ArticlePDF Available

Abstract and Figures

Both liberals and conservatives accuse their political opponents of partisan bias, but is there empirical evidence that one side of the political aisle is indeed more biased than the other? To address this question, we meta-analyzed the results of 51 experimental studies, involving over 18,000 participants, that examined one form of partisan bias -- the tendency to evaluate otherwise identical information more favorably when it supports one’s political beliefs or allegiances than when it challenges those beliefs or allegiances. Based on previous literature, two hypotheses were tested: an asymmetry hypothesis (predicting greater partisan bias in conservatives than liberals) and a symmetry hypothesis (predicting equal levels of partisan bias in liberals and conservatives). Mean overall partisan bias was robust (r = .245) and there was strong support for the symmetry hypothesis: liberals (r = .235) and conservatives (r = .255) showed no difference in mean levels of bias across studies. Moderator analyses reveal this pattern to be consistent across a number of different methodological variations and political topics. Implications of the current findings for the ongoing ideological symmetry debate, and the role of partisan bias in scientific discourse and political conflict are discussed. **ACCESS FULL TEXT HERE: http://journals.sagepub.com/eprint/FHAK8VDgpGuJv3qm843D/full
Content may be subject to copyright.
Running head: BIAS IS BIPARTISAN 1
At Least Bias is Bipartisan:
A Meta-Analytic Comparison of Partisan Bias in Liberals and Conservatives
Peter H. Ditto Brittany S. Liu
University of California, Irvine Kalamazoo College
Cory J. Clark Sean P. Wojcik
Florida State University Upworthy
Eric E. Chen Rebecca H. Grady Jared B. Celniker Joanne F. Zinger
University of California, Irvine
Perspectives on Psychological Science, in press
Abstract
Both liberals and conservatives accuse their political opponents of partisan bias, but is there
empirical evidence that one side of the political aisle is indeed more biased than the other? To
address this question, we meta-analyzed the results of 51 experimental studies, involving over
18,000 participants, that examined one form of partisan bias -- the tendency to evaluate
otherwise identical information more favorably when it supports one’s political beliefs or
allegiances than when it challenges those beliefs or allegiances. Based on previous literature, two
hypotheses were tested: an asymmetry hypothesis (predicting greater partisan bias in
conservatives than liberals) and a symmetry hypothesis (predicting equal levels of partisan bias
in liberals and conservatives). Mean overall partisan bias was robust (r = .245) and there was
strong support for the symmetry hypothesis: liberals (r = .235) and conservatives (r = .255)
showed no difference in mean levels of bias across studies. Moderator analyses reveal this
pattern to be consistent across a number of different methodological variations and political
topics. Implications of the current findings for the ongoing ideological symmetry debate, and the
role of partisan bias in scientific discourse and political conflict are discussed.
Keywords: bias, motivated reasoning, ideology, politics, meta-analysis
Word Count: 190
Correspondence concerning this article should be addressed to Peter H. Ditto, Department of
Psychology & Social Behavior, 4201 Social & Behavioral Sciences Gateway, University of
California, Irvine, CA 92697-7085. Email: phditto@uci.edu.
https://doi.org/10.1177/1745691617746796
Perspectives on Psychological Science
1 –19
© The Author(s) 2018
Reprints and permissions:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/1745691617746796
www.psychologicalscience.org/PPS
We asked 951 U.S. visitors to the website YourMorals
.org (http://www.yourmorals.org) how well they
thought the term “biased” described the average Demo-
crat and the average Republican. Respondents describ-
ing themselves as Democrats saw the average Republican
as substantially more biased than the average Democrat.
Republican respondents expressed the mirror image
belief that the average Democrat was substantially more
biased than the average Republican (see Fig. 1).1
This finding should be unsurprising to even a casual
observer of contemporary U.S. politics. A few hours
watching cable news or reading accounts of political
events on any of hundreds of partisan websites will
reveal a pervasive narrative in which political allies are
characterized as rational, informed, and reasonable,
whereas political opponents are described as irrational
“low-information voters” blinded by partisan bias.
These recriminations are distinctly mutual, to the point
that politicians and pundits from both the left and right
rely on the same colorful phrases to capture how the
other side is “drinking the Kool-Aid” (e.g., Huffington,
2002; O’Reilly, 2005) or suffering from one form or
another of “derangement syndrome” (Horowitz, 2008;
Krauthammer, 2003; Raimondo, 2016).
In this article, we take such reciprocal accusations
of partisan bias as our starting point and examine
whether either side’s accusations are correct. Is there
empirical evidence to support the contention that one
746796PPSXXX10.1177/1745691617746796Ditto et al.Bias Is Bipartisan
research-article2018
Corresponding Author:
Peter H. Ditto, Department of Psychology & Social Behavior, 4201
Social & Behavioral Sciences Gateway, University of California, Irvine,
CA 92697-7085
E-mail: phditto@uci.edu
At Least Bias Is Bipartisan: A
Meta-Analytic Comparison of Partisan
Bias in Liberals and Conservatives
Peter H. Ditto1, Brittany S. Liu2, Cory J. Clark3,
Sean P. Wojcik1, Eric E. Chen1, Rebecca H. Grady1,
Jared B. Celniker1, and Joanne F. Zinger1
1Department of Psychology & Social Behavior, University of California, Irvine; 2Department
of Psychology, Kalamazoo College; and 3Department of Psychology, Florida State University
Abstract
Both liberals and conservatives accuse their political opponents of partisan bias, but is there empirical evidence that
one side of the political aisle is indeed more biased than the other? To address this question, we meta-analyzed the
results of 51 experimental studies, involving over 18,000 participants, that examined one form of partisan bias—
the tendency to evaluate otherwise identical information more favorably when it supports one’s political beliefs or
allegiances than when it challenges those beliefs or allegiances. Two hypotheses based on previous literature were
tested: an asymmetry hypothesis (predicting greater partisan bias in conservatives than in liberals) and a symmetry
hypothesis (predicting equal levels of partisan bias in liberals and conservatives). Mean overall partisan bias was robust
(r = .245), and there was strong support for the symmetry hypothesis: Liberals (r = .235) and conservatives (r = .255)
showed no difference in mean levels of bias across studies. Moderator analyses reveal this pattern to be consistent
across a number of different methodological variations and political topics. Implications of the current findings for
the ongoing ideological symmetry debate and the role of partisan bias in scientific discourse and political conflict are
discussed.
Keywords
bias, motivated reasoning, ideology, politics, meta-analysis
2 Ditto et al.
side of the political aisle is more biased than the other?
Or is partisan bias a bipartisan problem, or perhaps
little problem at all?
Assessing the magnitude of partisan bias across the
political spectrum is a challenging task, ill-suited to
examination in a single survey or experiment. Conse-
quently, we report a targeted meta-analytic comparison
of the magnitude of one particular variety of partisan
bias in liberals and conservatives—the tendency to
evaluate otherwise identical information more favorably
when it supports one’s political beliefs or allegiances
than when it challenges those beliefs or allegiances—
examining results from 51 different experimental tests
involving more than 18,000 participants.
Defining Partisan Bias
At the broadest level, partisan bias refers to a general
tendency for people to think or act in ways that unwit-
tingly favor their own political group or cast their own
ideologically based beliefs in a favorable light. Politi-
cally involved individuals, of course, hold many beliefs
that favor their chosen political party or ideology, and
many engage in actions deliberately intended to pro-
mote the political groups they identify with and the
political beliefs they hold. Our focus is on cases in
which this favoritism is less conscious and intentional,
such that people are generally unaware that their politi-
cal affinities have affected their judgments or behavior.
This kind of partisan bias can take many forms and can
occur at multiple levels of the information processing
sequence, including selectively exposing oneself to
information that supports one’s own political group or
views (e.g., Iyengar & Hahn, 2009; Stroud, 2008), selec-
tively remembering information that supports one’s
own political group or views (e.g., Frenda, Knowles,
Saletan, & Loftus, 2013), and, most prototypically, selec-
tively evaluating information in ways that support one’s
own political group or views (e.g., Lord, Ross, & Lepper,
1979).
In real-world political discourse, partisan bias is
often labeled as hypocrisy, in that the individual applies
different (and harsher) standards when evaluating the
judgments and behavior of political opponents than
when evaluating similar or identical judgments or
behaviors displayed by political allies. Analogously, the
classic approach to empirical examination of partisan
bias is to ask participants to evaluate “matched” infor-
mation: information that is as identical as possible in
every way except that in one case it favors the partici-
pant’s political affinities (politically congenial informa-
tion), and in the other it challenges those affinities
(politically uncongenial information). For example,
Lord etal. (1979) recruited participants with strong
attitudes either in support of or in opposition to capital
punishment and asked them to rate the methodological
quality of fictitious but realistic empirical studies exam-
ining whether the death penalty deters homicide. Two
versions of the studies were created: one with results
supporting the deterrent efficacy of capital punishment,
and one with results showing that capital punishment
actually increased rather than decreased homicide rates.
On average, participants (regardless of whether they
supported capital punishment or opposed it) rated the
studies as better quality research when the results sup-
ported their views on the efficacy of capital punishment
than when they challenged those views, despite the
fact that the methodologies of the studies were held
constant across conditions and only the results were
altered. Likewise, Cohen (2003) presented participants
self-identifying as Democrats or Republicans with iden-
tical welfare policies that were said to be strongly sup-
ported by either the majority of congressional Democrats
or the majority of congressional Republicans. Both
Democratic and Republican participants expressed
more positive views of the identical policy when it was
ostensibly supported by members of their own party
than by members of the opposition party.
These studies rely on a logic for demonstrating bias
that is ubiquitous, albeit typically implicit, in psycho-
logical research and grounded in the logic of expected
utility theory (von Neumann & Morgenstern, 1944).
According to Kahneman and Tversky (1984; see also
Keys & Schwartz, 2007), a fundamental axiom of all
analyses of rational choice is the principle of invari-
ance: Judgments should not be affected by trivial
changes (i.e., those irrelevant to the decision) in the
1
2
3
4
5
6
7
Democratic
Participants
Republican
Participants
Level of Agreement (7 = Strongly Agree)
“Republicans Are Biased”
“Democrats Are Biased”
Fig. 1. Mean level of agreement with how well the term “biased”
describes the average Democrat and the average Republican, pre-
sented separately for Democratic and Republican participants (N =
951). Error bars indicate ±1 SE.
Bias Is Bipartisan 3
way information is presented. If a decision about oth-
erwise identical alternatives is affected by whether
those alternatives are presented in terms of the number
of lives lost versus the number saved (Tversky &
Kahneman, 1981), or if a judgment about otherwise
identical behavior differs on the basis of whether it was
enacted by an African American person or by a White
person (Duncan, 1976), then some deviation from ratio-
nality (i.e., bias) is implied. Analogously, if the identical
scientific study or policy proposal is evaluated differ-
ently depending on whether it reflects positively on
liberals or conservatives, partisan bias is implied. The
magnitude of that bias (i.e., the divergence between
how that study or policy is evaluated when it is politi-
cally congenial vs. politically uncongenial) can be mea-
sured to determine whether the bias is significantly
more pronounced for participants on the political left
or the political right.
The simplicity of this analysis is belied by the fact
that applying the logic of invariance in actual empirical
studies faces a number of challenges (Kahan, 2016;
Keys & Schwartz, 2007). Information supplied to par-
ticipants must be experimentally manipulated and care-
fully matched to rule out inadvertent informational
differences between conditions. This minimizes the
possibility that the manipulated information (e.g.,
frame, race, politics) itself conveys relevant information
that could plausibly account for differential judgments
from a Bayesian or related normative perspective. Still,
the difficulty of ruling out counterexplanations that are
based on rational cognitive factors such as expectations
(“priors in Bayesian terms) has vexed research on
motivated perception and reasoning for decades (Ditto,
2009; Erdelyi, 1974; Miller & Ross, 1975; Nisbett & Ross,
1980; Tetlock & Levi, 1982). Such counterexplanations
are notoriously difficult to rule out completely, but their
plausibility is reduced to the extent that (a) the politi-
cally congenial information and politically uncongenial
information presented to participants are matched in
every way possible except for their agreement with a
participant’s political beliefs or allegiances, and (b) par-
ticipants’ evaluations are specifically focused on the
validity or quality of the matched information provided
rather than on a general assessment of the information’s
conclusion (“updating” in Bayesian terms).
Evidence for Asymmetrical Partisan Bias
Interest in locating bias along the political spectrum
has deep roots in psychology, stretching back at least
to work by Adorno and colleagues on the authoritarian
personality (Adorno, Frenkel-Brunswik, Levinson, &
Sanford, 1950). Although the specifics have evolved
over the years (Altemeyer, 1981, 1996), the essential
thesis of this research tradition is that deep-seated con-
flicts (psychodynamic and/or interpersonal) predispose
some people to extreme “conservative” views character-
ized by conventionalism, antipathy toward minority
groups, a preference for strong authoritarian leaders,
and rigid black-or-white, good-or-bad thinking.
Recent research in political psychology has updated
and reinforced this notion that conservative political
views are tied to biased thinking and, in particular,
resistance to novel or threatening information. For
example, political conservatism has been described as
a form of motivated social cognition associated with a
host of personal dispositions related to resistance to
change (dogmatism; low levels of openness to experi-
ence; and high need for order, structure, and closure;
Jost, Glaser, Kruglanski, & Sulloway, 2003). Other work
has found associations between conservatism and threat
sensitivity (Hibbing, Smith, & Alford, 2014; Lilienfeld &
Latzman, 2014), avoidant search strategies (Shook &
Fazio, 2009), shallow system 1 thinking (Eidelman,
Crandall, Goodman, & Blanchar, 2012), valuation of
group loyalty (Graham etal., 2013), and self-enhancement
motivation (Wojcik, Hovasapian, Graham, Motyl, & Ditto,
2015). All of these factors could plausibly manifest them-
selves as a stronger tendency among political conserva-
tives than among political liberals to favor information
that supports rather than challenges their political
affinities.
