values might reinforce or validate those values
(e.g., Lakoff, 2002). For instance, if liberals reframe support
for same-sex marriage in terms of the more conservative
moral principles, their arguments may legitimize those prin-
ciples, facilitating their use to support conservative positions
Figure 7. Depiction of key results from Studies 3 through 6 demonstrating how moral reframing affected political attitudes of the
targeted group.
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from

Feinberg and Willer 15
in the future. Relatedly, advocates might be reluctant to use
moral reframing because they feel doing so compromises
their own beliefs, a position apparently taken by a small
number of participants in the above auxiliary study.
Furthermore, moral reframing could lead advocates to assim-
ilate to, or soften in their views of values they previously
opposed.
Finally, our theorizing and results may seem somewhat
contradictory to claims that, conservatives endorse all five
moral foundations equally whereas liberals only endorse the
harm and fairness foundations. Although descriptive results
for the MFQ suggest conservatives do endorse the five foun-
dations equally (Graham et al., 2011; Haidt, 2012), impor-
tantly, the questionnaire was not designed or scaled for
intra-personal comparisons across the moral domains, but
rather for inter-personal comparisons. Furthermore, if con-
servatives endorsed all five foundations equally, we should
have found that they were less likely to suffer from the moral
empathy gap, and any morally grounded argument should
have persuaded them. But we did not find those effects.
Moreover, results from a large political psychology literature
challenge the notion that conservatives endorse the fairness
and care foundations as much as they do the other founda-
tions. Indeed much of this literature suggests conservatism is
based fundamentally on the acceptance and legitimation of
inequality, and the willingness to prioritize the ingroup’s
goals over the welfare of those in the outgroup (e.g., Caprara
et al., 2006; Jost et al., 2003; Lakoff, 2002; Pratto, Sidanius,
Stallworth, & Malle, 1994). If conservatives endorsed the
five moral foundations equally, we would expect them to
take an ambivalent stance on many morally polarizing issues
like same-sex marriage and immigration, yet they do not.
Overall, even if conservatives advocate strongly for things
like fairness, equality, and social justice on paper, their polit-
ical stances suggest they are also willing to discount these
values in forming their political attitudes when other moral
values, such as loyalty, authority, and purity are at stake (e.g.,
conservative attitudes toward war, tax policy, and same-sex
marriage).
Conclusion
Although fundamental to social order and group solidarity
(Haidt, 2012), a darker side of morality is revealed in the
domain of politics, where it often divides those who do not
share the same convictions (Skitka & Morgan, 2009).
Morality contributes to political polarization because moral
convictions lead individuals to take absolutist stances and
refuse to compromise (Ditto & Koleva, 2011; Marietta,
2008; Skitka et al., 2005). Recognizing morality’s influence
on political attitudes, our research presents a means for polit-
ical persuasion that, rather than challenging one’s moral val-
ues, incorporates them into the argument. As a result,
individuals see value in an opposing stance, reducing the
attitudinal gap between the two sides. This technique not
only substantiates the power of morality to shape political
thought but also presents a potential means for political
coalition formation.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect
to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, author-
ship, and/or publication of this article.
Notes
1. Making English the official language of the United States is
supported more among conservatives than liberals (Rasmussen
Reports, 2010), and was added to the official Republican plat-
form in 2012.
2. We also tested the effects of a full regression model that
included political ideology, fairness score, experimental con-
dition, and both Political Ideology × Condition and Fairness
Score × Condition interaction terms. The two interaction terms
for this model were both non-significant, likely due to issues
of multicollinearity (the Variance Inflation Factor for each is
greater than 5). Even so, the Fairness Score × Condition interac-
tion appeared to be a more robust predictor, β = .18, p = .132,
relative to the Political Ideology × Condition interaction, β =
−.099, p = .570.
Supplemental Material
The online supplemental material is available at http://pspb.sage-
pub.com/supplemental.
References
Aquino, K., & Reed, A., II. (2002). The self-importance of moral
identity. Journal of Personality and Social Psychology, 83,
1423-1440.
Atran, S., & Axelrod, R. (2008). Reframing sacred values.
Negotiation Journal, 24, 221-246.
Avnet, T., & Higgins, E. T. (2006). How regulatory fit affects
value in consumer choices and opinions. Journal of Marketing
Research, 43, 1-10.
Cacioppo, J. T., Petty, R. E., & Sidera, J. A. (1982). The effects of a
salient self-schema on the evaluation of proattitudinal editori-
als: Top-down versus bottom-up message processing. Journal
of Experimental Social Psychology, 18, 324-338.
Caprara, G. V., Schwartz, S., Capanna, C., Vecchione, M., &
Barbaranelli, C. (2006). Personality and politics: Values, traits,
and political choice. Political Psychology, 27, 1-28.
Cesario, J., Grant, H., & Higgins, E. T. (2004). Regulatory fit
and persuasion: Transfer from “feeling right.” Journal of
Personality and Social Psychology, 86, 388-404.
