ArticlePDF Available

Political Differences in Free Will Belief Are Associated With Differences in Moralization

Authors:

Abstract and Figures

In fourteen studies, we tested whether political conservatives’ stronger free will beliefs were linked to stronger and broader tendencies to moralize, and thus a greater motivation to assign blame. In Study 1 (meta-analysis of five studies, n=308,499) we show that conservatives have stronger tendencies to moralize than liberals, even for moralization measures containing zero political content (e.g., moral badness ratings of faces and personality traits). In Study 2, we show that conservatives report higher free will belief, and this is statistically mediated by the belief that people should be held morally responsible for their bad behavior (n=14,707). In Study 3, we show that political conservatism is associated with higher attributions of free will for specific events. Turning to experimental manipulations of our hypothesis, we show that: when conservatives and liberals see an action as equally wrong there is no difference in free will attributions (Study 4); when conservatives see an action as less wrong than liberals, they attribute less free will (Study 5); and specific perceptions of wrongness account for the relation between political ideology and free will attributions (Study 6a and 6b). Finally, we show that political conservatives and liberals even differentially attribute free will for the same action depending on who performed it (Studies 7a-d). Our results suggest political differences in free will belief are at least partly explicable by conservatives’ tendency to moralize, which strengthens motivation to justify blame with stronger belief in free will and personal accountability.
Content may be subject to copyright.
Everett, Clark, et al. (2020, JPSP:PPID)
1
Political differences in free will belief are associated with
differences in moralization
Jim Albert Charlton Everett*1, Cory J. Clark*2, Peter Meindl3, Jamie B. Luguri4, Brian D.
Earp,5, Jesse Graham6, Peter H. Ditto7, & Azim F. Shariff8
1 School of Psychology, University of Kent
2 Department of Psychology, Durham University.
3 Department of Psychology, Calvin College.
4 The Law School, University of Chicago.
5 Department of Psychology, Yale University.
6 Department of Management, University of Utah.
7 Department of Psychological Science, University of California, Irvine.
8 Department of Psychology, University of British Columbia.
Manuscript Forthcoming in the
Journal of Personality and Social Psychology: Personality Processes and Individual Differences
Abstract
In fourteen studies, we tested whether political conservatives’ stronger free will beliefs were
linked to stronger and broader tendencies to moralize, and thus a greater motivation to assign
blame. In Study 1 (meta-analysis of five studies, n=308,499) we show that conservatives have
stronger tendencies to moralize than liberals, even for moralization measures containing zero
political content (e.g., moral badness ratings of faces and personality traits). In Study 2, we
show that conservatives report higher free will belief, and this is statistically mediated by the
belief that people should be held morally responsible for their bad behavior (n=14,707). In
Study 3, we show that political conservatism is associated with higher attributions of free will
for specific events. Turning to experimental manipulations to test our hypotheses, we show
that: when conservatives and liberals see an action as equally wrong there is no difference in
free will attributions (Study 4); when conservatives see an action as less wrong than liberals,
they attribute less free will (Study 5); and specific perceptions of wrongness account for the
relation between political ideology and free will attributions (Study 6a and 6b). Finally, we show
that political conservatives and liberals even differentially attribute free will for the same action
depending on who performed it (Studies 7a-d). These results are consistent with our theory
that political differences in free will belief are at least partly explicable by conservatives’
tendency to moralize, which strengthens motivation to justify blame with stronger belief in
free will and personal accountability.
free will; morality; blame; motivated cognition; political psychology
Author Note
* Indicates shared first-authorship. Correspondence can be directed at either:
Jim A.C. Everett at Keynes College, Canterbury, Kent or j.a.c.everett@kent.ac.uk or
Cory J. Clark at Department of Psychology, Durham or cory.j.clark@durham.ac.uk
Everett, Clark, et al. (2020, JPSP:PPID)
2
“We must reject the idea that every time a law's broken, society is guilty rather than
the lawbreaker. It is time to restore the American precept that each individual is
accountable for his actions.”
Ronald Reagan (1968)
Liberals and conservatives characteristically view the relationship between the
individual and society in different terms. Whereas liberal (i.e., left-wing) ideology has often
focused on the role of social institutions and other external forces in shaping individual
behavior, conservative (i.e., right wing) thinking tends to emphasize the importance of
personal responsibility (Eidelman et al., 2012; Jost et al., 2008; Skitka et al., 2002; Skitka &
Tetlock, 1992, 1993). According to the conservative view, individuals should take
responsibility for the course of their own lives and refrain from expecting others to solve
their problems. In addition to being explicitly championed by prominent conservative
leaders (Cameron, 2010; Reagan, 1968; Thatcher, 1981), a focus on personal responsibility
seems to pervade the thinking of everyday conservatives as well (Carey & Paulhus, 2013).
Research has shown that conservatives are more likely than liberals to make dispositional
attributions of responsibility in a number of key areas, including poverty (Zucker & Weiner,
1993), unemployment (Feather, 1985), obesity (Crandall, 1994), and even intelligence (Skitka
et al., 2002).
In addition to judging that others are more responsible for their actions, recent
research by Carey and Paulhus (2013) has suggested that conservatives also believe that
others have more free will. Political conservatism is not merely associated with thinking that
others are more responsible for their specific actions, but also with thinking that they have
more autonomous control over their behavior in general. Across three studies, Carey and
Paulhus (2013) found that belief in free will was associated with traditional conservative
attitudes as well as with an increased importance attached to the three ‘conservative’ moral
foundations (loyalty, authority, sanctity). Why might this be so?
We suggest that the relationship between political orientation and free will belief
might be parsimoniously explained by motivated social cognition. This hypothesis is derived
from two areas of research. First, recent research has demonstrated that free will beliefs
are motivated by desires to punish others (Clark et al., 2014) and to justify holding them
morally responsible (Clark, Baumeister, & Ditto, 2017), which recently has been replicated
and confirmed in meta-analyses (Clark, Winegard, & Shariff, 2019). Second, political
conservatives have a tendency to moralize a wider scope of actions than their liberal
counterparts (Graham et al., 2009, 2011, 2013). Combining these two areas of research, we
suggest that conservatives report greater belief in free will and attribute more free will to
people than do liberals because conservatives recognize a wider spectrum of transgressions
for which moral responsibility must be assigned and moral blame attributed.