Evidence for Symmetrical Partisan Bias
No analogous research tradition has championed a
hypothesis of greater bias in liberals than in conserva-
tives. There is, however, considerable theory and data
to suggest that conservatives do not have a monopoly
on bias. The psychological literature is replete with
examples of motivated reasoning, particularly in the
form of self- and group-enhancing biases, and these
biases have been found in a multitude of different pop-
ulations and contexts (Alicke, 1985; Billig & Tajfel, 1973;
Darley & Gross, 1983; Ditto, 2009; Hastorf & Cantril,
1954; Kunda, 1990; Mercier & Sperber, 2011). There is
little reason to expect political liberalism to provide
immunity against motivated reasoning and some reason
to expect that political and moral judgments in general
may be particularly vulnerable to motivational and
affective influence (Ditto, Pizarro, & Tannenbaum, 2009;
Haidt, 2001).
More specifically, just as the “rigidity-of-the-right”
hypothesis underlying work on the authoritarian person-
ality was challenged almost immediately by arguments
that extreme ideologues at both ends of the political
4 Ditto et al.
spectrum tend toward cognitive inflexibility (Rokeach,
1956; Shils, 1954; Taylor, 1960), recent research confirms
that many tendencies often viewed as particularly char-
acteristic of conservative thought are found in liberals
too, if you look in the right place. The central theme of
this work is that all people are motivated to defend core
beliefs and moral commitments, but because beliefs,
commitments, and moral sensitivities differ across the
political spectrum (e.g., Graham etal., 2013), similar
motivations will lead liberals and conservatives to direct
bias and intolerance toward different topics and targets
(e.g., Brandt, Reyna, Chambers, Crawford, & Wetherell,
2014). This analysis suggests that, in terms of any specific
political judgment, either liberal or conservative bias
could be magnified depending on how that judgment
affects each side’s core commitments (Crawford, 2012,
2014), but that if judgments were aggregated across
politically relevant topics, both sides would reveal an
equal proclivity to bend information in their political
favor. Implicit in this analysis is that academic psychol-
ogy’s particular focus on bias in political conservatives
is partly a function of the blind spots (Pronin, 2007) of
a scientific discipline that is overwhelmingly composed
of political liberals (Duarte etal., 2015; Inbar & Lammers,
2012).
The Current Study
How bias is distributed across the political spectrum is
clearly a matter of current empirical debate. Two dif-
ferent hypotheses can be supported by evidence in the
literature: an asymmetry hypothesis that predicts greater
partisan bias in conservatives than liberals and a sym-
metry hypothesis that predicts that levels of partisan bias
will not differ between liberals and conservatives. It is
also possible, of course, that partisan bias could be
greater in liberals than in conservatives even though
this hypothesis has not received extensive attention in
the literature. The current study seeks to evaluate these
hypotheses in a targeted meta-analytic comparison of
the magnitude of one prototypical form of partisan bias
among liberals and conservatives in the United States.
We selected meta-analysis as our approach to take
advantage of the wealth of data on partisan bias that
have already been collected. Meta-analysis also allows
us to examine partisan biases across studies using dif-
fering operationalizations of acceptance of or resistance
to political information, left versus right political orien-
tation, and judgments about a variety of political topics.
Given the challenges of differentiating partisan bias
from some form of rational belief updating, we restricted
our analysis to studies in which the strongest inferences
about bias can be made: experimental studies similar
to those conducted by Lord and colleagues (1979) and
Cohen (2003) that used matched-information designs
to explore partisan biases in the processing of politi-
cally congenial and politically uncongenial information.
These studies come from many different labs, including
scholars who support both the symmetry and asym-
metry perspectives. Our goal is to provide a thorough
representation of the extant psychological research
regarding susceptibility to partisan bias in liberals and
conservatives.
Method
We conducted literature searches using PsycINFO,
Psych Articles, and Worldwide Poli Sci databases. We
searched for the following terms anywhere in the main
text: “bias* assim*,“confirm* bias*,“my* bias*,“bias*
evaluat*,” “motiv* reason*,” and “motiv* skeptic*.” We
also searched for the reverse construction of each term
(e.g., “assim* bias*”). We included the term polit* in
each search to limit our results to studies with political
content. An initial search was conducted in October
2012 and was updated in October 2014 and December
2016. In an effort to locate studies that fit our inclusion
criteria but were not published or did not fall under
our literature search terms, we performed a search of
the Social Science Research Network (an online reposi-
tory that contains both published and unpublished
works), emailed the Society for Personality and Social
Psychology listserv and well-known researchers in the
field requesting articles fitting our criteria, and searched
the reference lists of articles that fit our inclusion cri-
teria. Two additional articles were suggested by one of
the reviewers of the initial version of this article. These
searches returned a total of more than 1,500 articles,
book chapters, and dissertations.
After an initial culling of articles that were clearly
inappropriate for inclusion (e.g., nonempirical pieces),
each remaining article was evaluated by at least two
members of our research team to determine whether it
met four inclusion criteria. In rare instances of disagree-
ment, decisions were resolved through discussion with
the whole group.
Inclusion criteria
The four criteria for a study to be included in our analy-
sis were the manipulation of political congeniality, the
measurement of left-right political orientation, a mea-
sure of information evaluation, and a sample composed
of U.S. participants.
Manipulation of political congeniality. Included
studies had to manipulate (either within or between sub-
jects) whether participants were presented with stimuli
Bias Is Bipartisan 5
that either (a) supported or opposed their political beliefs
(e.g., Lord etal., 1979) or (b) associated a particular pol-
icy or behavior with the participant’s own party or the
opposing political party (e.g., Cohen, 2003). Political-
congeniality manipulations included fictional scientific
studies with results supporting either liberal or conserva-
tive beliefs, examples of similar behavior demonstrated
by liberal or conservative actors, and identical policies
endorsed by Democratic or Republican politicians. We
excluded studies in which the manipulated information
was only loosely matched, such as studies presenting
participants with persuasive essays for liberal and conser-
vative positions that differed substantially in their content
(e.g., Taber & Lodge, 2006).
Measure of left-right political orientation. Included
studies had to measure participants’ self-reported place-
ment on a left/liberal to right/conservative dimension of
political orientation. Variations included measures of
liberal-conservative ideology, Democratic or Republican
party affiliation, and endorsement of specific attitudes
with a clear left-right dimension (e.g., in favor of or
against gun control). We did not include studies that
measured only personality dimensions associated with
political ideology (e.g., right-wing authoritarianism) or
that equated conservative ideology with prejudicial atti-
tudes (e.g., toward African Americans). Studies were also
excluded if they included only one ideological group
(e.g., conservatives only); deriving estimates of bias from
liberals and conservatives evaluating the same closely
matched stimuli most effectively leverages the power of
matched information designs to isolate and compare the
magnitude of partisan bias.
Information evaluation measure. Studies needed to
measure participants’ evaluation of the validity, quality, or
acceptance of the matched politically congenial and polit-
ically uncongenial information. Examples of information-
evaluation measures included ratings of a scientific study’s
methodological quality, approval or disapproval of a politi-
cal actor’s behavior, and endorsement of specific policy
proposals presented in the stimulus materials. Studies
were not included if their only evaluation measure was
endorsement of a general political attitude (e.g., attitude
toward capital punishment after reading a study on capi-
tal punishment) given the vulnerability of general attitu-
dinal measures to normative counterexplanation.2
U.S. sample. Although we have no reason to doubt the
generality of political bias, our particular interest is on
liberal-conservative differences in the context of U.S. pol-
itics. Because of this focus and the difficulties of defining
liberal and conservative in different national contexts, we
included only studies with participants from the United
States.
Of the articles evaluated, 48 included data that
metall four inclusion criteria. Because the majority of
qualifying articles were interested in documenting the
existence of partisan bias in general rather than cross-
ideological comparisons of bias, only 11 of 48 articles
included enough information to calculate separate lib-
eral and conservative effect sizes. For the remaining
articles, we contacted authors and asked them to pro-
vide additional analyses or to provide data that we
could use to perform these analyses ourselves. For 10
articles, the relevant data were no longer available or
the authors did not respond to our requests. For articles
with multiple studies, each unique sample was counted
as an individual study and contributed one effect size
in the main analyses. If a study included judgments about
multiple topics manipulated between subjects (i.e., some
participants responded to materials about gun control
and others responded to materials about capital punish-
ment; e.g., MacCoun & Paletz, 2009), effect sizes for each
topic were entered as a separate “study.” Our final sam-
ple included effect sizes from 51 studies culled from 38
articles, with a total of 18,815 participants (for the full
list of included studies, see Table 1).
Primary analyses
Among the 51 studies included, the statistical tests used
to assess the differential evaluation of politically con-
genial and politically uncongenial information varied a
great deal. Reported statistics included ts or Fs with
their degrees of freedom; βs with their standard errors;
χ2s with their sample sizes; and means, standard devia-
tions, and their sample sizes. For each study, we used
the available statistics to compute a Pearson’s r effect
size for overall partisan bias (roverall). Positive values
reflect the degree to which both liberal and conserva-
tive participants responded more positively to politi-
cally congenial information than to politically
uncongenial information. We examined support for the
symmetry versus asymmetry hypotheses in two ways.
First, we calculated separate partisan-bias effect sizes
for liberals and conservatives for each study (rlib and
rcon).3 Positive rlib values indicate that liberals evaluated
liberal-friendly stimuli more positively than they did
conservative-friendly stimuli. Likewise, positive rcon val-
ues indicate that conservatives evaluated conservative-
friendly stimuli more positively than they did
liberal-friendly stimuli. Second, we calculated an rdiff
effect size from each study reflecting the degree to
which rcon and rlib differ within each study. We assigned
positive rdiff values to indicate that rcon was greater than
6 Ditto et al.
Table 1. Effect-Size Estimates and Study Characteristics for All Studies
Study roverall rdiff r lib rcon Pol Man Sam Info Topic
Bolsen etal. (2014) .361***
(153)
.042 .326**
(81)
.400***
(72)
P So Rep NS Environmental – general
Bergan (2012) .094
(110)
–.268** .338**
(59)
–.203
(51)
P So Stu NS Abortion
Bullock (2011) .290***
(1,633)
.184*** .106**
(803)
.449***
(830)
P So Rep NS Health care
Claassen & Ensley (2016) .194***
(549)
.031 .167**
(297)
.227***
(252)
P So Rep NS Campaign tricks
Christenson & Kriner (2017) .384***
(354)
.042 .351***
(195)
.423***
(159)
P So Rep NS Presidential behavior
Ciuk & Yost (2016) .271*
(77)
.009 .263
(39)
.280
(38)
P So Ad Com NS Environmental – general
Cohen (2003) .696***
(79)
.009 .692***
(48)
.702***
(31)
I So Stu NS Welfare
Crawford & Xhambazi (2015) .254**
(163)
–.005 .260**
(115)
.249
(48)
I So Ad Ol NS Protest
Crawford (2012) .418***
(161)
–.112 .479***
(110)
.268
(51)
I So Ad Ol NS Presidential behavior
Crawford etal. (2013) .292***
(211)
–.380*** .536***
(134)
–.201
(77)
I Co Stu, Ad S Multiple social policies
Crawford etal. (2014) .126
(157)
–.149 .205*
(121)
–.159
(36)
I So Ad Ol NS Presidential behavior
Crawford, Kay, & Duke (2015)
Sample 1
.254**
(153)
.009 .249**
(112)
.268
(41)
I So Ad Ol NS Multiple social policies
Crawford etal. (2015) Sample 2 .181*
(162)
–.029 .197*
(121)
.130
(41)
I So Ad Ol NS Multiple social policies
Druckman (2001) .419***
(239)
.001 .419***
(149)
.420***
(90)
P So Stu NS Asian disease scenario
Furgeson etal. (2008a) .145
(114)
.092 .086
(80)
.286
(34)
C Co Law Stu NS Constitution
interpretation
Furgeson etal. (2008b) .241***
(270)
.014 .234**
(209)
.266*
(61)
C Co Stu NS Tax policy
Groenendyk (2013) .001
(161)
–.070 .065
(88)
–.078
(73)
P So Rep NS Outsourcing
Hawkins & Nosek (2012) Study 1 .208***
(895)
–.014 .218***
(592)
.189**
(303)
P So Ad Ol NS Welfare
Hawkins & Nosek (2012) Study 2 .245***
(928)
.023 .229***
(590)
.274***
(338)
P So Ad Ol NS Education policy
Kahan (2013) .239***
(1,062)
–.014 .253***
(550)
.225***
(512)
C Co Rep NS Global warming
Kahan etal. (2016) .162***
(723)
.084* .078
(362)
.243***
(361)
C So Rep NS Multiple topics
Kahan etal. (2011) .339***
(1,466)
.035 .307***
(736)
.370***
(729)
C Co Rep S Multiple social policies
Kahan etal. (2012) .366***
(200)
.015 .353***
(102)
.380***
(98)
C So Rep NS Protest
Kahan, Peters, etal. (2017) .212***
(397)
.146** .072
(205)
.352***
(193)
C Co Rep S Gun control
Kahan, Jamieson etal. (2017) .116***
(1,391)
.092*** .026
(714)
.208***
(677)
C Co Ad Ol NS Zika virus
Kam (2005) .302***
(166)
.085 .248**
(112)
.412**
(54)
P So Stu NS Food irradiation policy
Kopko etal. (2011) .041
(100)
.001 .042
(60)
.040
(40)
P So Stu NS Ballots
(continued )
Bias Is Bipartisan 7
Study roverall rdiff r lib rcon Pol Man Sam Info Topic
Lai & Nosek (2012)a.096*
(545)
.039 .065
(334)
.144*
(211)
P So Ad Ol NS Education policy
Liu (2014)a Study 1 .245***
(381)
.074 .220***
(335)
.428**
(46)
I Co Ad Ol S Abstinence
Liu (2014)a Study 2 .170***
(433)
–.065 .120*
(363)
.021
(70)
I Co Ad Ol S Capital punishment
Liu (2014)a Study 3 .366***
(537)
.100* .325***
(440)
.538***
(97)
I Co Ad Ol S Gun control
Lopez (1994)a Study 1 .116
(126)
.210* –.105
(61)
.314*
(65)
IA Co Stu S Capital punishment
Lopez (1994)a Study 2 .076
(47)
.420** –.310
(26)
.526*
(21)
IA Co Stu S Capital punishment
Lord etal. (1979) .643***
(48)
.176 .518
(24)
.740***
(24)
IA Co Stu S Capital punishment
MacCoun & Paletz (2009) Sample
1
.012
(156)
–.260*** .270*
(78)
–.248*
(78)
I Co Rep S Gun control
MacCoun & Paletz (2009) Sample
2
.290***
(148)
–.120 .409***
(67)
.186
(81)
I Co Rep S Capital punishment
MacCoun & Paletz (2009) Sample
3
.562***
(134)
.060 .518***
(61)
.596***
(73)
I Co Rep S Medical marijuana
MacCoun & Paletz (2009) Sample
4
.237***
(171)
.074 .175
(97)
.317**
(74)
I Co Rep S Education policy
Malka & Lelkes (2010) .233***
(322)
.008 .224**
(134)
.240**
(188)
I So Rep NS Farm subsidies
Mullinix (2016) .495***
(759)
–.065 .541***
(399)
.441***
(360)
P So Rep NS Multiple social policies
Munro & Munro (2014) .080
(106)
.083 .009
(62)
.181
(44)
P Co Stu S Scientific evidence
Nawara (2011)a.019
(158)
.032 –.008
(94)
.059
(64)
P So Stu NS Presidential behavior
Scurich & Shniderman (2014)
Study 1
.223*
(125)
–.129 .359**
(56)
.108
(69)
IA Co Ad Ol S Capital punishment
Scurich and Shniderman (2014)
Study 2
.349***
(128)
.078 .300**
(87)
.448**
(41)
IA Co Ad Ol S Abortion
B. K. Smith (2014) .042
(179)
.001 .041
(124)
.043
(57)
P So Ad Ol NS NSA policy
C. T. Smith, Ratliff, & Nosek
(2012) Study 1
.238***
(559)
–.009 .244***
(374)
.226**
(185)
P So Ad Ol NS Welfare
C. T. Smith, Ratliff, & Nosek
(2012) Study 2
.209***
(509)
.042 .190***
(410)
.290**
(99)
P So Ad Ol NS Welfare
Tannenbaum, Fox, & Rogers
(2014) Study 1
.104**
(238)
–.009 .109
(172)
.088
(66)
I So Ad Ol NS Public policy
Tannenbaum, Fox, & Rogers
(2014) Study 2
.147**
(366)
–.095 .210**
(249)
.007
(117)
I Co Ad Ol NS Public policy
Tannenbaum, Fox, & Rogers
(2014) Study 3
.199
(88)
.035 .169
(50)
.240
(38)
I Co Bur NS Public policy
Tannenbaum, Fox, & Rogers
(2014) Study 4
.389*
(30)
–.048 .451
(11)
.356
(19)
I Co May NS Public policy
Note: Values in parentheses indicate the number of participants. Positive roverall values indicate greater bias; positive rdiff values indicate
conservatives show more bias than liberals. Pol = political orientation; P = party; C = composite; I = ideology; A = issue attitude; Man =
manipulation type; So = source; Co = content; Sam = sample type; Rep = representative; Stu = students; Ad = adults; C = community; Ol = online;
Bur = bureaucrats; May = U.S. mayors; Info = information type; S = scientific; NS = nonscientific.