Clary, E. G., Snyder, M., Ridge, R. D., Miene, P. K., & Haugen,
J. A. (1994). Matching messages to motives in persuasion: A
functional approach to promoting Volunteerism1. Journal of
Applied Social Psychology, 24, 1129-1146.
Day, M. V., Fiske, S. T., Downing, E. L., & Trail, T. E. (2014).
Shifting liberal and conservative attitudes using moral
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from

16 Personality and Social Psychology Bulletin
foundations theory. Personality and Social Psychology
Bulletin, 40(12), 1559-1573.
Ditto, P. H., & Koleva, S. P. (2011). Moral empathy gaps and the
American culture war. Emotion Review, 3, 331-332.
Downs, J. S., Holbrook, M. B., Sheng, S., & Cranor, L. F.
(2010). Are your participants gaming the system? Screening
Mechanical Turk workers. In Proceedings of the 28th inter-
national conference on Human Factors in Computing Systems
(pp. 2399-2402). New York: ACM Press.
Edwards, K. (1990). The interplay of affect and cognition in atti-
tude formation and change. Journal of Personality and Social
Psychology, 59(2), 202.
Fabrigar, L. R., & Petty, R. E. (1999). The role of the affective
and cognitive bases of attitudes in susceptibility to affectively
and cognitively based persuasion. Personality and Social
Psychology Bulletin, 25, 363-381.
Feinberg, M., & Willer, R. (2013). The moral roots of environmen-
tal attitudes. Psychological Science, 24, 56-62.
Feinberg, M., Willer, R., Antonenko, O., & John, O. P. (2012).
Liberating reason from the passions: Overriding intuitionist
moral judgments through emotion reappraisal. Psychological
Science, 23, 788-795.
Graham, J., Haidt, J., & Nosek, B. A. (2009). Liberals and conser-
vatives rely on different sets of moral foundations. Journal of
Personality and Social Psychology, 96, 1029-1046.
Graham, J., Nosek, B. A., Haidt, J., Iyer, R., Koleva, S., & Ditto, P.
H. (2011). Mapping the moral domain. Journal of Personality
and Social Psychology, 101, 366-385.
Haddock, G., Maio, G. R., Arnold, K., & Huskinson, T. (2008).
Should persuasion be affective or cognitive? The moderating
effects of need for affect and need for cognition. Personality
and Social Psychology Bulletin, 34, 769-778.
Haidt, J. (2001). The emotional dog and its rational tail: A social
intuitionist approach to moral judgment. Psychological Review,
108, 814-834.
Haidt, J. (2007). The new synthesis in moral psychology. Science,
316(5827), 998-1002.
Haidt, J. (2012). The righteous mind: Why good people are divided
by politics and religion. New York: Pantheon.
Haidt, J., & Hersh, M. A. (2001). Sexual morality: The cultures
and emotions of conservatives and liberals. Journal of Applied
Social Psychology, 31, 191-221.
Haidt, J., & Joseph, C. (2004). Intuitive ethics: How innately pre-
pared intuitions generate culturally variable virtues. Daedalus,
133(4), 55-66.
Hofmann, W., Wisneski, D. C., Brandt, M. J., & Skitka, L. J. (2014).
Morality in everyday life. Science, 345, 1340-1343.
Janoff-Bulman, R. (2009). To provide or protect: Motivational
bases of political liberalism and conservatism. Psychological
Inquiry, 20, 120-128.
Janoff-Bulman, R., Sheikh, S., & Baldacci, K. G. (2008). Mapping
moral motives: Approach, avoidance, and political orientation.
Journal of Experimental Social Psychology, 44, 1091-1099.
Johnson, J. A. (2005). Ascertaining the validity of individual pro-
tocols from web-based personality inventories. Journal of
Research in Personality, 39, 103-129.
Jost, J. T., Glaser, J., Kruglanski, A. W., & Sulloway, F. J.
(2003). Political conservatism as motivated social cognition.
Psychological Bulletin, 129, 339-375.
Kidwell, B., Farmer, A., & Hardesty, D. M. (2013). Getting liberals
and conservatives to go green: Political ideology and congruent
appeals. Journal of Consumer Research, 40(2), 350-367.
Lakoff, G. (2002). Moral politics: How liberals and conservatives
think. Chicago, IL: University of Chicago Press.
Lavine, H., & Snyder, M. (1996). Cognitive processing and the
functional matching effect in persuasion: The mediating
role of subjective perceptions of message quality. Journal of
Experimental Social Psychology, 32, 580-604.
Lee, A. Y., & Aaker, J. L. (2004). Bringing the frame into focus:
The influence of regulatory fit on processing fluency and per-
suasion. Journal of Personality and Social Psychology, 86,
205-218.
Luntz, F. I. (2007). Words that work: It’s not what you say, it’s what
people hear. New York, NY: Hyperion.