Motivated Beliefs in Free Will
What do we mean by “free will?” In this paper, we draw on an understanding of free
will that has both been articulated by philosophers and seems to track the intuitions of lay-
people. In line with previous empirical work in this area, we use the term “free will” to refer
to an autonomous choice of action that a person performs in the absence of substantial
Everett, Clark, et al. (2020, JPSP:PPID)
3
internal and external constraints (Baumeister & Monroe, 2014; Paulhus & Carey, 2011),
where this ability to choose renders one morally responsible for their actions (Nichols,
2007; Nichols & Knobe, 2007). Free will, in other words, can be understood as
responsibility-making autonomy. Note that the concept of free will is distinct from the
concept of attributions in social psychology (e.g., Skitka et al., 2002; Zucker & Weiner,
1993), and this can broadly be related to the philosophical distinction between reasons and
causes. Attributions are reasons, and help answer the question of what the reason is for why
a person performed a given action. In social psychology, work on attribution has focused on
two main kinds of reasons: dispositional attributions (the person did it because of the kind
of person they are); and situational attributions (the person did it because of the situation
they were placed in). In contrast, the concept of free will relates to causes, which can
partially include reasons but also ultimate level causal factors (e.g., it was determined by
genes). To illustrate: it is perfectly plausible to say that someone stole something because
they are a selfish person (a dispositional attribution), but that because their selfishness was
genetically determined (an attribution of free will), they did not have free and thus were not
personally responsible.
Assuming this definition of free will of responsibility-making autonomy, what would it
mean for belief in free will to be “motivated,” as we suggested? Motivated social cognition
refers to the well-documented tendency for desired conclusions to organize judgment
processes in a top-down fashion that favors evidence for the conclusions people prefer
(Ditto et al., 2009). When reasoning about the world, people often act more like intuitive
lawyers than intuitive scientists, such that their desired beliefs influence their actual beliefs
(Baumeister & Newman, 1994; Haidt, 2001, 2012). In moral reasoning, desires to blame and
to hold individuals morally responsible compel people to produce rational explanations that
would justify their moral judgments (Alicke, 2000; Clark et al., 2015). Indeed, a growing
body of research has demonstrated that the desire to hold individuals morally accountable
for their immoral behaviors can lead to motivated judgments that such immoral behaviors
are intended, under the agent’s control, and freely chosen (Alicke, 1992, 2000; Alicke, Rose,
& Bloom, 2011; Clark et al., 2014; Clark, Bauman, Kamble, & Knowles, 2017; Clark,
Winegard, & Baumeister, 2019; Cushman, Knobe, & Sinnott-Armstrong, 2008; Hamlin &
Baron, 2014; Knobe, 2003; Knobe & Fraser, 2008; Leslie, Knobe, & Cohen, 2006; Phillips &
Knobe, 2009).
But how might belief in free will, specifically, be seen as a form of motivated social
cognition? Across five studies, Clark et al. (2014) used a range of methods experimental,
correlational, and archival to test the hypothesis that a key motivation underlying belief in
human free will is the desire to hold others morally responsible for their behavior. For
example, telling students that a fellow classmate had cheated on a recent exam increased
belief in free will on a standard measure of global free will belief; and countries with higher
homicide rates were also found to express higher levels of free will belief. Clark et al (2014)
concluded that free will belief is not an abstract, invariant phenomenon, but is rather linked,
at least in part, to a motivated desire to hold others morally responsible for their wrongful
behaviors, the strength of which varies across time and situation.
Everett, Clark, et al. (2020, JPSP:PPID)
4
The focus on wrongful behaviors may have a straightforward explanation. Put simply,
across a broad range of psychological phenomena, “bad is stronger than good” (Baumeister
et al., 2001, p. 1), meaning that people tend to notice, and give greater weight to, negative
actions and outcomes than positive ones. For example, research has repeatedly shown a
praise-blame asymmetry in judgments of intentional action: people are more inclined to say
that a behavior with negative side-effects was performed intentionally than an identical
action with positive side-effects (Knobe, 2003; Pettit & Knobe, 2009). Motivated judgments
of others’ behavior are most pronounced in – and perhaps even driven by cases in which
the behavior is seen as harmful (Alicke et al., 2008). All else being equal, the desire to blame
another for bad behavior is more potent than the desire to praise another for their good
behavior (Clark, Shniderman, Baumeister, Luguri, & Ditto, 2018). As Baumeister et al.
(2001) note, while a general explanation for this effect is hard to come by given its inherent
generality across a broad range of psychological phenomena, it is likely that a tendency to
pay greater attention to bad actions and outcomes than good ones will have been
evolutionarily adaptive because survival often requires more urgent attention to possible
bad outcomes (e.g., a predator behind you) than possible good outcomes (e.g., a berry bush
behind you).
Political Differences in Morality
What could explain a greater belief in free will among conservatives compared to
liberals? It is possible that this difference simply reflects a fundamental underlying political
difference in perceptions of how much freedom and control people have over their
behavior and life outcomes. But given the work of Clark et al. (2014) on free will belief
being driven by a desire to blame, we predicted that conservatives’ greater endorsement of
belief in free will might be dueat least in part—to conservatives’ stronger tendency to
moralize. According to a large body of research, political conservatives and liberals differ in
many more ways than their preferred political candidates (Amodio et al., 2007; Jost et al.,
2009), including at the cognitive level. Most importantly for the purposes of this paper, a
growing body of evidence has demonstrated that political orientations are associated with
differences in judgments in the moral domain, both in terms of what constitutes a moral
issue in the first place as well as how wrong a particular action/behavior is.
The social intuitionist approach suggests that moral judgment is largely a motivated
phenomenon: moral judgment is triggered by quick moral intuitions, and moral reasoning
largely serves as a post-hoc rationalization of these intuitive judgments (Haidt, 2001). Moral
Foundations Theory (Graham et al., 2011, 2013) is rooted in work on motivated cognition
and the social intuitionist approach to morality (Baumeister & Newman, 1994; Ditto et al.,
2009; Haidt, 2001, 2012) and posits that conservatives have a broader moral domain than
do liberals. In other words, liberals are less inclined to perceive a variety of actions as
morally relevant than conservatives (Graham et al., 2009) a finding demonstrated by
analyzing data drawn from participants from 11 different world regions (Graham et al.,
2011), and from life narrative interviews with politically engaged adults (McAdams et al.,
2008). Similarly, we note, political conservatism is associated with more punitiveness in
general (e.g., Carroll et al., 1987; Sargent, 2004). So while liberals can and do find some
actions more morally wrong than conservatives (Frimer et al., 2017), the available evidence
Everett, Clark, et al. (2020, JPSP:PPID)
5
does suggest that conservatives tend to view a broader range of actions as having moral
significance,
1
and therefore as more suitable for moral judgment and blame. We
demonstrate this ourselves in Study 1 of the manuscript, using a meta-analysis of five new
studies with a combined n of 308,499, establishing the conservative tendency to moralize.