aThese studies are unpublished.
*p < .05. **p < .01. ***p < .001.
Table 1. (Continued)
8 Ditto et al.
rlib in a given study (and negative rdiff values to indicate
that rlib was greater than rcon) in line with the asymmetry
hypothesis described above. All aggregate r effect sizes
were computed with the Comprehensive Meta-Analysis
software (Ver. 3.0; Biostat, Inc., Englewood, NJ), which
converts r effect sizes to Fisher z values, and were
analyzed using random-effects models.
Moderator analyses
Because of the relatively modest number of studies
included in our analysis, we limited our examination
of potential moderators to five. The moderators we
chose to examine were three common methodological
variations found in existing studies (the nature of the
manipulation, the nature of the political-orientation
measure, and the nature of the sample) and two
additional variables we suspected might moderate the
magnitude of partisan-bias effects. At least two mem-
bers of our research team coded each study for each
moderator.
Political-congeniality manipulation. We coded for
whether the manipulation of political information entailed
varying the content or the source of the stimuli being
evaluated. For instance, Lord etal. (1979) manipulated
the content of the political information by showing par-
ticipants evidence that either supported or challenged
the effectiveness of capital punishment. Cohen (2003)
manipulated the source of political information when he
showed participants the same welfare policy but varied
whether that policy was endorsed by congressional
Democrats or congressional Republicans.
Political orientation measure. Measures of political
orientation were coded for whether they were based on
liberal-conservative ideology, Democratic-Republican party
affiliation, or liberal-conservative position on an issue-specific
attitude.
Sample. We coded for whether the sample was drawn
from a student population, a convenience sample of
adults online, or a nationally representative sample.
Type of information. We coded for whether the infor-
mation was presented in the form of scientific data (e.g.,
Lord etal., 1979) or nonscientific information such as a
description of a specific policy (e.g., Cohen, 2003) or the
behavior of a political actor (e.g., Crawford, 2012).
Political topic. We coded for the specific topic repre-
sented in the political-congeniality manipulation. Among
the 51 studies, six political topics were used in three or
more studies, allowing us to aggregate and compare their
results: capital punishment (k = 6), presidential behavior
(k = 5),4 welfare policy (k = 4), environmental policy (k =
4), abortion (k = 3), and gun control (k = 3).5
Results
Table 1 presents mean effect sizes for overall partisan
bias (roverall), partisan bias separately for liberals and
conservatives (rlib, rcon), and the relative magnitude of
liberal and conservative partisan bias (rdiff) for all 51
studies. Table 1 also shows how each study was coded
on the five moderator variables.
Overall partisan bias ranged from rs = .001 to .696;
some studies showed very little partisan bias and others
showed a great deal of bias. There was also a substantial
range of effect sizes for rlib, rcon, and rdiff, indicating that
studies ranged from showing substantially greater bias
for liberals than for conservatives to showing substan-
tially greater bias for conservatives than for liberals.
Table 2 displays aggregated r effect-size analyses for
the main hypotheses with random-effects models. There
was a statistically significant small- to medium-sized
mean effect of overall partisan bias (roverall = .245, p <
.001; 95% confidence interval, or CI = [.208, .280]),
which suggests that people in general showed a clear
tendency to evaluate politically congenial stimuli more
favorably than similarly structured politically unconge-
nial stimuli.
The average effect sizes for rlib and rcon differed sig-
nificantly from zero, indicating that liberal and conser-
vative participants were both biased in favor of
information that supported their particular political
beliefs and allegiances. The results provided support
for the symmetry hypothesis: The mean levels of liberal
and conservative bias were very similar in magnitude
(rlib = .235, 95% CI = [.192, .276]; rcon = .255, 95% CI =
[.205, .304]) and the aggregate rdiff effect size across all
51 studies was extremely small and was not significantly
different from zero (rdiff = .009, p = .55, 95% CI = [.020,
.038]; see Table 2), indicating no difference in degree
of bias between liberals and conservatives. In other
words, whether partisan bias was aggregated separately
for liberals and conservatives or compared within each
study and then aggregated, our results suggest that
liberals and conservatives were both significantly biased
in favor of information that supported their ideological
beliefs and groups, and the two groups were biased to
very similar degrees.
Moderator analyses
There was significant heterogeneity within roverall and
rdiff effect sizes (see results of QW tests for homogeneity
in Table 2), so we tested whether any of our coded
Bias Is Bipartisan 9
variables moderated our main findings. These modera-
tor analyses should be interpreted cautiously, however,
because the relatively small number of studies exam-
ined in subgroups creates the possibility of confound-
ing among the moderators (e.g., many of the studies
examining a particular political topic may also rely on
a particular methodological approach).
Overall, none of our analyses revealed statistically
significant differences for any of our moderator vari-
ables for either overall magnitude of partisan bias
(roverall), magnitude of bias in liberals and conservatives
separately (rlib, rcon), or the relative magnitude of bias
in liberals and conservatives (rdiff). It is noteworthy that
the overall partisan-bias effect was significant for every
subgroup for all five moderator variables examined. All
statistics for the moderator analyses are reported in
Table 3.
Although we found no significant moderators in our
analysis, the prediction intervals associated with our
mean effect sizes (presented in Table 2) suggest that
the true effects of partisan bias—for liberals, for con-
servatives, and for both groups combined—are likely
to vary widely from study to study, such that true effects
range from nonexistent (very close to zero) to fairly
large. Furthermore, the true effects for the difference
between conservatives and liberals are also likely to
vary, ranging from liberals being slightly more biased
than conservatives to conservatives being slightly more
biased than liberals. These wide prediction intervals
underscore the fact that moderators of these effects are
likely to exist even though we were not able to identify
these moderators in our study.
Publication bias
We addressed the possibility of publication bias in mul-
tiple ways. First, we sought out and included both pub-
lished (k = 42) and unpublished studies (k = 9) of
partisan bias.
Second, we looked at whether publication in a peer-
reviewed source moderated effect size. Published stud-
ies showed a larger mean partisan-bias effect size (roverall =
.266, p < .001) than did unpublished studies (roverall =
.139, p = .003; QB = 6.35, p = .012), but the mean effect
sizes in both sets of studies were significantly greater
than zero. Moreover, our primary interest in this project
was not whether overall bias exists, but rather the rela-
tive magnitude of bias in conservatives and liberals. In
this case, publication status did not moderate results.
Conservatives and liberals showed equivalent levels of
relative bias in both published (rdiff = .001, p = .95) and
unpublished studies (rdiff = .054, p = .11; QB = 2.15, p =
.14).
Third, we used funnel plots to visually assess pub-
lication bias by plotting Fisher’s transformation of the
effect size for each study on the horizontal axis against
the natural log of its sample size on the vertical axis,
and we used linear regression to test the slope through
the points in the funnel plot (Sterne, Becker, & Egger,
2005). Symmetrical funnel plots with a nonsignificant
Table 2. Mean Effect Size Estimates Across All Studies for Overall Partisan Bias, Difference in
Partisan Bias Between Liberals and Conservatives, Liberal Partisan Bias, and Conservative Partisan
Bias
Variable k
Random effects model
Homogeneity
among studies
r
95% CI 95% PI
Lower
bound
Upper
bound
Lower
bound
Upper
bound QW(50) Tau
Overall partisan bias (roverall) 51 .245*** .208 .280 0.003 0.486 307.96*** 0.120
Partisan bias for liberals (rlib) 51 .235*** .192 .276 –0.038 0.508 244.70*** 0.136
Partisan bias for conservatives (rcon) 51 .255*** .205 .304 –0.059 0.569 224.33*** 0.156
Difference in bias between
conservatives and liberals (rdiff)
51 .009 –.020 .038 –0.175 0.175 100.41*** 0.083
Note: Positive roverall, rlib, and rcon values indicate that participants demonstrate bias; positive rdiff values indicate that
conservatives show more bias than liberals. The Q statistic (also known as Cochrane’s Q) is the weighted sum of
squared differences between the observed effects and the weighted average effect. Tau, used for computing prediction
intervals (PIs), is an estimate of the standard deviation of the distribution of true effect sizes, assuming that those
effect sizes are normally distributed. The 95% PI gives the range in which the point estimate of 95% of all true effects
(including those from both previously completed and future studies) are expected to fall, assuming that true effect sizes
are normally distributed through the domain (Borenstein, Hedges, Higgins, & Rothstein, 2009). CI = confidence interval;
QW = within-studies heterogeneity.
***p < .001.
10 Ditto et al.
slope indicate that publication bias is not an issue.
Asymmetry in the funnel plot with a negative slope
indicates publication bias because studies with small
sample sizes showing null or negative effects are absent
from the sample of studies. There was no evidence of
publication bias for either overall partisan bias (roverall
β = .12, p = .42) or the relative degree of bias in con-
servatives and liberals (rdiff β = .01, p = .97).
Discussion
The clearest finding from this meta-analysis was the
robustness of partisan bias. A tendency for participants
to find otherwise identical information more valid and
compelling when it confirmed rather than challenged
their political affinities was found across a wide range
of studies using different kinds of samples, different
operationalizations of political orientation and political
congeniality, and across multiple political topics. The
mean effect for overall partisan bias was modest in size,
but statistically significant partisan-bias effects were
found in 39 of 51 samples and in every subgroup com-
pared in our moderator analyses. That is, the tendency
to evaluate politically congenial information more char-
itably than politically uncongenial information was
found whether the study manipulated congeniality via
the source of the information or its content; whether
political orientation was operationalized as ideology,
party affiliation, or a specific attitude about a particular
political issue; whether the sample was composed of
students, adults opting into an online study, or a rep-
resentative sample of U.S. citizens; whether the infor-
mation evaluated was scientific or nonscientific; and
across several different politically charged topics. None
of this should be surprising given the extensive body
of research confirming a pervasive human tendency
toward motivated reasoning and self- and group
enhancement (Brown & Kobayashi, 2002; Kunda, 1990;
Mercier & Sperber, 2011; Sedikides, Gaertner, & Vevea,
2005; Stanovich, West, & Toplak, 2013). People are less
skeptical consumers of information that they want to
believe than of information that they do not want to
Table 3. Moderator Analyses for Partisan-Bias Effect-Size Estimates
Moderator variable k
Overall partisan bias
Difference in bias
between conservatives
and liberals
rlib rcon
roverall QBp rdiff QBp
Political orientation 1.387 .500 5.464 .065
Issue attitude 5 .289** .136 .168 .421***
Party 19 .222*** .017 .212*** .243***
Ideology 19 .271*** –.048 .297*** .221***
Manipulation type 0.168 .682 0.061 .805
Source 27 .251*** .007 .246*** .253***
Content 24 .236*** .015 .221*** .259***
Sample 3.773 .152 1.346 .510
Representative 16 .281*** .021 .263*** .300***
Students 12 .251*** .052 .197** .314***
Online 19 .208*** .006 .208*** .207***
Topic 2.233 .816 3.414 .636
Capital punishment 6 .248*** .056 .196.300**
Presidential behavior 5 .285*** –.018 .298*** .254**
Welfare 4 .324*** .002 .298*** .316***
Environmental 4 .334*** .034 .310*** .362***
Abortion 3 .192** –.052 .226** .137
Gun control 3 .210* .005 .225* .238
Scientific 0.706 .401 0.030 .862
Not scientific 35 .235*** .010 .226*** .241***
Scientific 16 .268*** .017 .256*** .297***
Note: Results are for random-effect moderator analyses. Positive roverall values indicate greater overall partisan
bias; positive rdiff values indicate that conservatives show more bias than liberals. Moderator analyses were
performed on roverall and rdiff, but liberal and conservative partisan bias (rlib and rcon, respectively) are also
shown for reference. QB = between-studies heterogeneity.
p < .10. *p < .05. **p < .01. ***p < .001.