Marietta, M. (2008). From my cold, dead hands: Democratic con-
sequences of sacred rhetoric. Journal of Politics, 70, 767-779.
Mason, W., & Suri, S. (2012). Conducting behavioral research on
Amazon’s Mechanical Turk. Behavior Research Methods, 44,
1-23.
McAdams, D. P., Albaugh, M., Farber, E., Daniels, J., Logan, R.
L., & Olson, B. (2008). Family metaphors and moral intuitions:
How conservatives and liberals narrate their lives. Journal of
Personality and Social Psychology, 95, 978-990.
Moral foundations questionnaire: 30-item full version item key
(2008, July). Retrieved from http://www.moralfoundations.
org/questionnaires
Morgan, G. S., Skitka, L. J., & Wisneski, D. C. (2010). Moral and
religious convictions and intentions to vote in the 2008 presi-
dential election. Analyses of Social Issues and Public Policy,
10, 307-320.
Petty, R. E., & Wegener, D. T. (1998). Matching versus mismatch-
ing attitude functions: Implications for scrutiny of persuasive
messages. Personality and Social Psychology Bulletin, 24,
227-240.
Pew Research Center. (2013). Many more democrats than repub-
licans say protecting environment a top priority. Retrieved
from http://www.pewresearch.org/daily-number/many-more-
democrats-than-republicans-say-protecting-environment-a-
top-priority/
Pew Research Center. (2014). Political polarization in the
American public. Retrieved from http://www.people-press.
org/2014/06/12/political-polarization-in-the-american-public/
Pratto, F., Sidanius, J., Stallworth, L. M., & Malle, B. F. (1994).
Social dominance orientation: A personality variable predict-
ing social and political attitudes. Journal of Personality and
Social Psychology, 67, 741-763.
Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling
strategies for assessing and comparing indirect effects in multi-
ple mediator models. Behavior Research Methods, 40, 879-891.
Rasmussen Reports. (2010). 87% say English should be U.S. offi-
cial language. Retrieved from http://www.rasmussenreports.
com/public_content/politics/general_politics/may_2010/87_
say_english_should_be_u_s_official_language
Schwartz, S. H., & Bilsky, W. (1987). Toward a universal psycho-
logical structure of human values. Journal of Personality and
Social Psychology, 53, 550-562.
Schwarz, N., & Clore, G. L. (1983). Mood, misattribution, and judg-
ments of well-being: Informative and directive functions of
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from

Feinberg and Willer 17
affective states. Journal of Personality and Social Psychology,
45, 513-523.
Shweder, R. A., Much, N. C., Mahapatra, M., & Park, L. (1997).
The "big three" of morality (autonomy, community, and divin-
ity), and the "big three" explanations, of suffering. In A. Brandt
& P. Rozin (Eds.), Morality and Health (pp. 119-169). New
York: Routledge.
Skitka, L. J., & Bauman, C. W. (2008). Moral conviction and politi-
cal engagement. Political Psychology, 29, 29-54.
Skitka, L. J., Bauman, C. W., & Sargis, E. G. (2005). Moral con-
viction: Another contributor to attitude strength or something
more? Journal of Personality and Social Psychology, 88,
895-917.
Skitka, L. J., & Morgan, G. S. (2009). The double-edged sword of
a moral state of mind. In D. Narvaez & D. K. Lapsley (Eds.),
Personality, identity, and character: Explorations in moral psy-
chology (pp. 355–375). New York: Cambridge University Press.
Skitka, L. J., & Mullen, E. (2002). The dark side of moral convic-
tion. Analyses of Social Issues and Public Policy, 2, 35-41.
Tetlock, P. E., Kristel, O. V., Elson, S. B., Green, M. C., & Lerner,
J. S. (2000). The psychology of the unthinkable: Taboo trade-
offs, forbidden base rates, and heretical counterfactuals.
Journal of Personality and Social Psychology, 78, 853-870.
Thorisdottir, H., Jost, J. T., Liviatan, I., & Shrout, P. E. (2007).
Psychological needs and values underlying left-right political
orientation: Cross-national evidence from Eastern and Western
Europe. Public Opinion Quarterly, 71, 175-203.
Waytz, A., Gray, K., Epley, N., & Wegner, D. (2010). Causes
and consequences of mind perception. Trends in Cognitive
Sciences, 14, 383-388.
Wheeler, S. C., Petty, R. E., & Bizer, G. Y. (2005). Self-schema
matching and attitude change: Situational and dispositional
determinants of message elaboration. Journal of Consumer
Research, 31(4), 787-797.
Wright, J. C., Cullum, J., & Schwab, N. (2008). The cognitive and
affective dimensions of moral conviction: Implications for
attitudinal and behavioral measures of interpersonal tolerance.
Personality and Social Psychology Bulletin, 34, 1461-1476.
at UNIV TORONTO on October 7, 2015psp.sagepub.comDownloaded from