Practical Significance
If our hypothesis is correct, this could help explain a variety of political
disagreements between liberals and conservatives regarding the degree to which various
groups (e.g., the unemployed, the homeless, prisoners, women with unwanted pregnancies,
etc.) are responsible for their plights, and thus how they should be treated by governmental
policies. Moreover, our hypothesized results might help explain why these kinds of
disagreements seem so intractable. If policy differences result from affect-based moral
intuitions about responsibilityand liberals and conservatives have different moral
intuitionsit may be difficult or impossible for liberals and conservatives to agree on the
“correct” policy solutions. And these disagreements should not be expected to be solved by
collecting more and better data.
The Present Research
We conducted a series of 14 studies to explore this motivated cognition account
whereby conservatives ascribe more free will because they have a broader moral domain,
and thus more often perceive actions as appropriately subject to moral judgment. In other
words, we predicted that because liberals think that a narrower range of phenomena
constitute moral issues, they have a narrower range of actions for which judgments of blame
and attributions of free will are deemed to be appropriate. If this prediction is correct, 1)
higher tendencies to blame should account for conservatives’ stronger belief in free will, and
2) political liberals and conservatives should differ in attributions of free will only when
there are corresponding differences in the extent to which they perceive actions to be
morally wrong. In short, we should find that conservatives believe more in free will and
ascribe more free will generally (Studies 1-3), but in instances where differences in
perceived moral wrongness can be removed or reversed, differences in free will ascriptions
should similarly be absent or in the opposite direction, respectively (Studies 4-7).
Experimental materials, pre-registrations, data, analysis code, and results can be seen at the
Open Science Framework: https://osf.io/ny82c/
Meta-analysis of Studies 1a-1e
This opening meta-analysis was conducted to establish the relationship between
more conservative political ideology and moralization. We combined data from five studies
drawn from a variety of populations with a total n of 308,499. Each study included a
measure of political ideology and one or more measures of moralization.
1
While aspects of our argument are adjacent to Moral Foundations Theory (MFT), we do not depend on the
specifics of MFT and will avoid these arguments in this paper. We believe the findings from MFT support our
assertion that conservatives moralize more than liberals, but we do not rely exclusively on MFT to make this
point. We establish this ourselves in Study 1. Moreover, we take no stance on the kinds or categories of moral
intuitions that liberals or conservatives might have, nor whether these kinds or categories can all fit under a
“harm” umbrella or whether they reflect distinct moral modules.
Everett, Clark, et al. (2020, JPSP:PPID)
6
Studies 1a-1e Methods
Table 1 contains summary details (sample information, ideology and moralization
measures) for Studies 1a-1e and full details of all study methods are available at the OSF.
Ethics Statement. Studies 1a and 1c received ethical approval from the University
of Southern California’s Institutional Review Board (“Morality Studies”; UP-07-00393).
Studies 1d and 1e were covered by a separate application from the same institution
("Reading, Thoughts, and Behavior"; UP-12-00388). In Study 1b, we conducted secondary
data analysis of a publicly available dataset provided by the “Measuring Morality” project
based at Duke University.
Participants. Studies 1a and 1c were conducted on yourmorals.org, a survey website
on which participants (mostly from the U.S., but some from across the globe) complete
surveys in exchange for response feedback. For Study 1b, data were drawn from the
Measuring Morality Survey from a nationally representative panel of adult participants
maintained by Knowledge Networks. Studies 1d and 1e were conducted on Amazon
Mechanical Turk (Mturk) with U.S. participants. Mturk is not perfectly representative of the
U.S. as a whole, but it is diverse with respect to age, sex, race, education, SES, and ideology
more so than student samples are (Paolacci & Chandler, 2014).
Ideology. Ideology was self-reported on 7-point scales in all studies. In Studies 1a,
1c, and 1d, the poles were “Very liberal” to “Very conservative.” In Study 1b, these were
“Extremely liberal” to “Extremely conservative.” In Study 1e, participants self-reported their
political ideology on economic issues, social issues, foreign policy issues, and overall on the
same scale as Study 1b, which were combined into an index of ideology.
Moralization. Full details of each moralization measure are available at the OSF. Some
of the moralization measures can be criticized for containing politically relevant items (e.g.,
religious items), and so the meta-analysis was conducted twice, once with all moralization
measures, and once with the politically irrelevant measures only. The measures that were
excluded in the politically irrelevant version of the meta-analysis have an asterisk by their name
in the descriptions below and in Table 1.
In Study 1a, moralization was measured using The Moral Foundations Questionnaire*
(MFQ; Graham et al., 2009). Study 1b included three moralization measures: (1) The
Moralization of Everyday Life Scale (MELS; Lovett et al., 2012), (2) The Moral Foundations
Sacredness Scale* (MFSS; Graham & Haidt, 2012), and (3) The Ethical Values Assessment* (EVA;
Padilla-Walker & Jensen, 2016).
In Study 1c, moralization was measured by having participants evaluate how important
it was for a person to possess various characteristics in order to be a morally good person.
Participants completed a random subset of 45-47 characteristics from a list of 92. This list was
created by combining previous studies that attempted to create representative lists of
moralized characteristics (Aquino & Reed II, 2002; Cawley et al., 2000; Lapsley & Lasky, 2001;
K. D. Smith et al., 2007; Walker & Pitts, 1998) and conducting a large scale pretest in which
4,565 yourmorals.org participants were asked to report a) their most important moral values
and b) behaviors for which they most often morally judged people.
Study 1d measured moralization by having participants view 16 pictures of faces and
rate how morally bad the depicted person probably was. The 16 pictures were taken from a
Everett, Clark, et al. (2020, JPSP:PPID)
7
larger set of pictures shown to effectively capture four main quadrants of social perception
(e.g. Harris & Fiske, 2006): Warmth/Incompetence; Coldness/Incompetence;
Coldness/Competence; Warmth/Competence. Study 1e measured moralization by having
participants rate how morally bad 30 personality traits are. These 30 traits were the individual
personality facets from the NEO PI-R (Costa & McCrae, 1992a, 1992b).
Meta-analysis Methods
We employed many of the suggested procedures outlined by Goh, Hall, and Rosenthal
(2016) to conduct our meta-analysis. Because the studies were drawn from different
populations and used different measures of moralization, we report random effects (Goh et
al., 2016; see also Hedges & Vevea, 1998).We included one effect size for each study (i.e., the
effect sizes for the three moralization measures in Study 1b were averaged; Card, 2012), thus
five r effect sizes (the correlations between more conservative ideology and higher
moralization) were included. We conducted one-sample t-tests of the effect sizes first with all
moralization measures and then with the potentially politically confounded ones removed.