Bias Is Bipartisan 11
believe (Ditto & Lopez, 1992), and this pattern is as
evident in the political realm as it is in other realms of
life that evoke strong emotions, preferences and social
allegiances.
The question of ideological symmetry
A corollary of the general robustness of partisan bias
was specific support for the symmetry hypothesis. Our
meta-analysis contributes to a long-standing and ongo-
ing debate regarding the psychological similarities and
differences between people occupying the left and right
ends of the ideological spectrum (Adorno etal., 1950;
Brandt etal., 2014; Crawford, 2017; Jost, 2017; Jost
etal., 2003; Rokeach, 1956). Contrary to the view that
political conservatives are particularly prone to defen-
siveness and cognitive rigidity (Adorno etal., 1950; Jost
etal., 2003), our analysis found that when partisan bias
was aggregated across studies, topics, and methodologi-
cal details, both liberals and conservatives were biased
in favor of information that confirmed their political
beliefs, and the two groups were biased to very similar
degrees.
Given the pervasiveness of motivated reasoning and
the strong tribal animosities between left and right that
have long characterized U.S. politics, it might seem odd
to expect people on one side of the political divide to
be substantially more or less evenhanded in their judg-
ments than people on the other side. And yet there is
a large and growing body of literature, including con-
siderable experimental work, associating political con-
servatism with a broad array of motivational orientations
suggestive of cognitive rigidity and resistance to nega-
tive or threatening information (Hibbing etal., 2014;
Jost, 2017). This work is compelling, but it is important
to note that these studies focus their comparisons on
individual differences in general motivational proclivi-
ties (e.g., need for order, tolerance for ambiguity),
whereas our meta-analysis examined specific judgment
outcomes (e.g., the differential evaluation of politically
congenial and politically uncongenial information). As
such, the two sets of studies do not directly contradict
each other, but the question clearly arises as to why
the differential motivational tendencies of liberals and
conservatives documented in past research were not
found to manifest themselves in differential susceptibil-
ity to partisan bias in our meta-analysis.
One possibility is that the asymmetrical psychologi-
cal propensities of liberals and conservatives have their
primary impact not on susceptibility to bias in general
but rather on the topics about which the two groups
are likely to be biased (Brandt etal., 2014; Crawford,
2012, 2014). Greater commitment to attitude positions
is associated with more selective processing and resis-
tance to persuasion (Krosnick, 1988; Pomerantz,
Chaiken, & Tordesillas, 1995; Zuwerink & Devine,
1996), and moral commitments may be particularly
potent in rousing psychological defenses (Mullen &
Skitka, 2006; Skitka, Bauman, & Sargis, 2005). By this
account, conservatives’ heightened discomfort with
uncertainty and threat might reveal itself, not in more
biased processing of information about any political
topic, but rather in relatively pronounced bias about
information that threatens or assuages those (or other)
particularly conservative concerns. A recent study, for
example, found political conservatism to be associated
with greater credulity to information about personal or
societal risks (e.g., attacks by terrorists or sharks) but
not personal or societal benefits (e.g., the health advan-
tages of carrots or cats; Fessler, Pisor, & Holbrook,
2017). Analogously, liberals by this account might be
expected to show particularly biased responses to infor-
mation bearing on their core concerns about protection
for vulnerable groups and societal inequality. Uhlmann,
Pizarro, Tannenbaum, and Ditto (2009), for example,
found political conservatives to be unaffected by the
race of an individual to be sacrificed in a moral dilemma,
whereas liberals did discriminate on the basis of race:
Liberals were significantly less likely to sacrifice an
individual with a stereotypically African American name
than a stereotypically White name (for similar findings,
see Norton, Vandello, & Darley, 2004).6 Our meta-
analysis found only nonsignificant differences in bias
across political topics, but future research with greater
statistical power and topics chosen to map onto the
known psychological and moral sensitivities of liberals
and conservatives (e.g., Graham etal., 2013; Jost etal.,
2003) would be a more compelling test of the topic-
specific bias hypothesis.
Another possibility is that the psychological differ-
ences between liberals and conservatives have their
effects on aspects of the information processing
sequence other than the biased evaluation of political
information. The studies examined in our meta-analysis
all confronted participants with information that either
supported or challenged their political beliefs, a “strong
situation” (Mischel, 1977) likely to evoke motivated
responding in most or all people, and one that pre-
cludes the choice generally available in the natural
environment to direct one’s attention toward or away
from particular kinds of information. It is possible, then,
that the choice of what information to seek out or avoid
is where conservatives’ relative reticence toward novel
and threatening information has its effects, rather than
how that information is processed once it is confronted.
Research in the selective-exposure tradition has pro-
duced several studies suggesting that the tendency to
preferentially seek out information that supports rather
than challenges political views is more pronounced in
conservatives than in liberals (Barberá, Jost, Nagler,
12 Ditto et al.
Tucker, & Bonneau, 2015; Messing & Westwood, 2014;
Nam, Jost, & Van Bavel, 2013; Rodriguez, Moskowitz,
Salem, & Ditto, 2017). It is also true, however, that
several studies have revealed no political differences in
selective-exposure tendencies (Collins, Crawford, &
Brandt, 2017; Frimer, Skitka, & Motyl, 2017; Iyengar &
Hahn, 2009; Stroud, 2008), and a few have suggested
greater selective exposure among liberals than among
conservatives (Bakshy, Messing, & Adamic, 2015;
Knobloch-Westerwick & Meng, 2009).
Much like the pattern seen in our meta-analysis, the
literature on selective exposure reveals considerable
variability across studies in the relative magnitude of
bias in liberals and conservatives; the clearest conclu-
sion to be drawn from the extant data concerns the
proneness of both sides to favor politically congenial
over politically uncongenial information. Research on
political selective exposure, however, is a step ahead
of work on the biased processing of political informa-
tion in its recognition of important boundary conditions
and contextual influences on political bias such as
information utility and attitude importance (Garrett &
Stroud, 2014; Knobloch-Westerwick & Kleinman, 2012;
Knobloch-Westerwick & Meng, 2009). Similar contex-
tual factors have been found to moderate motivated-
reasoning processes outside of the political domain
(Kunda, 1990; Lerner & Tetlock, 1999; Neuberg & Fiske,
1987), and exploring their operation in political moti-
vated reasoning has the potential to clarify predictions
regarding when and in whom partisan bias is most
likely to be found, including variability over time and
political climate (Federico & Malka, 2018).
Of course, further research is needed to thoroughly
investigate all of the speculation above. This research
would ideally include new experimental studies (e.g.,
comparing the magnitude of partisan bias across topics
that differ in attitude importance or moral conviction
for liberals and conservatives), longitudinal studies
(where data are available) to track changes in political
congeniality biases over time and historical context, as
well as additional meta-analyses (e.g., comparing selec-
tive exposure tendencies in liberals and conservatives).
The swelling body of research examining the psycho-
logical underpinnings of liberalism and conservatism
should be particularly helpful in generating testable
hypotheses.
There are almost certainly both important symme-
tries and important asymmetries between liberal and
conservative psychology, and research exploring this
complicated web of commonality is inaptly character-
ized as pursuing “Swiss-style neutrality” or some kind
of false moral equivalency between liberal and conser-
vative ideology (Jost, 2017). Different psychological
processes contribute to different manifestations of bias,
and there are complexities to political ideology that
belie the simple unidimensional (liberal-conservative)
characterization relied on here (e.g., Crawford, Jussim,
Cain, & Cohen, 2013; Iyer, Koleva, Graham, Ditto, &
Haidt, 2012; Malka & Soto, 2015). All this complexity
must be considered in any comprehensive treatment of
the ideological symmetry question, and given that com-
plexity, a simple portrait of the psychological superior-
ity of one ideology over another seems unlikely to
emerge. Moreover, psychological comparisons are com-
pletely independent of, and in no way preclude,
thoughtful assessments of the superiority or inferiority
of political ideologies at a social, economic, or moral
level. Psychological equivalency does not imply moral
equivalency, despite a fundamental human tendency to
conflate descriptive evaluations with prescriptive ones
(Ditto & Liu, 2016; Hume, 1740/1985; Liu & Ditto, 2013).
Political psychologists, ourselves included, face a
unique challenge, highlighted ironically by the findings
of pervasive partisan bias presented here, to prevent
our own political views from influencing how we con-
duct and interpret our research. We agree with Jost
(2017) that a preference for finding commonalities
between ideologies is no less problematic than a prefer-
ence for showing one particular ideology to be psycho-
logically (or morally) superior to others, and we
encourage all researchers interested in partisan bias to
take every precaution to avoid falling prey to the very
phenomenon we seek to understand.
Limitations
Our meta-analysis was more targeted than some
because of our desire to focus on studies that provide
the most compelling evidence for partisan bias: experi-
mental studies using a matched information design to
examine differential evaluation of politically congenial
and politically uncongenial information (Kahan, 2016).
We could have cast our net more broadly to include
studies using correlational data or other experimental
designs or examining other kinds of partisan biases
(e.g., selective exposure, hostile media bias). Instead,
we felt that given the long-established difficulties of
disentangling motivated bias from normative decision
processes (Ditto, 2009; Kahan, 2016; Tetlock & Levi,
1982), focusing only on the highest quality studies as
a first step would provide the most accurate and modest
yardstick to compare bias across groups.
This does not mean, however, that bias is always the
sole explanation for differences found in studies with
carefully matched stimulus materials. For example, sev-
eral studies included in our meta-analysis demonstrate
significant differences in how positively an identical
policy was evaluated, even when the only difference
Bias Is Bipartisan 13
between conditions was a single word indicating
whether one’s own party or the opposing party
endorsed that policy (e.g., Malka & Lelkes, 2010). At
one level this can be construed as bias: A person favors
the very same policy that they would have rejected if
the other party had proposed it. But party labels can
also be thought of as cues, and favoring a policy sup-
ported by people one agrees with on many other issues
can be thought of as a sensible heuristic strategy rather
than a bias (Bullock, 2011; Leeper & Slothuus, 2014).
This interpretational ambiguity, of course, is just one
example of the formidable challenge of ruling out nor-
mative counterexplanation that transcends the study of
bias in political judgment. In our meta-analysis, studies
that manipulate the political content of information
rather than its source are (arguably) less vulnerable to
this ambiguity, and our analysis shows the mean effect
of partisan bias to be equally strong in the former (r =
.236) and in the latter (r = .251). But ultimately, there
is an empirical catch-22 at the heart of all research on
motivated reasoning. Because contextual information
must be manipulated to produce differential motiva-
tions to accept or reject a given piece of target informa-
tion, the informational differences between conditions
that are a necessary part of the motivational manipula-
tion are always a potential cause of any differential
judgments between those same conditions. As long as
information is used to manipulate motivation, the
entanglement between the two (and the potential con-
founding that inevitably results) will always persist, at
least to some degree.
Minimizing the plausibility of normative explanations
for putative bias effects is important in scientific
research, and restricting our meta-analysis to only the
most carefully designed experiments was our attempt
to do that here. But it is important to recognize that in
the real world of politics, as in virtually every real-world
situation, prior belief and motivated bias are naturally
confounded (Ditto, Munro, Apanovich, Scepansky, &
Lockhart, 2003), and both are likely to play a role in
partisan resistance to politically challenging informa-
tion. When confronted with the latest Republican tax-
reform plan, for example, most Democrats approach
that plan both expecting it to be bad policy (based on
prior information to which they have been exposed
about the ineffectiveness of tax cuts, almost certainly
shaped by selective-exposure tendencies) and moti-
vated to perceive it as bad policy, either because aspects
of the policy offend their moral sensitivities or because
of their general antipathy toward the Republicans who
proposed it. This natural coalition of belief and motiva-
tion may help to explain why the bias we observe under
tightly controlled experimental conditions seems so
subtle compared with the seemingly blatant hypocrisy
people often perceive in their real-world political
antagonists.
Another key limitation of our study was our decision
to treat political orientation dichotomously rather than
continuously. This decision flowed primarily from our
focus on matched information designs in which political
congeniality was defined by whether information con-
firms or challenges participants’ existing political views
or allegiances, making the inclusion of individuals with
moderate or politically independent views in continu-
ous analyses problematic. Included studies also used
varied operationalizations of left versus right ideology,
many measuring or reporting it only dichotomously, so
adopting a dichotomous approach allowed us to include
the maximum number of studies in our analyses. Still,
our approach of comparing the magnitude of liberal
and conservative bias in reactions to information
manipulated to either challenge or support partisan
beliefs raises important issues about the equivalency
of stimulus materials across experimental conditions
(for examination of one such issue, see the Supplemen-
tal Material available online) as well as the extent to
which our liberal and conservative samples were
equally extreme in their ideological commitments.
Future work should consider how to best gauge bias
across the continuous spectrum of ideology, most criti-
cally for the ability to evaluate what is likely to be an
important role for ideological extremity in fomenting
partisan bias.
Finally, it is important to consider whether the politi-
cal views of researchers may have influenced the sam-
ple of studies available for our meta-analysis, especially
in a field so disproportionally composed of individuals
whose sympathies lie with one particular political per-
spective (Inbar & Lammers, 2012). In most meta-
analyses, the file-drawer problem is a straightforward
matter of gauging the extent to which null results are
underrepresented in the published data. The current
case is more complicated in that (a) our primary result
of interest is a null finding (no difference in magnitude
of bias between liberals and conservatives) and (b) it
is plausible to consider whether a particular pattern of
affirmative results—those showing strong liberal bias—
might be underrepresented in the literature as well.