Results
As can be seen in summary Table 1, more conservative political ideology was positively
and significantly related to greater moralization in every single study (ps < .027), including
those containing zero political content. The meta-analysis revealed a small to medium effect
size overall, r = .27, p = .003; which remained small to medium and significant with the politically
irrelevant measures only, r = .24, p = .012.
Everett, Clark, et al. (2020) JPSP:PPID
8
Moralization meta-analysis summary and results
Sample
n
M
age
%
Male
Ideology Measure
Moralization Measure
yourmorals.org
303,553
40.88
55
7-point "Very liberal" to "Very
conservative"
Moral Foundations
Questionnaire*
Measuring
Morality
1,516
46.76
48
7-point “Extremely Liberal” to
“Extremely Conservative”
Moralization of Everyday
Life Scale
Survey
Moral Foundations
Sacredness Scale*
Ethical Values Assessment*
Study 1b Overall
Study 1b Politically
Irrelevant Overall
yourmorals.org
2,987
38.38
56
7-point "Very liberal" to "Very
conservative"
Importance of morally good
characteristics
mturk
179
33.95
47
7-point "Very liberal" to "Very
conservative"
Moral badness of 16 faces
mturk
264
37.11
38
7-point “Extremely Liberal” to
“Extremely Conservative” Index on economic
Moral badness of 30
personality facets
social, foreign policy, and overall
Random Effects (all measures)
Random Effects (politically irrelevant only)
Note. Asterisk indicates the moralization measure contains items with political significance. The Politically Irrelevant Results do not include these measures.
Everett, Clark, et al. (2020, JPSP:PPID)
9
Discussion
This meta-analysis adds to the body of research demonstrating that
conservatives have stronger tendencies to moralize than liberals. This was true overall, and
crucially, was true even for moralization measures containing zero political content (e.g., moral
badness ratings of faces and personality traits). Thus, these results support a premise of our
present hypothesis and justifies our examination of whether conservative moralization may
explainin part—conservatives’ stronger beliefs in free will and personal responsibility.
Study 2
Study 2 employed individual difference measures to examine the relationships among
political ideology, free will belief, and moralization. We conducted a preliminary test of our
hypothesis by assessing whether tendencies to hold others morally responsible for
blameworthy actions mediate the relationship between conservative ideology and belief in
free will. The Free Will and Determinism Scale-Plus (FAD-Plus; Paulhus & Carey, 2011) is
probably the most widely used free will belief scale, but it is routinely criticized for confounding
free will and moral responsibility for blameworthy actions (Clark et al., 2014; Clark, Winegard,
& Baumeister, 2019; Monroe & Ysidron, 2019). Items such as “Criminals are totally responsible
for the bad things they do” and “People must take full responsibility for any bad choices they
make” measure beliefs about how blameworthy people generally are for bad actions more
than beliefs about freedom and control. Other items, however, exhibit higher prima facie
validity, such as “People have complete control over the decisions they make” and “Strength
of mind can always overcome the body’s desires.” We did not have access to large scale data
with the modern FAD-Plus, but we did have access to large scale data with the original FAD
(Paulhus & Margesson, 1994), which is identical on five of the seven items and similarly contains
four face valid items and three general blame items. We decided to leverage the flaws of this
scale to examine whether tendencies to blame (using the blameworthy scale items) statistically
mediate the relationship between conservative political ideology and free will belief (using the
face valid items).
2
There are two limitations to this approach. First, the FAD items were likely
selected by scale developers to intercorrelate and this weakens their appropriateness for
mediation analysis. Second, causal orders cannot be inferred from mediation analyses of this
sort (see our later studies, which experimentally test whether manipulating moral blame
motives influence free will judgments, for such causal evidence). Nevertheless, this analysis is
at least suggestive that a proportion of the relationship between more conservative political
ideology and free will beliefs can be linked to general blame tendencies.
Method
Participants (n = 14,708; 38% Female; Mage = 34.80, SD = 15.88; 73% from the U.S., the
remaining from more than 100 countries) were recruited through yourmorals.org. This study
received ethical approval from the IRB of UC Irvine (“Moral Psychology on the Internet”;
Protocol #2007-5740). Demographic information is collected at registration including age, sex,
education, religious attendance, and political orientation. After registration, visitors self-select
2
We have not conducted any analyses on the individual items of the FAD here, but our data are openly
available for any researchers interested in this.
Everett, Clark, et al. (2020, JPSP:PPID)
10
to take one or more surveys from a list of over 50. For inclusion, participants had to have
completed the free will subscale from the Free Will and Determinism scale (FAD; Paulhus &
Margesson, 1994), which was broken down into one free will belief subscale ( = .69)
containing four items (e.g., “People have complete control over the decisions they make.”) and
one blame subscale ( = .71) containing three items (e.g., “Criminals are totally responsible
for the bad things they do.”), each rated on a 5-point scale from “strongly disagree” to “strongly
agree”. All participants who completed the FAD were included if they also reported their
political ideology on a 7-point scale from “Extremely Liberal” to “Extremely Conservative.
The sample leaned liberal (M = 2.99, SD = 1.69), with 66.4% below the moderate midpoint,
13.8% at the moderate midpoint, and 19.7% above the moderate midpoint.
Results
More conservative ideology was positively correlated with free will belief, r = .351, p
<.001, and blame tendencies, r = .427, p <.001, and free will belief and blame tendencies were
positively correlated, r = .527, p <.001. We conducted two bootstrap mediation analyses
(10,000 resamples) testing whether blame tendencies statistically mediated the relationship
between ideology and free will beliefs both with and without demographic controls (age, sex,
education, and religious attendance). Both mediations were highly significant, without controls
(see Figure 1), b = .100, se = .002, 95% CI [.095-.104], Z = 41.52, p <.0001, and with all controls,
b = .097, se = .003, 95% CI [.090-.103], Z = 29.15, p <.0001 (being younger, female, less
educated, and more religious all significantly predicted stronger free will belief; being older,
female, and less educated significantly predicted stronger blame tendencies [religiosity did not
significantly predict blame tendencies]).
Figure 1. Indirect effect of more conservative political ideology on free will belief
through blame in Study 2. Note. *** indicates p < .001.
Discussion
Study 2 provided individual difference evidence consistent with our hypothesis.
Stronger beliefs that people are morally responsible for their bad behaviors statistically
mediated the relationship between more conservative ideology and stronger beliefs in free
willboth with and without relevant demographic controls. This suggests that higher
tendencies to blame account for a proportion of the relationship between more conservative
Everett, Clark, et al. (2020, JPSP:PPID)
11
political ideology and free will beliefs. Of course, correlational studies cannot supply evidence
of causal relationships and so our later studies manipulate blame desires directly to test how
these differentially impact the free will judgments of conservatives and liberals.