First, we made active attempts to uncover and
include data from unpublished sources and conducted
moderator analyses comparing the relative effect size
of conservative and liberal bias in published and
unpublished studies, which revealed no significant dif-
ferences. Second, suppression of evidence of liberal
bias (either active or passive) seems unlikely in that
very few of the studies included in our meta-analysis
were specifically focused on comparing the magnitude
of liberal and conservative bias; most did not even
14 Ditto et al.
report the relevant data or comparisons. Still, we should
note again that although we can find no evidence that
the strength of liberal bias was underestimated in the
current study, research on partisan bias is naturally
fraught with the potential for that same partisan bias
to influence the research process at multiple levels,
from study design and construction of stimulus materi-
als to the analysis and reporting of relevant data. New
methods being promoted to enhance the reproducibility
of empirical findings in the field of psychology (e.g.,
Cumming, 2014; Simmons, Nelson, & Simonsohn, 2011)
should help combat all forms of research bias, including
those flowing from researchers’ political commitments
(Ditto, Wojcik, Chen, Grady, & Ringel, 2015).
Conclusion
It is common in political discourse to hear politicians
and pundits contrast the biased opinions of their politi-
cal opponents with their own side’s impartial view of
the facts. Our meta-analysis suggests instead that par-
tisan bias is a bipartisan problem and that we may
simply recognize bias in others better than we see it in
ourselves (Pronin, 2007). This same myopia toward our
own side’s biases may also help explain why a field
dominated by liberal researchers has been so much
more focused on the biased perceptions of the political
right than the political left. This meta-analysis raised
more questions than it answered in terms of the specific
determinants of partisan bias, and future research may
suggest that our assessment of the magnitude of bias
in each side may be imprecise (see the confidence and
prediction intervals in Table 2) or historically variable.
What is most clear from the data, however, is that both
liberals and conservatives show a consistent tendency
to be less skeptical consumers of information that sup-
ports than challenges their political beliefs. The fact
that neither side is immune to partisan bias may be the
more important point than whether one side falls prey
to it slightly more than the other.
Using different standards to evaluate information
when it supports your political views than when it
challenges them represents an obvious problem in
terms of normative standards of judgment. Still, it can
be argued that in terms of individual self-interest, a
tendency to adjust one’s political views to fit with
norms of important social or cultural groups makes
good sense (Kahan, 2013). But partisan bias represents
a practical problem as well. It is increasingly clear in
contemporary U.S. politics that liberals and conserva-
tives often hold dramatically different factual beliefs
about key political issues (Frankovic, 2016; Rampell,
2016). The processing biases documented in our meta-
analysis, particularly in conjunction with partisan
selective-exposure effects, are likely to be an important
contributor to these “alternative facts” by leading politi-
cal partisans to readily accept “facts” that support their
side’s positions rather than to carefully scrutinize them.
These differences in factual belief can in turn contribute
to political conflict and governmental dysfunction by
making compromise and negotiation more difficult and
fueling corrosive political stereotypes of the other side
as deluded, hypocritical, or just plain dumb (Ditto &
Liu, 2016; Kennedy & Pronin, 2008).
One solution many in the academy might suggest is
the provision of clear scientific data to provide impartial
answers to disputed questions of fact and to use as a
foundation for evidence-based policy prescriptions. Our
data, however, present a potential obstacle for this pro-
posed solution in that our moderator analyses revealed
that political partisans responded to information com-
posed of scientific data in just as biased a fashion as
they responded to nonscientific arguments. Rather than
being the final arbiter of truth—the impartial political
referee that many people seem to crave—empirical data
may simply provide “grist for a motivated cognitive mill”
(Ditto & Lopez, 1992, p. 579). Together with a growing
body of evidence suggesting that increased knowledge
and expertise in a topic area exacerbates rather than
ameliorates political bias (Kahan etal., 2012; Liu &
Ditto, 2013; Taber & Lodge, 2006), the prognosis for
eradicating partisan bias with harder data and better
education does not seem particularly rosy.7
Sophisticated strategies informed by psychological
science need to be developed to combat our political
prejudices (e.g., Feinberg & Willer, 2013; Fernbach,
Rogers, Fox, & Sloman, 2013) and to begin to build a
less polarized, more civil, and more evidence-based
political culture. The evidence available right now, both
scientific and anecdotal, suggests that this will not be
easy. But a crucial first step is to recognize our collec-
tive vulnerability to perceiving the world in ways that
validate our political affinities.
Action Editor
Brad J. Bushman was the Action Editor for this article.
Author Contributions
All the authors contributed to finding articles, evaluat-
ing whether studies fit inclusion criteria, extracting data,
computing effect sizes, and coding moderators. Analyses
were performed by R. H. Grady and E. E. Chen. J. F.
Zinger served primarily as a meta-analysis consultant. P.
H. Ditto oversaw the entire research project and drafted
the manuscript, and all the authors provided critical revi-
sions. All the authors approved the final manuscript for
submission.
Bias Is Bipartisan 15
Declaration of Conflicting Interests
The author(s) declared that there were no conflicts of interest
with respect to the authorship or the publication of this article.
Supplemental Material
Additional supporting information can be found at http://
journals.sagepub.com/doi/suppl/1745691617746796
Notes
1. The interaction between participants’ party identification and
target of judgment was significant, F(1, 949) = 525.65, p < .001,
η2 = .34.
2. An individual’s prior level of support or opposition to capital
punishment (for example) should not rationally affect how that
individual judges the methodological quality of any particular
study examining the effectiveness of capital punishment, given
that the quality of any specific study is independent of the gen-
eral attitude. On the other hand, prior attitude could plausibly
affect the general level of support or opposition to capital pun-
ishment expressed after exposure to a particular study, even if
no biased judgment occurred. For example, a participant begin-
ning a study opposed to capital punishment might still be more
opposed to capital punishment after reading a study support-
ing it than would a participant beginning the study supporting
capital punishment, simply because the two individuals began
with different attitudes. Thus, a study supporting the efficacy
of capital punishment that led all participants to update their
attitudes about capital punishment to the same degree (i.e.,
no bias) would still leave a capital-punishment opponent with
stronger negative beliefs compared with a capital-punishment
supporter, simply because the former began the study with
more or stronger negative beliefs than the latter. For a simi-
lar but more technical treatment of the rationality of Bayesian
updating in the context of political judgment, see Kahan (2016).
3. If studies did not dichotomize ideological groups, then we
divided the groups according to their ratings, using the scale
midpoint as the cut-off.
4. These are studies that described the behavior of a U.S.
president (e.g., approval of electronic surveillance measure;
Christenson & Kriner, 2016) and manipulated whether the pres-
ident was a Democrat or a Republican.
5. For moderator analyses involving a political topic, we cal-
culated separate rs for each topic regardless of whether the
topic was manipulated between or within subjects. If topic was
manipulated within subjects, the effect size for only one topic
per sample was used in moderator analyses so that responses
from the same participants would not contribute to multiple
effect sizes.
6. Also relevant here is another study reported by Uhlmann
etal. (2009) that examined judgments about the morality of
civilian collateral damage caused by the actions of either the
U.S. military or the Iraqi military. In this case, the judgments of
political liberals were unaffected by the nationality of the per-
petrators, whereas conservatives were significantly more forgiv-
ing when American actions led to unintended civilian deaths
than when Iraqi actions did. This fits well with data showing
that conservatives place greater moral value on loyalty and
patriotism than do liberals (Graham etal., 2013).
7. These findings also suggest another testable explanation for
why the motivational differences between liberals and conser-
vatives do not produce differential patterns of partisan bias.
Liberals’ relative tendency to engage in effortful, system 2 think-
ing (reflected in their higher scores on measures of integrative
complexity, cognitive reflection, and need for cognition; Jost,
2017) may offer them little protection from (and perhaps even
some vulnerability to) biased political judgment (Kahan, 2013,
2016).
References
References marked with an asterisk indicate studies included
in the meta-analysis.
Adorno, T. W., Frenkel-Brunswik, E., Levinson, D. J., &
Sanford, R. N. (1950). The authoritarian personality. New
York, NY: Harper and Row.
Alicke, M. D. (1985). Global self-evaluation as determined
by the desirability and controllability of trait adjectives.
Journal of Personality and Social Psychology, 49, 1621–
1630. doi:10.1037/0022-3514.49.6.1621
Altemeyer, B. (1981). Right-wing authoritarianism. Winnipeg,
Canada: University of Manitoba Press.
Altemeyer, B. (1996). The authoritarian specter. Cambridge,
MA: Harvard University Press.
Bakshy, E., Messing, S., & Adamic, L. (2015). Exposure to
ideologically diverse news and opinion on Facebook.
Science, 348, 1130–1132. doi:10.1126/science.aaa1160
Barberá, P., Jost, J. T., Nagler, J., Tucker, J. A., & Bonneau, R.
(2015). Tweeting from left to right: Is online political
communication more than an echo chamber? Psy cho logi-
cal Science, 26, 1531–1542. doi:10.1177/095679761559
4620
*Bergan, D. E. (2012). Partisan stereotypes and policy atti-
tudes. Journal of Communication, 62, 1102–1120.
doi:10.1111/j.1460-2466.2012.01676.x
Billig, M., & Tajfel, H. (1973). Social categorization and simi-
larity in intergroup behaviour. European Journal of Social
Psychology, 3, 27–52. doi:10.1002/ejsp.2420030103
*Bolsen, T., Druckman, J. N., & Cook, F. L. (2014). The influence
of partisan motivated reasoning on public opinion. Political
Behavior, 36, 235–262. doi:10.1007/s11109-013-9238-0
Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein,
H. R. (2009). Introduction to meta-analysis. Chichester,
England: Wiley.
Brandt, M. J., Reyna, C., Chambers, J. R., Crawford, J. T.,
& Wetherell, G. (2014). The ideological-conflict hypoth-
esis: Intolerance among both liberals and conservatives.
Current Directions in Psychological Science, 23, 27–34.
doi:10.1177/0963721413510932
Brown, J. D., & Kobayashi, C. (2002). Self-enhancement in
Japan and America. Asian Journal of Social Psychology,
5, 145–168. doi:10.1111/1467-839X.00101
*Bullock, J. G. (2011). Elite influence on public opinion in an
informed electorate. American Political Science Review,
105, 496–515. doi:10.1017/S0003055411000165
*Christenson, D. P., & Kriner, D. L. (2017). Constitutional
qualms or politics as usual? The factors shaping public
support for unilateral action. American Journal of Political
Science, 61, 335–349. doi:10.1111/ajps.12262
16 Ditto et al.
*Ciuk, D. J., & Yost, B. A. (2016). The effects of issue salience,
elite influence, and policy content on public opinion.
Political Communication, 33, 328–345. doi:10.1111/
ajps.12262
*Claassen, R. L., & Ensley, M. J. (2016). Motivated reasoning
and yard-sign-stealing partisans: Mine is a likable rogue,
yours is a degenerate criminal. Political Behavior, 38,
317–335. doi:10.1007/s11109-015-9313-9
*Cohen, G. L. (2003). Party over policy: The dominating
impact of group influence on political beliefs. Journal of
Personality and Social Psychology, 85, 808–822. doi:10
.1037/0022-3514.85.5.808
Collins, T. P., Crawford, J. T., & Brandt, M. J. (2017). No
evidence for ideological asymmetry in dissonance avoid-
ance. Social Psychology, 48, 123–134. doi:10.1027/1864-
9335/a000300
*Crawford, J. T. (2012). The ideologically objectionable prem-
ise model: Predicting biased political judgments on the
left and right. Journal of Experimental Social Psychology,
48, 138–151. doi:10.1016/j.jesp.2011.10.004
Crawford, J. T. (2014). Ideological symmetries and asymme-
tries in political intolerance and prejudice toward political
activist groups. Journal of Experimental Social Psychology,
55, 284–298. doi:10.1016/j.jesp.2014.08.002
Crawford, J. T. (2017). Are conservatives more sensitive
to threat than liberals? It depends on how we define
threat and conservatism. Social Cognition, 35, 354–373.
doi:10.1521/soco.2017.35.4.354
*Crawford, J. T., Jussim, L., Cain, T. R., & Cohen, F. (2013).
Right-wing authoritarianism and social dominance orien-
tation differentially predict biased evaluations of media
reports. Journal of Applied Social Psychology, 43, 163–74.
doi:10.1111/j.1559-1816.2012.00990.x
*Crawford, J. T., Kay, S. A., & Duke, K. E. (2015). Speaking
out of both sides of their mouths: Biased political
judgments within (and between) individuals. Social
Psychological & Personality Science, 6, 422–430. doi:
10.1177/1948550614566858
*Crawford, J. T., Wiley, S., & Ventresco, N. (2014). Examining
Americans’ attitudes toward drone strikes on the eve of
the 2012 presidential election. Analyses of Social Issues
and Public Policy, 14, 46–60. doi:10.1111/asap.12030
*Crawford, J. T., & Xhambazi, E. (2015). Predicting political
biases against the Occupy Wall Street and Tea Party move-
ments. Political Psychology, 36, 111–121. doi:10.1111/
pops.12054
Cumming, G. (2014). The new statistics: Why and how.
Psychological Science, 25, 7–29. doi:10.1177/09567976
13504966
Darley, J. M., & Gross, P. H. (1983). A hypothesis-confirming
bias in labeling effects. Journal of Personality and Social
Psychology, 44, 20–33. doi:10.1037/0022-3514.44.1.20
Ditto, P. H. (2009). Passion, reason, and necessity: A quantity
of processing view of motivated reasoning. In T. Bayne &
J. Fernandez (Eds.), Delusion, self-deception, and affective
influences on belief formation (pp. 23–53). New York,
NY: Psychology Press.
Ditto, P. H., & Liu, B. S. (2016). Moral coherence and politi-
cal conflict. In P. Valdesolo & J. Graham (Eds.), Social
psychology of political polarization (pp. 102–122). New
York, NY: Routledge.
Ditto, P. H., & Lopez, D. F. (1992). Motivated skepticism: Use
of differential decision criteria for preferred and non-
preferred conclusions. Journal of Personality and Social
Psychology, 63, 568–584. doi:10.1037/0022-3514.63.4.568
Ditto, P. H., Munro, G. D., Apanovich, A. M., Scepansky,
J. A., & Lockhart, L. K. (2003). Spontaneous skepti-
cism: The interplay of motivation and expectation in
responses to favorable and unfavorable medical diag-
noses. Personality and Social Psychology Bulletin, 29,
1120–1132. doi:10.1177/0146167203254536
Ditto, P. H., Pizarro, D. A., & Tannenbaum, D. (2009).
Motivated moral reasoning. In D. M. Bartels, C. W.