Study 3
In Study 1 we have demonstrated that conservatives moralize a much broader range
of actions than liberals, and in Study 2 that conservatives report higher general belief in free
will, and that this association of ideology with free will belief is partially mediated by beliefs
that others are morally responsible for their bad actions. In Study 3 we wanted to turn away
from reports of general, abstract belief in free will to look at attributions of free will for specific
events, using the opportunity to look at attributions of free will for both positive and negative
events.
To our knowledge, no research has been conducted on a comparison of political
differences in attributions of free will for positive events. This is surprising, though, because
the comparison between attributions for positive and negative events sets up an interesting
test of two alternative explanations for what may be driving the existing liberal-conservative
differences in endorsement of free will. If conservatives are dispositionally inclined to hold a
stronger belief that people have free will than liberals, then conservatives should endorse
and attribute greater free will regardless of whether their actions lead to good or bad
outcomes. However, if differences in free will belief are in part motivated by moral
judgments of wrongness and a desire to blame, then we should see an interaction between
political affiliation and the positivity or negativity of the event. Since conservatives have
stronger tendencies to moralize, they should show a stronger negativity bias in their free
will attributionsthat is, attributing more free will to actions with negative rather than
positive outcomesthan would liberals. In Study 3, we test this.
Method
Open Science and Ethics Statement. Our design, hypotheses, and analysis plan
were all pre-registered at the Open Science Framework. For this study and all subsequent
ones, we report all measures, manipulations, and exclusions, and results, analysis code, and
experiment materials are available for download at: https://osf.io/ny82c/. This study received
ethical approval from the IRB of the University of British Columbia (“Social Impacts of
Emerging Technology”; Protocol #H18-02727).
Participants. We originally recruited 146 American participants online using MTurk.
On the recommendation of an anonymous reviewer and editor, we subsequently conducted a
pre-registered second wave of data collection to maximise power. We report in the main text
results using the original sample combined with the new one, which gave us a final sample of
444 participants (188 female, Mage = 35). On average, participants were slightly left-of-center
on a 1-7 scale (M = 3.89, SD = 1.65), with 188 Democrats and 131 Republicans (the remaining
were neither).
Design. We had a fully within-subjects design, where participants were asked to rate
how much free will someone had for six distinct situations. Three of these were negative (the
material living conditions of the homeless, drug addicts' addictions, a man imprisoned for
participating in gang violence) and three were positive (the financial success of investment
Everett, Clark, et al. (2020, JPSP:PPID)
12
bankers, students’ admission to elite universities, a famous musician winning a Grammy award).
These items were selected simply because they appeared to be unambiguously positive or
negative, but later studies will more systematically pretest the selected target actions. For each
item, participants rated how much (a) responsibility (1 = Not at all personally responsible to 7 =
Completely personally responsible);
3
(b) control (1 = Not at all in control; 7 = Completely in control);
and (c) free will (1 = No free will at all; 7 = Complete free will) the different groups or individuals
concerned had for their current situation. These scores were aggregated to form a reliable
overall measure of free will attributions for both positive ( = .83) and negative situations (
= .82).
Analysis. We collected and analyzed one wave of data prior to our initial submission
for publication. At the request of a reviewer and our editor, we then conducted another wave
of data collection before completing the final analyses. To account for this sequential testing
and ensure full transparency, we have taken three approaches. First, we have simply reported
the results in full for the original sample in the supplemental results at the OSF. Second, we
have employed the technique of adjusting our significance levels for sequential testing.
Following the guide provided by Lakens (2014), we have used a linear alpha spending technique
to obtain a revised significance level of p ≤ .038 that accounts for the sequential testing. While
an increasingly common technique in social psychology that easily allows the reader to see
whether our key result is significant at this revised level, the strategy is formally inappropriate
though still acceptable because we did not pre-specify that we would conduct two waves
of data collection and analysis. Given this, the third strategy we have used is to compute a p-
augmented statistic (Sagarin et al., 2014), a technique explicitly designed for cases where the
sample size has been increased post-hoc after initial data analysis. This statistic consists of a
range of values greater than p .05 and represents the magnitude of the resulting Type I error
inflation as a result of our increased data collection.
Results
We used regression to look at whether political orientation (continuous between-
subjects) predicted free will attributions, and whether this differed for the positive and negative
events (as a within-subjects variable, 0 = positive, 1 = negative). In this and all subsequent studies,
we report semi-partial rs (the proportion of the variance in free will attributions uniquely
explained by the indicated predictor) as estimates of effect sizes (except for the mixed within-
between interaction here, for which the semi partial r is not available).
Results using the full sample revealed a significant interaction between valence
condition and political ideology, b = -.03, SE = .01, t = -2.16, p = .031, 95% CI [-.06, .00], Pseudo-
= 0.49, paugmented = [.053, .073]. Though a small effect, as can be seen in see Figure 2, this
interaction was significant even at our lower threshold of significance, p < .038, accounting for
the sequential data collection (Lakens, 2014). This interaction supplemented a main effect of
3
A reviewer questioned the appropriateness of a responsibility question in our index of free will attributions.
Here, we are precisely interested in motivated increases in free will attributions for purposes of increased
personal responsibility, so it seemed appropriate for the question at hand. But we also cross-checked our main
results removing this responsibility item, and this did not impact the statistical significance of any of our main
results. Thus, even removing this item, the interpretation of our data remains unchanged.
Everett, Clark, et al. (2020, JPSP:PPID)
13
ideology whereby conservative ideology was associated with greater free will attributions
overall, b = .26, SE = .02, t = 12.06, p < .001, 95% CI [.22, .31], and no overall difference in
attributions of free will for the positive or negative events, b = -.02, SE = .03, t = 0.58, p = .56,
95% CI [-.04, .07]. Specifically, the relationship between conservative ideology and attributions
of free will was stronger for negative events, b = .30, SE = .03, t = 11.24, p < .001, 95% CI
[.24, .35], R2 = 0.22, semi-partial r = 0.47, than for positive events, b = .23, SE = .03, t = 8.78, p
< .001, 95% CI [.18, .28], R2 = 0.15, semi-partial r = 0.39 (see Figure 2).
Figure 2. Attributions of free will as a function of participant political ideology, for
positive and negative events separately. (Study 3).