Bauman, L. J. Skitka, & D. L. Medin (Eds.), The psychol-
ogy of learning and motivation (Vol. 50, pp. 307–338).
Burlington, MA: Academic Press. doi:10.1016/S0079-7421
(08)00410-6
Ditto, P. H., Wojcik, S. P., Chen, E. E., Grady, R. H., & Ringel,
M. (2015). Political bias is tenacious. Behavioral & Brain
Sciences, 38, Article e140. doi:10.1017/S0140525X14001186
*Druckman, J. N. (2001). Using credible advice to overcome
framing effects. Journal of Law, Economics, & Organi-
zation, 17, 62–82. doi:10.1093/jleo/17.1.62
Duarte, J. L., Crawford, J. T., Stern, C., Haidt, J., Jussim, L., &
Tetlock, P. E. (2015). Political diversity will improve social
psychological science. Behavioral & Brain Sciences, 38,
Article e130. doi:10.1017/S0140525X14000430
Duncan, B. L. (1976). Differential social perception and attri-
bution of intergroup violence: Testing the lower limits of
stereotyping of Blacks. Journal of Personality and Social
Psychology, 34, 590–598. doi:10.1037/0022-3514.34.4.590
Eidelman, S., Crandall, C. S., Goodman, J. A., & Blanchar, J. C.
(2012). Low-effort thought promotes political conser-
vatism. Personality and Social Psychology Bulletin, 38,
808–820. doi:10.1177/0146167212439213
Erdelyi, M. H. (1974). A new look at the new look: Perceptual
defense and vigilance. Psychological Review, 81, 1–25.
doi:10.1037/h0035852
Federico, C. M., & Malka, A. (2018). The contingent, contextual
nature of the relationship between needs for security and
certainty and political preferences: Evidence and impli-
cations. Advances in Political Psychology, 39(Suppl. 1),
3–48. doi:10.1111/pops.12477
Feinberg, M., & Willer, R. (2013). The moral roots of envi-
ronmental attitudes. Psychological Science, 24, 56–62.
doi:10.1177/0956797612449177
Fernbach, P. M., Rogers, T., Fox, C. R., & Sloman, S. A.
(2013). Political extremism is supported by an illusion
of understanding. Psychological Science, 24, 939–946.
doi:10.1177/0956797612464058
Fessler, D. M. T., Pisor, A. C., & Holbrook, C. (2017).
Political orientation predicts credulity regarding puta-
tive hazards. Psychological Science, 28, 651–660. doi:
10.1177/0956797617692108
Frankovic, K. (2016, December 27). Belief in conspiracies
largely depends on political identity. YouGov. Retrieved
from https://today.yougov.com/news/2016/12/27/belief-
conspiracies-largely-depends-political-iden/
Bias Is Bipartisan 17
Frenda, S. J., Knowles, E. D., Saletan, W., & Loftus, E. F.
(2013). False memories of fabricated political events.
Journal of Experimental Social Psychology, 49, 280–286.
doi:10.1016/j.jesp.2012.10.013
Frimer, J. A., Skitka, L. J., & Motyl, M. (2017). Liberals and
conservatives are similarly motivated to avoid exposure
to one another’s opinions. Journal of Experimental Social
Psychology, 72, 1–12. doi:10.1016/j.jesp.2017.04.003
*Furgeson, J. R., Babcock, L., & Shane, P. M. (2008a). Behind
the mask of method: Political orientation and constitu-
tional interpretive preferences. Law and Human Behavior,
32, 502–510. doi:10.1007/s10979-007-9112-x
*Furgeson, J. R., Babcock, L., & Shane, P. M. (2008b). Do
a law’s policy implications affect beliefs about its con-
stitutionality? An experimental test. Law and Human
Behavior, 32, 219–227. doi:10.1007/s10979-007-9102-z
Garrett, R. K., & Stroud, N. J. (2014). Partisan paths to expo-
sure diversity: Differences in pro- and counterattitudi-
nal news consumption. Journal of Communication, 64,
680–701. doi:10.1111/jcom.12105
Graham, J., Haidt, J., Koleva, S., Motyl, M., Iyer, R., Wojcik, S. P.,
& Ditto, P. H. (2013). Moral foundations theory: The pra-
gmatic validity of moral pluralism. In P. Devine & A.
Plant (Eds.), Advances in Experimental Social Psychology
(Vol. 47, pp. 55–130). San Diego, CA: Academic Press.
doi:10.1016/B978-0-12-407236-7.00002-4
*Groenendyk, E. (2013). Competing motives in the parti-
san mind: How loyalty and responsiveness shape party
identification and democracy. Oxford Scholarship Online.
doi:10.1093/acprof:oso/9780199969807.001.0001.
Haidt, J. (2001). The emotional dog and its rational tail: A social
intuitionist approach to moral judgment. Psychological
Review, 108, 814–834. doi:10.1037/0033-295X.108.4.814
Hastorf, A., & Cantril, H. (1954). They saw a game: A case
study. Journal of Personality and Social Psychology, 49,
129–134. doi:10.1037/h0057880
*Hawkins, C. B., & Nosek, B. A. (2012). Motivated inde-
pendence? Implicit party identity predicts politi-
cal judgments among self-proclaimed independents.
Personality and Social Psychology Bulletin, 38, 1437–1452.
doi:10.1177/0146167212452313
Hibbing, J. R., Smith, K. B., & Alford, J. R. (2014). Differences
in negativity bias underlie variations in political ideology.
Behavioral & Brain Sciences, 37, 297–307. doi:10.1017/
S0140525X13001192
Horowitz, D. (2008, December 8). Get over Obama derange-
ment syndrome. Politico. Retrieved from http://www
.politico.com/story/2008/12/get-over-obama-derange-
ment-syndrome-016306
Huffington, A. (2002, August 16). Wacko in Waco. Salon. Retrieved
from http://www.salon.com/2002/08/16/waco_13/
Hume, D. (1985). A treatise of human nature. London,
England: Penguin. (Original work published 1740)
Inbar, Y., & Lammers, J. (2012). Political diversity in social
and personality psychology. Perspectives on Psychological
Science, 7, 496–503. doi:10.1177/1745691612448792
Iyengar, S., & Hahn, K. S. (2009). Red media, blue media:
Evidence of ideological selectivity in media use. Journal
of Communication, 59, 19–39. doi:10.1111/j.1460-
2466.2008.01402.x
Iyer, R., Koleva, S., Graham, J., Ditto, P. H., & Haidt, J. (2012).
Understanding libertarian morality: The psychological dis-
positions of self-identified libertarians. PLOS ONE, 7(8),
Article e42366. doi:10.1371/journal.pone.0042366
Jost, J. T. (2017). Ideological asymmetries and the essence of
political psychology. Political Psychology, 38, 167–208.
doi:10.1111/pops.12407
Jost, J. T., Glaser, J., Kruglanski, A. W., & Sulloway, F. (2003).
Political conservatism as motivated social cognition. Psy-
chological Bulletin, 129, 339–375. doi:10.1037/0033-
2909.129.3.339
*Kahan, D. M. (2013). Ideology, motivated reasoning, and cog-
nitive reflection. Judgment and Decision Making, 8, 407–
424. Retrieved from http://journal.sjdm.org/13/13313/
jdm13313.pdf
Kahan, D. M. (2016). The politically motivated reasoning par-
adigm, part 1: What politically motivated reasoning is and
how to measure it. In R. A. Scott, S. M. Kosslyn, & M. C.
Buchmann (Eds.), Emerging Trends in the Social and
Behavioral Sciences: An Interdisciplinary, Searchable, and
Linkable Resource. doi:10.1002/9781118900772.etrds0417
*Kahan, D. M., Hoffman, D. A., Braman, D., & Evans, D.
(2012). They saw a protest: Cognitive illiberalism and
the speech-conduct distinction. Stanford Law Review, 64,
851–906. Retrieved from http://digitalcommons.law.yale
.edu/fss_papers/4692
*Kahan, D. M., Hoffman, D., Evans, D., Devins, N., Lucci, E.,
& Cheng, K. (2016). “Ideology” or “situation sense”? An
experimental investigation of motivated reasoning and
professional judgment. University of Pennsylvania Law
Review, 164(2), 349–440. Retrieved from https://scholar
ship.law.upenn.edu/penn_law_review/vol164/iss2/2
*Kahan, D. M., Jamieson, K. H., Landrum, A., & Winneg, K.
(2017). Culturally antagonistic memes and the Zika virus:
An experimental test. Journal of Risk Research, 20, 1–40.
doi:10.1080/13669877.2016.1260631
*Kahan, D. M., Jenkins-Smith, H., & Braman, D. (2011).
Cultural cognition of scientific consensus. Journal of Risk
Research, 14, 147–174. doi:10.1080/13669877.2010.511246
*Kahan, D. M., Peters, E., Dawson, E. C., & Slovic, P. (2017).
Motivated numeracy and enlightened self-government.
Behavioural Public Policy, 1, 54–86. doi:10.1017/
bpp.2016.2
Kahan, D. M., Peters, E., Wittlin, M., Slovic, P., Ouellette, L. L.,
Braman, D., & Mandel, G. N. (2012). The polarizing
impact of science literacy and numeracy on perceived
climate change risks. Nature Climate Change, 2, 732–735.
doi:10.1038/nclimate1547
Kahneman, D., & Tversky, A. (1984). Choices, values,
and frames. American Psychologist, 39, 341–350. doi:
10.1037/0003-066X.39.4.341
*Kam, C. D. (2005). Who toes the party line? Cues, values, and
individual differences. Political Behavior, 27, 163–182.
doi:10.1007/s11109-005-1764-y
Kennedy, K. A., & Pronin, E. (2008). When disagreement gets
ugly: Perceptions of bias and the escalation of conflict.
18 Ditto et al.
Personality and Social Psychology Bulletin, 34, 833–848.
doi:10.1177/0146167208315158
Keys, D. J., & Schwartz, B. (2007). “Leaky” rationality: How
research on behavioral decision making challenges norma-
tive standards of rationality. Perspectives on Psychological
Science, 2, 162–180. doi:10.1111/j.1745-6916.2007.00035.x
Knobloch-Westerwick, S., & Kleinman, S. B. (2012).
Preelection selective exposure: Confirmation bias ver-
sus informational utility. Communication Research, 39,
170–193. doi:10.1177/0093650211400597
Knobloch-Westerwick, S., & Meng, J. (2009). Looking the
other way: Selective exposure to attitude-consistent and
counterattitudinal political information. Communication
Research, 36, 426–448. doi:10.1177/0093650209333030
*Kopko, K. C., Mckinnon, S., Budziak, J., Devine, C. J., &
Nawara, S. P. (2011). In the eye of the beholder? Motivated
reasoning in disputed elections. Political Behavior, 33,
271–290. doi:10.1007/s11109-010-9133-x
Krauthammer, C. (2003, December 5). Bush derangement
syndrome. Townhall. Retrieved from http://townhall
.com/columnists/charleskrauthammer/2003/12/05/bush_
derangement_syndrome
Krosnick, J. A. (1988). The role of attitude importance in
social evaluation: A study of policy preferences, presi-
dential candidate evaluations, and voting behavior.
Journal of Personality and Social Psychology, 55, 196–210.
doi:10.1037//0022-3514.55.2.196
Kunda, Z. (1990). The case for motivated reasoning.
Psychological Bulletin, 108, 480–498. doi:10.1037/0033-
2909.108.3.480
*Lai, C. K., & Nosek, B. A. (2012). Party over policy (and race).
Unpublished data set.
Leeper, T. J., & Slothuus, R. (2014). Political parties, motivated
reasoning, and public opinion formation. Advances in
Political Psychology, 35, 129–156. doi:10.1111/pops.12164
Lerner, J. S., & Tetlock, P. E. (1999). Accounting for the effects
of accountability. Psychological Bulletin, 125, 255–275.
doi:10.1037/0033-2909.125.2.255
Lilienfeld, S. O., & Latzman, R. D. (2014). Threat bias, not neg-
ativity bias, underpins differences in political ideology.
Behavioral & Brain Sciences, 37, 318–319. doi:10.1017/
S0140525X1300263X
*Liu, B. S. (2014). The expertise paradox: Examining the role
of different aspects of expertise in biased evaluation of sci-
entific information (Unpublished doctoral dissertation).
University of California, Irvine.
Liu, B. S., & Ditto, P. H. (2013). What dilemma? Moral evaluation
shapes factual belief. Social Psychological & Personality
Science, 4, 316–323. doi:10.1177/1948550612456045
*Lopez, D. F. (1994). The affective mediation of expectancy
processes (Unpublished doctoral dissertation). Kent State
University, Kent, OH.
*Lord, C. G., Ross, L., & Lepper, M. R. (1979). Biased assimi-
lation and attitude polarization: The effects of prior
theories on subsequently considered evidence. Journal
of Personality and Social Psychology, 37, 2098–2109.
doi:10.1037/0022-3514.37.11.2098
*MacCoun, R. J., & Paletz, S. (2009). Citizens’ perceptions of
ideological bias in research on public policy controversies.
Political Psychology, 30, 43–65. doi:10.1111/j.1467-
9221.2008.00680.x
*Malka, A., & Lelkes, Y. (2010). More than ideology: Conservative–
liberal identity and receptivity to political cues. Social Justice
Research, 23, 156–188. doi:10.1007/s11211-010-0114-3
Malka, A., & Soto, C. J. (2015). Rigidity of the economic
right? Menu-independent and menu-dependent influ-
ences of psychological dispositions on political attitudes.
Current Directions in Psychological Science, 24, 137–142.
doi:10.1177/0963721414556340
Mercier, H., & Sperber, D. (2011). Why do humans reason?
Arguments for an argumentative theory. Behavioral & Brain
Sciences, 34, 57–111. doi:10.1017/S0140525X10000968
Messing, S., & Westwood, S. J. (2014). Selective exposure in the
age of social media: Endorsements trump partisan source
affiliation when selecting news online. Communication
Research, 41, 1042–1063. doi:10.1177/0093650212466406
Miller, D. T., & Ross, M. (1975). Self-serving biases in the attri-
bution of causality: Fact or fiction? Psychological Bulletin,
82, 213–225. doi:10.1037/h0076486
Mischel, W. (1977). The interaction of person and situation.
In D. Magnusson & N. S. Endler (Eds.), Personality at the
crossroads: Current issues in interactional psychology (pp.