Discussion
In Study 3 we looked at whether political ideology was associated with attributions of
free will for specific events. We found that political ideology indeed predicted higher free will
attributions overall, and this relationship appeared stronger for negative situations than
positive ones (though, contra our pre-registered hypothesis, the political difference was only
reducedrather than absent entirelyamong positive situations). Conceptually, the capacity
for free will should hold whether one experiences a good or bad outcome, and so if ideology
is genuinely related to an abstract belief in free will, there should have been no difference
depending on the valence of the outcomes. That we observed a small-but-significant
interaction whereby conservatism predicted higher attributions of free will to a stronger
degree for the negative than the positive events, however, suggests that free will attributions
are not merely reflecting some dispositional variance in a belief in human autonomy, but instead
a more basic, social psychological phenomenon likely one relating to blame.
One important limitation of this study, though, is that we did not control for
perceptions of wrongness of the events. We assume that Republicans attributed more free will
to the negative eventsbeing homeless, being a drug addict, and being imprisoned for gang
violencebecause these are all situations that conservatives tend to find more morally
Everett, Clark, et al. (2020, JPSP:PPID)
14
objectionable than do liberals (e.g., Graham et al., 2009). If free will judgments only differ by
political ideology when there are differences in perceived moral wrongness, then these
differences should be eliminated when the perceived wrongness of the event is matched for
both conservatives and liberals.
Study 4
Study 4 sought to further chart the boundaries of the association between free will
and political ideology by looking at attributions of free will for specific events that were pre-
tested to be equally immoral for liberals and conservatives. If political differences in free will
belief are mainly derived from differences in moralizing, these free will differences should be
eliminated when looking at items judged as equally wrong by both liberals and conservatives.
On the other hand, if political ideology is more abstractly and generally related to free will
belief, then conservatism should be associated with higher free will attributions across
different events, even those that are judged as equally wrong by both liberals and conservatives.
While most prior work has focused on attributions for morally negative events only
(see Study 3), we wanted to look at free will attributions for a range of both moral and non-
moral, and positive and negative events to test the generality of our hypothesis. If our
hypothesis that political ideology is related to free will belief because of differences in the
scope of the moral domain is correct, we should observe roughly equivalent attributions of
free will between liberals and conservatives for any type of action that has roughly equal moral
significance for liberals and conservatives.
Method
Pre-Testing. We pretested items for use in Study 4 that did not significantly differ in
perceived morality (or valence) based on political orientation. To do this, we recruited 109
MTurk participants (34 female, Mage = 36), though five participants were excluded for failing
two attention checks (e.g., “Please click Scale Point 1 to confirm you’re paying attention”),
leaving 104 participants in the analysis. On a 1-7 scale (7 indicating stronger conservatism),
participants were on average slightly left-of-center (M = 3.53), with slightly liberal views on
social issues (M = 3.19) and moderate views on economic issues (M = 3.88). All main political
positions were represented, with 21 Republicans, 39 Democrats, and 43 Independents.
Participants were required to rate a series of 51 events and occurrences, 1) for how
positive or negative it was (-100 = negative; 0 = neutral; 100 = positive); and 2) for how moral it
was: is the action morally bad, morally good, or irrelevant to morality? (-100 = morally bad; 0 =
morally irrelevant; 100 = morally positive). All of these items were devised by the researchers to
be as politically neutral as possible, including a range of both positive and negative items, and
moral and non-moral items. The list of all items, along with mean scores and correlations with
ideology, can be seen on the OSF. To select the final items to use for the main study we used
a statistical cut-off point (r < .10; p > .40) to ensure conservatives and liberals did not rate the
item as differentially positive or negative, or differentially morally relevant or not. Based on
this, we selected 20 items: five that were moral and positive; five that were moral and negative;
five that were non-moral and positive; and five that were non-moral and negative.
Open Science and Ethics Statement. For the main study, we report all measures,
manipulations, and exclusions. Results, analysis code, and experiment materials are available
Everett, Clark, et al. (2020, JPSP:PPID)
15
for download at the OSF. This study received ethical approval from the IRB of the University
of Oregon (“The Effect of Culture on Attitudes and Outcomes”; Protocol #10162012.023).
Participants. For the main study, 647 American participants completed the survey
online using MTurk. 18 participants were excluded from analysis for taking the survey more
than once, leaving a final sample of 629 participants (269 female; Mage = 35). Our sample size
was determined by available funds and a goal to recruit at least 150 participants per each of
the four conditions.
Design. Our experimental design had a 2 (morality: morally relevant vs. non-moral) x
2 (valence: positive vs. negative) experimental design, where in each condition participants
were given the five events obtained from pre-testing and asked to indicate how much free will
they perceived the actor to have. Some participants read events that were - equally to liberal
and conservatives - seen as positive and morally-relevant (e.g., “Working one day a week at a
soup kitchen”); others read events that were negative and morally-relevant (e.g., “Spreading
malicious rumours about a co-worker”); others read events that were positive but non-moral
(e.g., “Making money from a smart investment”); and the remaining participants read events
that were negative but non-moral (e.g., “Failing a college exam”). Full items can be seen on the
OSF. The order of the events and the five dependent variables for each event were all
randomized for each participant to avoid potential order effects.
Measures. For each of the five events, participants rated the degree of free will that
they perceived the actor to have using five items: “How much control would someone have
over __”; How much responsibility would someone have for __”; “To what extent would
someone who did __ have exercised free will?”; “To what extent is someone who __
performing an action that is freely chosen?”; “To what extent would someone who did __ have
been able to have made other choices and not done this?” rated on 7-point scales from “not at
all” to “very much.Within each condition these five items showed high internal consistency
(s > .88) and so were aggregated together to form an overall measure of free will attributions
(i.e., all 5 DVs for all 5 items).
Participants indicated their political ideology on two scales: one measuring
social conservatism, and one measuring economic conservatism, rated on 7-point scales from
“very liberal” to “very conservative,” which were combined into an index of overall political
ideology ( = .73). Finally, participants completed the free will subscale of the FAD+.