333–352). Hillsdale, NJ: Erlbaum.
Mullen, E., & Skitka, L. J. (2006). Exploring the psycho-
logical underpinnings of the moral mandate effect:
Motivated reasoning, group differentiation, or anger?
Journal of Personality and Social Psychology, 90, 629–643.
doi:10.1037/0022-3514.90.4.629
*Mullinix, K. J. (2016). Partisanship and preference forma-
tion: Competing motivations, elite polarization, and issue
importance. Political Behavior, 38, 383–411. doi:10.1007/
s11109-015-9318-4
*Munro, G. D., & Munro, C. A. (2014). “Soft” versus “hard”
psychological science: Biased evaluations of scientific evi-
dence that threatens or supports a strongly held political
identity. Basic and Applied Social Psychology, 36, 533–
543. doi:10.1080/01973533.2014.960080
Nam, H. H., Jost, J. T., & Van Bavel, J. J. (2013). “Not for all
the tea in China!” Political ideology and the avoidance of
dissonance-arousing situations. PLOS ONE, 8(4), Article
e59837. doi:10.1371/journal.pone.0059837
*Nawara, S. P. (2011). The formation of responsibility attributions
and their role in shaping political behavior (Unpublished
doctoral dissertation). Ohio State University, Columbus.
Neuberg, S. L., & Fiske, S. T. (1987). Motivational influ-
ences on impression formation: Outcome dependency,
accuracy-driven attention, and individuating processes.
Journal of Personality and Social Psychology, 53, 431–444.
doi:10.1037/0022-3514.53.3.431
Nisbett, R. E., & Ross, L. (1980). Human inference: Strategies
and shortcomings of social judgment. Englewood Cliffs,
NJ: Prentice-Hall.
Norton, M. I., Vandello, J. A., & Darley, J. M. (2004). Casuistry
and social category bias. Journal of Personality and Social
Psychology, 87, 817–831. doi:10.1037/0022-3514.87.6.817
O’Reilly, B. (2005, September 9). Feeling sorry for O’Reilly.
Fox News. Retrieved from http://www.foxnews.com/
story/2005/09/09/feeling-sorry-for-oreilly.html
Bias Is Bipartisan 19
Pomerantz, E. M., Chaiken, S., & Tordesillas, R. (1995). Attitude
strength and resistance processes. Journal of Personality
and Social Psychology, 69, 408–419. doi:10.1037/0022-
3514.69.3.408
Pronin, E. (2007). Perception and misperception of bias in
human judgment. Trends in Cognitive Sciences, 11, 37–43.
doi:10.1016/j.tics.2006.11.001
Raimondo, J. (2016, December 27). Do you suffer from Trump
Derangement Syndrome? Los Angeles Times. Retrieved
from http://www.latimes.com/opinion/op-ed/la-oe-rai
mondo-trump-derangement-syndrome-20161226-story
.html
Rampell, C. (2016, December 28). Americans—especially
but not exclusively Trump voters—believe crazy, wrong
things. The Washington Post. Retrieved from https://www
.washingtonpost.com/news/rampage/wp/2016/12/28/
americans-especially-but-not-exclusively-trump-voters-
believe-crazy-wrong-things
Rodriguez, C. G., Moskowitz, J. P., Salem, R. M., & Ditto, P. H.
(2017). Partisan selective exposure: The role of party, ide-
ology and ideological extremity over time. Translational
Issues in Psychological Science, 3, 254–271. doi:10.1037/
tps0000121
Rokeach, M. (1956). Political and religious dogmatism: An
alternative to the authoritarian personality. Psychological
Monographs: General and Applied, 70(18), 1–43.
doi:10.1037/h0093727
*Scurich, N., & Shniderman, A. (2014). The selective allure
of neuroscientific explanations. PLOS ONE, 9(9), Article
e107529. doi:10.1371/journal.pone.0107529
Sedikides, C., Gaertner, L., & Vevea, J. L. (2005). Pancultural
self-enhancement reloaded: A meta-analytic reply to Heine
(2005). Journal of Personality and Social Psychology, 89,
539–551. doi:10.1037/0022-3514.89.4.539
Shils, E. (1954). Authoritarianism: “Right” and “Left.” In R.
Christie & M. Jahoda (Eds.), Studies in the scope and
method of “the authoritarian personality.” (pp. 24-49).
Glencoe, IL: Free Press.
Shook, N. J., & Fazio, R. H. (2009). Political ideology,
exploration of novel stimuli, and attitude formation.
Journal of Experimental Social Psychology, 45, 995–998.
doi:10.1016/j.jesp.2009.04.003
Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2011).
False-positive psychology: Undisclosed flexibility in
data collection and analysis allows presenting anything
as significant. Psychological Science, 22, 1359–1366.
doi:10.1177/0956797611417632
Skitka, L. J., Bauman, C. W., & Sargis, E. G. (2005). Moral con-
viction: Another contributor to attitude strength or some-
thing more? Journal of Personality and Social Psychology,
88, 895–917. doi:10.1037/0022-3514.88.6.895
*Smith, B. K. (2014). Cross-cutting concerns: The varying
effects of partisan cues in the context of social networks
(Master’s thesis). Portland State University, Portland, OR.
doi:10.15760/etd.1951
*Smith, C. T., Ratliff, K. A., & Nosek, B. A. (2012). Rapid
assimilation: Automatically integrating new informa-
tion with existing beliefs. Social Cognition, 30, 199–219.
doi:10.1521/soco.2012.30.2.199
Stanovich, K. E., West, R. F., & Toplak, M. E. (2013).
Myside bias, rational thinking, and intelligence. Current
Directions in Psychological Science, 22, 259–264.
doi:10.1177/0963721413480174
Sterne, J. A. C., Becker, B. J., & Egger, M. (2005). The fun-
nel plot. In H. R. Rothstein, A. J. Sutton, & M. Bornstein
(Eds.), Publication bias in meta-analysis: Prevention,
assessment, and adjustments (pp. 75–98). Hoboken, NJ:
Wiley.
Stroud, N. J. (2008). Media use and political predispositions:
Revisiting the concept of selective exposure. Political
Behavior, 30, 341–366. doi:10.1007/s11109-007-9050-9
Taber, C. S., & Lodge, M. (2006). Motivated skepticism in
the evaluation of political beliefs. American Journal
of Political Science, 50, 755–769. doi:10.1111/j.1540-
5907.2006.00214.x
*Tannenbaum, D., Fox, C. R., & Rogers, T. (2014). On the
misplaced politics of behavioural policy interventions.
Nature Human Behaviour, 1, Article 0130. doi:10.1038/
s41562-017-0130
Taylor, I. A. (1960). Similarities in the structure of extreme
social attitudes. Psychological Monographs: General and
Applied, 74(2), 1–36. doi:10.1037/h0093754
Tetlock, P. E., & Levi, A. (1982). Attribution bias: On the
inconclusiveness of the cognition-motivation debate.
Journal of Experimental Social Psychology, 18, 68–88.
doi:10.1016/0022-1031(82)90082-8
Tversky, A., & Kahneman, D. (1981). The framing of decisions
and the psychology of choice. Science, 211, 453–458.
doi:10.1126/science.7455683
Uhlmann, E. L., Pizarro, D. A., Tannenbaum, D., & Ditto, P. H.
(2009). The motivated use of moral principles. Judgment
and Decision Making, 4, 476–491.
Von Neumann, J., & Morgenstern, O. (1944). Theory of games
and economic behavior. New York, NY: Wiley.
Wojcik, S. P., Hovasapian, A., Graham, J., Motyl, M., & Ditto,
P. H. (2015). Conservatives report, but liberals display,
greater happiness. Science, 347, 1243–1246. doi:10.1126/
science.1260817
Zuwerink, J. R., & Devine, P. G. (1996). Attitude importance
and resistance to persuasion: It’s not just the thought that
counts. Journal of Personality and Social Psychology, 70,
931–944. doi:10.1037/0022-3514.70.5.931
... More recent studies have, therefore, relied on more stringent measures and designs. Challenging the intrinsic thesis, a meta-analysis of such studies revealed that liberals and conservatives are equally biased in their cognition of counterattitudinal information (Ditto et al., 2019). As an alternative, the contextual thesis proposes that the association between political leaning and science rejection is driven by an interplay between citizens' political identity and the politicization of science by elites (Newman et al., 2018;Nisbet et al., 2015). ...
... Finally, this examination established that a four-level model of polarization can expose ideological symmetries in science rejection that have previously remained largely undetected in observational studies. So far, the literature has been characterized by a paradox that, on the one hand, experimental studies have generally pointed out that liberals and conservatives are equally biased against counterattitudinal information (Ditto et al., 2019) while, on the other hand, observational research has quite consistently demonstrated that liberals have more trust in scientific claims (e.g. Hunt and Wald, 2020)-as well as more accurate beliefs (e.g. ...
Article
Full-text available
Citizens’ trust in science increasingly depends on their political leaning. Structural equation models on survey data from 10 European countries (N = 5,306) demonstrate that this science polarization can be captured by a model with four levels of generalization. Voters of populist parties distrust the system and elite in general, which indirectly fuels a broad science skepticism. At another level, right-wingers have less trust in science as a whole than left-wingers. After accounting for this general skepticism, left-wingers and right-wingers are, however, similarly prone to contest ideology-incongruent research fields and specific claims. These findings have three implications. First, research on science skepticism should carefully consider all four levels and their interplay. Second, the science polarization between populist and non-populist voters has fundamentally different origins than the effect of left-right ideology. Third, a four-level model can expose ideological symmetries in science rejection that have previously remained largely undetected in observational studies.
... When forming opinions and views about new information, people tend to engage in biased reasoning and information processing, motivated to align with pre-existing beliefs 8 . Numerous studies supported the notion that this tendency is not directed towards reaching accurate conclusions, but rather towards confirming and preserving desired views or ideological goals, for example positive views of one's political ingroup [34][35][36][37][38] . Motivated reasoning can exacerbate polarization, as it undermines processing of information that contradicts existing beliefs 39 and cements negative views of outgroups 40 . ...
Article
Full-text available
This work explored polarization over Israel’s Judicial Reform, introduced in January 2023. We find that the reform divided people into pro- and anti-reform camps, which differed in characteristics such as institutional trust, patriotism, and national identity. For example, the camps disagreed about trust in the government versus the judiciary. In line with motivated reasoning—biased reasoning processes used to reach desired conclusions—people’s pre-existing characteristics motivated polarized views of the reform as a threat to democracy (issue-based polarization) and negative emotions towards opponents (affective polarization). Further demonstrating a motivated process, pro-reform participants (the electorate majority), prioritized majority rule over other democratic features (e.g., minority rights) compared to anti-reform participants. Polarization differentially predicted downstream consequences (e.g., protest methods), indicating that the camps’ reactions were motivated by the extremity of their views and negative emotions. This work extends the understanding of potentially motivated polarization processes and their immediate downstream consequences.
... The problem is that the experiments do not allow us to tell. This kind of confound arguably afflicts experimental research on (politically) motivated reasoning in general and has been called "an empirical catch-22 at [its] heart"(Ditto et al. 2018; p. 285).Content courtesy of Springer Nature, terms of use apply. Rights reserved. ...
Article
Full-text available
In a series of very influential papers, Dan Kahan argues for “the identity protective cognition thesis”: the claim that politically motivated reasoning is a major factor explaining current levels of polarization over matters of fact, especially in the US. An important part of his case consists of experimental data supporting the claim that ideological polarization is more extreme amongst more numerate individuals. In this paper, we take a close look at how precisely this “numeracy effect” is supposed to come about. Working with Kahan’s own notion of motivated reasoning, we reconstruct the mechanism that according to him produces the effect. Surprisingly, it turns out to involve plenty of motivation to reason, but no motivated reasoning. This undermines the support he takes the numeracy effect to provide for the identity protective cognition hypothesis.
... They prefer a more limited governmental role, suggesting that private efforts and charities effectively solve social issues (Frimer, Skitka, & Motyl, 2017;Narveson, 2000). Conservatives believe that such programs reduce incentives for personal responsibility and self-sufficiency, advocating for an environment where individuals can choose their services without compulsory government involvement (Ditto et al., 2019). This ideological disparity underscores the fundamental differences in how liberals and conservatives envision the balance between government assistance and personal freedom in addressing social challenges. ...
Article
Full-text available
This research examines the factors that influence health-related stereotypes against immigrants, with a focus on how human capital, conservative ideology, and economic anxiety impact these perceptions across different immigrant groups in the United States. The study uses Ordinary Least Squares (OLS) regression analysis of data from the 2020 Collaborative Multi-racial Post-Election Survey to assess attitudes towards all immigrants, as well as specifically Mexican and Chinese immigrants. Objective: Health stereotypes often depict immigrants as carriers of diseases and burdens on society, which affect public opinion and policy. This study aims to analyze the roles of human capital, political ideology, and economic conditions in shaping these stereotypes. Method: Using OLS regression, the analysis explores the association between stereotyping and various predictors, including conservative ideology (measured objectively and subjectively), human capital, and economic anxiety. Disaggregating the data by immigrant group allows for nuanced insights into the interplay of these factors. Results: The results reveal moderate levels of stereotyping across all groups, with variations by immigrant origin. Objective measures of conservative ideology show a strong correlation with increased stereotyping, particularly against Mexican immigrants. Higher human capital is associated with reduced stereotyping, while economic anxiety consistently exacerbates negative stereotypes across all groups. Furthermore, Republican affiliation and racial identity are correlated with heightened stereotyping, especially against Chinese immigrants. Discussion: The findings highlight the complexity of the factors contributing to health-related stereotypes and emphasize the importance of addressing educational and economic elements to mitigate these biases. Tailored interventions that take into account the distinct dynamics of each immigrant group are crucial for reducing stereotypes and improving integration and health outcomes.
... While previous research has found no difference between liberals and conservatives in terms of partisan bias (Ditto et al., 2019), there is a moderate correlation between state-level tightness-looseness and statelevel conservative ideology (Harrington & Gelfand, 2014). To exclude conservatism as an alternative mechanism driving our results, we included the state-level "Conservative Advantage" measure provided by Gallup. ...