Results
First, we looked at the potential interactive effect of political orientation with
attributions of free will for events that were either positive or negative, and either morally
relevant or non-moral. We used a regression-based procedure to examine the effects of
morality (-1 = non-moral; 1 = morally relevant), valence (-1 = positive, 1 = negative), and
participant political orientation (centered) on attributions of free will. As expected, there was
no main effect of political ideology on ascriptions of free will, b = .00, SE =.02, t = -0.09, p =
0.93, 95% CI [-.04, .03], semi-partial r = -.00. Moreover, there was no three-way interaction of
morality, valence, and political orientation, b = -.03, SE =.02, t = -1.63, p = 0.10, 95% CI [-.07, .01],
R2 = 0.28, semi-partial r = -.06, and no interaction of political orientation with whether the
event was moral or non-moral, b = -.03, SE =.02, t = -1.51, p = 0.13, 95% CI [-.06, .01], semi-
partial r = -.05, or of political orientation with whether the event was positive or negative, b
Everett, Clark, et al. (2020, JPSP:PPID)
16
= .01, SE =.02, t = 0.59, p = 0.55, 95% CI [-.02, .05], semi-partial r = .02. That is, for these items
that we had pre-tested to be equally matched in morality (and valence) for liberals and
conservatives, there was no effect of political ideology on specific free will attributions. Indeed,
for the 10 items pre-tested to be equally morally relevant to liberals and conservatives, there
was not a single significant correlation between political orientation and free will attributions.
This was despite replicating our finding from Study 2 (and previous work by Carey and Paulhus,
2013), that conservative political ideology was significantly positively correlated with scores
on the free will subscale of the FAD+ (r = 0.29, p < .001), with conservatives showing higher
endorsement of both the more abstract (r = 0.24, p < .001) and moral responsibility focused
(r = 0.28, p < .001) items of the scale. Scores on the FAD+ were significantly correlated with
free will attributions across the events (r = 0.39, p < .001).
In other words, even though conservatives consistently report stronger beliefs
in free will than liberals, and stronger beliefs in free will are associated with stronger
attributions of free will in general, conservatives do not attribute more free will than liberals
to actions that they perceive as equally as morally wrong as liberals. Thus, when the moral
relevance of actions is controlled for, political differences in free will attributions are
eliminated.
Discussion
In Study 4 we explored whether we might be able to “break” the association
of political ideology and free will by looking at actions judged to be morally equivalent between
liberals and conservatives. If political ideology is related to free will belief primarily because of
differences in the scope of the moral domain, we reasoned, these differences should be
markedly reduced when looking at items judged as equally wrong by both liberals and
conservatives. Indeed, we found the differences were eliminated altogether: though we again
replicated that conservative ideology was significantly positively associated with an abstract
agreement that humans have free will, there were no differences in specific attributions of free
will for actions that were pre-tested to be equally morally valenced for liberals and
conservatives. These results are consistent with our hypothesis that differences in
conservatives’ and liberals’ perceptions of free will may be partially due to differences in
moralization, rather than representing any generalized, abstract belief that human behaviors
are freely chosen.
One potential concern with Study 4, however, is that we predicted (and found)
a null effect. Our findings are consistent with the idea that differences in conservatives’ and
liberals’ perceptions of free will are partially due to differences in moralization, and
inconsistent with the idea that conservatives attribute more free will regardless of the moral
content. Nonetheless, because it is impossible to prove a null hypothesis by rejecting it, we
cannot make firm conclusions about the motivated basis of free will on the basis of such a null
finding. In Study 5, then, we aimed to conduct a stronger test of our hypothesis by looking at
whether the association of conservatism with greater free will could be reversed, not just
eliminated.
Study 5
Everett, Clark, et al. (2020, JPSP:PPID)
17
If the typical tendency for conservatives to endorse stronger beliefs in free will is due
to a motivation to blame, rather than a stable personality characteristic, this tendency ought
to reverse when confronted with behaviors that conservatives perceive as less morally wrong
than their liberal counterparts. We tested this prediction in Study 5.
Method
Pre-Testing. We first pretested for items that political conservatives would see as
less morally wrong than liberals. American MTurk participants (n = 100; 44 female, Mage = 32)
rated how liberal or conservative they were on a 1-7 scale, with participants being slightly
liberal on average (M = 3.44), with slightly more liberal views on social issues (M = 3.07) and
more moderate views on economic issues (M = 3.80). All main political parties were
represented, with 13 Republicans, 47 Democrats, and 36 Independents. To help select our
items for the main study, participants rated the moral wrongness of a series of 25 events and
behaviors that were devised by the research team to specifically cover things that liberals
typically are more morally concerned about than conservatives (e.g., animal welfare, recycling,
prejudiced behavior [Graham et al. 2009]). The list of all items, along with mean scores and
correlations with ideology, can be seen on the OSF.
From participants’ ratings, we selected five items based on both the strength of
correlation between political ideology and wrongness (all ps< .05; all rs > .22), and the mean
wrongness (to avoid ceiling or floor effects: Ms between 3.62 and 4.95 on a 1-7 scale). It is
illuminating to note that, for 10 out of the 25 items, there were no significant correlations of
political orientation with ratings of wrongness. In other words, even when we intentionally
created items to capture things that liberals, but not conservatives, are typically outraged by,
conservatives judged 40% of the items as equally wrong as liberals did.
Open Science and Ethics Statement. For the main study, we report all measures,
manipulations, and exclusions. Results, analysis code, and experiment materials are available
for download from the OSF. This study received ethical approval from the IRB of the University
of Oregon (“The Effect of Culture on Attitudes and Outcomes”; Protocol #10162012.023).
Participants. For the main study, 513 American participants completed the survey
online via MTurk.
4
Participants were excluded from data analysis if they took the survey more
than once (n = 5) or failed one or both of two simple checks in which they had to select a
certain scale-point in the FAD+, and then indicate at the end of the survey which event they
did not answer questions about (n = 36). This left a final sample of 472 participants (237 female;
Mage = 37 years). As in Study 3, we used the same combination of sequential testing (Lakens,
2014) and a paugmented statistic (Sagarin et al., 2014) to account for the two stages of data
collection and analysis.
4
We originally recruited 204 participants and analyzed the results (after exclusions, n = 193), and our results
were not statistically significant. To explore whether this outcome was due to any problems with the specific
items, we decided to run another pilot study. However, when attempting to conduct this pilot, a technical error
resulted in recruitment of 275 more participants for the present study instead of the new pilot. We then re-
analyzed the data with the updated sample and found statistically significant results supporting our hypotheses,
suggesting that the non-significant effects in the initial analysis were explained by a lack of power. In the
interest of full transparency, we report both sets of results in full at the OSF, and then use the same combination
of sequential testing and the paugmented statistic to account for our waves of data collection.
Everett, Clark, et al. (2020, JPSP:PPID)
18
Design. In the main study, participants were given a list of five events that pre-testing
revealed would be perceived as more wrong by liberals: 1) “Robert sends a formal complaint
to his child’s school after finding that his child’s kindergarten teacher is transgendered”; 2)
“Sarah uses make-up products that are tested on animals”; 3) “In conversation with a fellow
student, John finds out that the student is gay. John immediately tells the student that he will
pray for him”; 4) “Garrett manages an upscale members club where only young and attractive
women are employed”; and 5) “Riley is remotely piloting military drones above Afghanistan.