Article
Full-text available
Do politicians' relational traits predict their bipartisan voting behavior? In this paper, we empirically test and find that relational individual dispositions, namely attachment orientations and conformity to cultural norms, can predict the bipartisan voting behavior of politicians in the United States House of Representatives and Senate. We annotated politicians' tweets using a machine learning approach paired with archival resources to obtain politicians' home-state looseness-tightness culture scores. Anxiously-attached politicians were less likely to be bipartisan than avoidantly-attached individuals. Bipartisan voting behavior was less likely in politicians whose home state was less tolerant of deviation from cultural norms. We discuss these results and possible implications, such as the preemptive assessment of politicians' bipartisanship likelihood based on attachment and state cultural pressure to adhere to group norms.
... The present investigation extended the existing literature by examining a wide range of possible patterns between political orientation and knowledge. One reason why prior research may have narrowly focused on determining whether the association between political orientation and knowledge is merely linear or quadratic is that the prevalent discourse in the social sciences centers around questions of ideological symmetry or asymmetry, seeking to understand whether left-wing and right-wing individuals differ, for example, in partisan biases 25 , moral judgments 26 , and belief superiority 27 . Yet this lack of scholarly attention on people at the political center is unwarranted given that many citizens hold moderate views, and these citizens are especially important for candidate selection 11 . ...
Article
Full-text available
Political knowledge is crucial for well-functioning democracies, with most scholars assuming that people at the political extremes are more knowledgeable than those at the center. Here, we adopt a data-driven approach to examine the relationship between political orientation and political knowledge by testing a series of polynomial curves in 45 countries (N = 63,544), spread over 6 continents. Contrary to the dominant perspective, we found no evidence that people at the political extremes are the most knowledgeable about politics. Rather, the most common pattern was a fourth-degree polynomial association in which those who are moderately left-wing and right-wing are more knowledgeable than people at the extremes and center of the political spectrum. This pattern was especially, though not exclusively, prevalent in Western countries. We conclude that the relationship between political orientation and political knowledge is more context-dependent and complex than assumed, and caution against (implicit) universal conclusions in social sciences.
Article
Full-text available
Political tribalism has increased dramatically in recent years. We explored partisan double-standards of Democratic and Republican voters across both hypothetical and real-worldscenarios. In Study 1, participants rated the perceived legitimacy of election outcomes inresponse to hypothetical and ambiguous results from the 2020 presidential election. In Study 2Part 1, college students and Amazon Turk volunteers rated their support of real-world presidential policies and actions. All policies/actions were attributed to Trump or Obama though they actually occurred under both presidents. In Study 2 Part 2, participants rated how bigoted various statements were; we manipulated who the utterances were attributed to (Trump v. Bill Clinton or Trump v. Martin Luther King [MLK]). Generally, Republican ratings were more favorable when statements were attributed to Trump vs. Democratic leaders while the opposite is true of Democrats. Crucially, these biases exist when evaluating identical information. Republicans and Democrats had a very small and very large tendency, respectively, to view statements as more bigoted under Trump vs. MLK. To the degree that this study can answer the question about which side is more guilty of double-standards, our results provide tentative evidence that this occurs under Democrats more than Republicans, though this overall difference may obscure important moderators. Our data provide evidence for tribal loyalty which may have significant social and political ramifications.Keywords: Tribalism, politics, bias, motivated reasoning, partisanship
Article
Full-text available
Our moral sense is a biological adaptation specific to our species. It has evolved to facilitate group living, regulate social interactions, and promote cooperation beyond kinship. Beliefs intensely associated with moral values can motivate positive societal progress. However, moral convictions can also lead to dogmatism, antagonistic attitudes, a polarized political climate, and violent collective actions. This article draws on theories and empirical evidence from anthropology, neuroscience, cognitive science, social psychology, economics, and political science to explain why and how moralization increases the strength of beliefs and attitudes –in certainty and importance. Moral convictions are perceived as absolute, universal, and definite beliefs or principles. They engage the brain circuit associated with reward processing and valuation, which motivates and reinforces behavior. Moral commitments act as signals to conform to the rules of one’s coalition, which causes predictable cognitive biases that benefit the self and the group. Moral convictions are held with great confidence. This leads to confirmation bias, which means that people are unwilling to seek or accept corrective information no matter how accurate the content of their beliefs. Once an issue is moralized, it becomes more challenging to engage in a cost-benefit analysis. A deeper understanding of attitudes, values, and moralization dynamics may allow social scientists and policymakers to identify and support pathways away from extremism and to strengthen and sustain engagement in benevolent forms of collective action.
Preprint
Full-text available
Many scholars conceive modern Western universities as places for open inquiry and relentless pursuit of truth. Yet in recent years, some scholars have expressed concerns about increasing censoriousness on college campuses. The present investigation tested whether people have heightened desires to censor information on campuses that is perceived as threatening to group equality or reinforcing of status hierarchies—specifically, information that portrays low-status groups unfavorably. Across four samples from three countries (U.S. adults, U.S. college-aged adults, undergraduates at a U.K. university, and undergraduates at a Hungarian university; total n=1,616) and three domains of group differences, we found that people were more censorious of statements that portray lower status groups unfavorably (women, Black people, Muslims) than otherwise identical statements that portray higher status groups unfavorably (men, White people, Christians). We also found that these differential standards in censorship preferences increased as participants self-identified as more politically liberal, perhaps reflecting Liberals’ greater aversion to inequality and protectiveness toward low-status groups. Such patterns (especially in conjunction with other recent work finding similar patterns) challenge the conventional wisdom that evaluative biases generally _harm_ low-status groups and _reinforce_ existing hierarchies. Our results suggest instead that, at least in recent years in modern Western societies, biases in information evaluations seem designed to _help_ low-status groups and to _eliminate_ or _reverse _existing hierarchies.
Article
Full-text available
Research on moral judgment has been dominated by rationalist models, in which moral judgment is thought to be caused by moral reasoning. The author gives 4 reasons for considering the hypothesis that moral reasoning does not cause moral judgment; rather, moral reasoning is usually a post hoc construction, generated after a judgment has been reached. The social intuitionist model is presented as an alternative to rationalist models. The model is a social model in that it deemphasizes the private reasoning done by individuals and emphasizes instead the importance of social and cultural influences. The model is an intuitionist model in that it states that moral judgment is generally the result of quick, automatic evaluations (intuitions). The model is more consistent than rationalist models with recent findings in social, cultural, evolutionary, and biological psychology, as well as in anthropology and primatology.
Article
Full-text available
The preference for media that confirms prior attitudes and beliefs is problematic in democratic societies based on dialogue and joint deliberation. Over the last decades, partisan selective exposure (PSE) is argued to have increased along with other indices of polarization. We address the question of the increase in PSE, and possible differences by party, ideology and ideological extremity. Using data from the Pew Research Survey, we analyzed self-reported media consumption in 8 nationally representative surveys from the period 2000–2012 (n = 23,381). We relied on previous research on ideological classification of media outlets to conduct confirmatory factor analyses establishing the existence of 2 different variables, conservative and liberal media consumption. We predicted latent variables of media consumption using Item Response Theory models and analyzed the trajectories running latent growth curve models. An unconditional growth model revealed a general and sustained increase in PSE across ideological groups over time. Republicans showed a greater increase over time than did Democrats, after controlling for demographics. Introducing ideological extremity in the model revealed no differences in the trajectories of PSE between liberals and extreme liberals, whereas subjects identified as “very conservative” show a much steeper increase in PSE than any other group, whereas conservatives showed the lowest growth over time. We discuss theoretical implications for ongoing debates about political polarization and ideological asymmetry.
Article
Full-text available
Nam, Jost, and van Bavel (2013) found that conservatives were more likely than liberals to avoid dissonance-arousing situations (viz., writing counter-attitudinal essays in a high-choice situation). A close replication of this original research was unsuccessful, as both liberals and conservatives avoided writing counter-attitudinal essays to similar degrees. We conducted an additional experiment that aimed to conceptually replicate Nam et al. (2013), and to examine whether people whose ideology is threatened might be more likely to avoid dissonance-arousing situations. Again, liberals and conservatives were equally likely to avoid writing counter-attitudinal essays. Threat had no effect on these decisions. A meta-analysis of Nam et al.'s (2013) two studies, the two studies presently reported, and a third supplemental study provide no evidence for asymmetry in dissonance avoidance.
Article
Full-text available
Government agencies around the world have begun to embrace the use of behavioural policy interventions (such as the strategic use of default options), which has inspired vigorous public discussion about the ethics of their use. Since any feasible policy requires some measure of public support, understanding when people find behavioural policy interventions acceptable is critical. We present experimental evidence for a ‘partisan nudge bias’ in both US adults and practising policymakers. Across a range of policy settings, people find the general use of behavioural interventions more ethical when illustrated by examples that accord with their politics, but view those same interventions as more unethical when illustrated by examples at odds with their politics. Importantly, these differences disappear when behavioural interventions are stripped of partisan cues, suggesting that acceptance of such policy tools is not an inherently partisan issue. Our results suggest that opposition to (or support for) behavioural policy interventions should not always be taken at face value, as people appear to conflate their attitudes about general purpose policy methods with their attitudes about specific policy objectives or policy sponsors.
Article
Full-text available
Ideologically committed people are similarly motivated to avoid ideologically crosscutting information. Although some previous research has found that political conservatives may be more prone to selective exposure than liberals are, we find similar selective exposure motives on the political left and right across a variety of issues. The majority of people on both sides of the same-sex marriage debate willingly gave up a chance to win money to avoid hearing from the other side (Study 1). When thinking back to the 2012 U.S. Presidential election (Study 2), ahead to upcoming elections in the U.S. and Canada (Study 3), and about a range of other Culture War issues (Study 4), liberals and conservatives reported similar aversion toward learning about the views of their ideological opponents. Their lack of interest was not due to already being informed about the other side or attributable election fatigue. Rather, people on both sides indicated that they anticipated that hearing from the other side would induce cognitive dissonance (e.g., require effort, cause frustration) and undermine a sense of shared reality with the person expressing disparate views (e.g., damage the relationship; Study 5). A high-powered meta-analysis of our data sets (N = 2417) did not detect a difference in the intensity of liberals' (d = 0.63) and conservatives' (d = 0.58) desires to remain in their respective ideological bubbles.
Article
Research on the dispositional origins of political preferences is flourishing, and the primary conclusion drawn from this work is that stronger needs for security and certainty attract people to a broad-based politically conservative ideology. Though this literature covers much ground, most integrative assessments of it have paid insufficient attention to the presence and implications of contingencies in the relationship between dispositional attributes and political attitudes. In this article, we review research showing that relationships between needs for security and certainty and political preferences vary considerably—sometimes to the point of directional shifts—on the basis of (1) issue domain and (2) contextual factors governing the content and volume of political discourse individuals are exposed to. On the basis of this evidence, we argue that relationships between dispositional attributes and political preferences vary in the extent to which they reflect an organic functional resonance between dispositions and preferences or identity-expressive motivation to adopt a political attitude merely because it is discursively packaged with other need-congruent attitudes. We contend that such a distinction is critical to gaining a realistic understanding of the origins and nature of ideological belief systems, and we consequently recommend an increased focus on issue-based and contextual variation in relationships between dispositions and political preferences.
Article
The motivated social cognition and negativity bias perspectives each posit that threat is especially related to political conservatism, such that threat causes people to adopt politically conservative beliefs, and that political conservatives are especially responsive to threatening stimuli. In this review, I argue that there is a kernel of truth to these perspectives, but that they each define both "threat" and "conservatism" too broadly. I review evidence supporting a Compensatory Political Behavior (CPB) Model, which posits that whereas liberals and conservatives are similarly influenced by and responsive to meaning threats, conservatives, and in particular social conservatives, are differentially influenced by and responsive to physical threats. The CPB model suggests that whereas some political beliefs are more deeply rooted in psychological predispositions, others reflect more surface-based ideological motives. I conclude with suggestions for future research to test the model's predictions regarding the relationship between threat and political ideology.
Article
We fielded an experiment in the 2012 Cooperative Congressional Election Study testing the theory that motivated reasoning governs reactions to news about misdeeds on the campaign trail. Treated subjects either encountered a fabricated news story involving phone calls with deceptive information about polling times or one involving disappearing yard signs (the offending party was varied at random). Control subjects received no treatment. We then inquired about how the treated subjects felt about dirty tricks in political campaigns and about all subjects’ trust in government. We find that partisans process information about dirty campaign tricks in a motivated way, expressing exceptional concern when the perpetrators are political opponents. However, there is almost no evidence that partisans’ evaluations of dirty political tricks in turn color other political attitudes, such as political trust.
Article
Individuals are not merely passive vessels of whatever beliefs and opinions they have been exposed to; rather, they are attracted to belief systems that resonate with their own psychological needs and interests, including epistemic, existential, and relational needs to attain certainty, security, and social belongingness. Jost, Glaser, Kruglanski, and Sulloway (2003) demonstrated that needs to manage uncertainty and threat were associated with core values of political conservatism, namely respect for tradition and acceptance of inequality. Since 2003 there have been far more studies on the psychology of left-right ideology than in the preceding half century, and their empirical yield helps to address lingering questions and criticisms. We have identified 181 studies of epistemic motivation (involving 130,000 individual participants) and nearly 100 studies of existential motivation (involving 360,000 participants). These databases, which are much larger and more heterogeneous than those used in previous meta-analyses, confirm that significant ideological asymmetries exist with respect to dogmatism, cognitive/perceptual rigidity, personal needs for order/structure/closure, integrative complexity, tolerance of ambiguity/uncertainty, need for cognition, cognitive reflection, self-deception, and subjective perceptions of threat. Exposure to objectively threatening circumstances—such as terrorist attacks, governmental warnings, and shifts in racial demography—contribute to modest “conservative shifts” in public opinion. There are also ideological asymmetries in relational motivation, including the desire to share reality, perceptions of within-group consensus, collective self-efficacy, homogeneity of social networks, and the tendency to trust the government more when one's own political party is in power. Although some object to the very notion that there are meaningful psychological differences between leftists and rightists, the identification of “elective affinities” between cognitive-motivational processes and contents of specific belief systems is essential to the study of political psychology. Political psychologists may contribute to the development of a good society not by downplaying ideological differences or advocating “Swiss-style neutrality” when it comes to human values, but by investigating such phenomena critically, even—or perhaps especially—when there is pressure in society to view them uncritically.