She has been ordered to target a compound believed to hold terrorists, but she knows there
may also be civilians. A few hours after she pilots her drone to attack the compound, she finds
out that 8 suspected terrorists and 5 children died in the resulting explosion”.
Each of the five events were presented in randomized order. Participants rated
how much free will, responsibility, control, and free choice the actor had, and these four items
were aggregated to form a measure of free will attributions, as for Study 3 (= .85).
Participants also rated how wrong they perceived the described action to be on a 1-7 scale
from “not at all wrong” to “very wrong. Last, participants again completed the free will
subscale of the FAD+ and reported their political ideology on the same two 7-point economic
and social conservatism used in previous studies.
Results
Supporting our hypothesis, for actions in which political liberals were more motivated
to assign blame (correlation of rated wrongness with ideology: r = -.47, p < .001), conservative
political ideology (mean-centred) negatively predicted free will, b = -.05, SE = .02, t = -2.73, p
= .007, 95% CI [-.09, -.02], R2 = 0.02, semi-partial r = -.12, paugmented = [.053, .054]. This interaction
was significant even at our lower threshold of significance, p < .036, accounting for the
sequential data collection (Lakens, 2014). That is, for these actions that political liberals saw as
more wrong, it was political liberalism that predicted greater free will attributions. This was
despite political conservativism again being significantly positively correlated with scores on
the free will subscale of the FAD+ (r = 0.19, p < .001), with conservatives showing higher
endorsement of both the more abstract (r = 0.16, p < .001) and moral responsibility focused
(r = 0.19, p < .001) items of the scale. That is, whereas political conservatives again reported
higher general, abstract belief in free will, when it came to attributing free will for specific
events that conservatives found less morally wrong than liberals, conservatives attributed less
free will.
Supporting the idea that differences in moralization underpin the specific free will
attributions, we found that when adding perceived moral wrongness (mean-centred) to the
model, political ideology no longer predicted ascriptions of free will, b = -.03, SE = .02, t = -
1.18, p = .24, 95% CI [-.07, .02], semi-partial r = .05, with only reported moral wrongness
significantly predicting free will attributions, b = .08, SE = .03, t = 2.61, p = .009, 95% CI [.02, .13],
semi-partial r = .12 thus implicating wrongness as the factor driving the relationship between
free will and political ideology, at least in part. Finally, we conducted a mediation analysis
(10,000 resamples) which revealed that wrongness judgments significantly statistically
mediated the relationship between political ideology and attributions of free will, b = -.03, se
= .01, 95% CI [-.05, -0.01], Z = -2.55, p =.011.
Discussion
Everett, Clark, et al. (2020, JPSP:PPID)
19
The results of Studies 4 and 5 suggest that the conservative tendency to ascribe more
free will than liberals does not (just) reflect a stable personality trait, but rather corresponds
to the perceived moral wrongness of actions. While conservatives consistently report higher
general, abstract belief in free will as measured by the FAD+, a different pattern emerges when
it comes to judgments of free will for specific events. For events that liberals and conservatives
were shown to see as equally morally wrong or praiseworthy, conservatives and liberals
attributed the same degree of free will (Study 4), and for events that liberals see as more
morally wrong than conservatives, it was liberals that attributed more free will (Study 5).
For Study 5, we wish to emphasize that these results only reached statistical
significance when using a larger sample than originally intended, indicating a smaller effect size
for the relationship between perceived moral wrongness and free will attributions for political
liberals than for political conservatives. Although we had not predicted this outcome in
advance, it seems consistent with our theoretical framework that liberals’ motivated free will
judgments would be weaker than conservatives’ due to their relatively weaker tendency to
perceive actions as morally significant. While it would be surprising if liberals did not exhibit
any motivated social cognition whatsoever (i.e., if they did not attribute more free will to
actions they perceived as wrong; Ditto et al., 2019a, 2019b), it would be equally surprising if
this tendency were equally as strong as it is for conservatives given the generally weaker
tendencies to moralize among liberals. A tempered or weaker effect, then, might be expected
for liberals.
Study 6
Thus far we have shown that conservatives moralize a wide variety of events more
than liberals (Study 1), and while they generally report greater free will beliefs than do liberals
(Studies 2-5), this tendency may be in part motivated by a greater desire to blame since
conservatives show no difference in free will attributions for actions seen as equally wrong as
liberals there is no difference in free will attributions (Study 4), and show lower free will
attributions for actions they find less wrong than liberals (Study 5). What we have not shown
yet, though, is whether a) conservatives would attribute more free will for events that they
found more morally wrong than liberals; and b) whether judgments of moral wrongness for
specific actions would mediate the relationship between conservative political orientation and
free will attributions.
In Study 6, therefore, we built upon the body of evidence from Studies 2-5 to look at
attributions of free will both for events that conservatives find more wrong, and for politically
neutral events in which there is no difference between conservatives and liberals’ perceptions
of wrongness. Our prediction was that conservatives would only attribute more free will for
the events they found morally wrong, and that this relationship between conservative ideology
and free will attributions would be statistically mediated by how wrong participants judged
the behavior to be. We test this in two separate studies: an initial test (Study 6a), and then a
later, pre-registered replication with a larger sample size (Study 6b).
Method
Pre-Testing. As in the previous studies, we pretested items for use in Study 6. Here,
however, we sought to identify items that either did or did not differ in perceived wrongness
Everett, Clark, et al. (2020, JPSP:PPID)
20
as a function of political orientation. American participants on MTurk (n = 110; 44 female, Mage
= 34) rated how liberal or conservative they were on a 1-7 scale. Participants were again
slightly liberal on average (M = 3.30), with more liberal views on social issues (M = 2.95) and
slightly more moderate views on economic issues (M = 3.65). All main political parties were
represented, with 22 Republicans, 42 Democrats, and 40 Independents. Participants were then
required to rate a series of 38 events and occurrences for how morally wrong they perceived
them to be.
In order to obtain politically neutral events, most of the items were drawn from the
30-item Moralization of Everyday Life Scale (MELS; Lovett et al., 2012). This consists of 30
common and everyday moral violations, the vast majority of which lack any clear connection
to political ideology (e.g., “Ava parks in a handicapped zone even though she is not
handicapped” or “Joseph starts smoking a cigarette in a non-smoking section of a restaurant”).
We primarily drew from the MELS to obtain the politically neutral items in order to reduce
any unconscious bias that could exist when creating items. In order to obtain items that
conservatives would judge as more wrong, we included 8 additional items that we devised
ourselves to parallel the format of the MELS items, but with content that we assumed would
be judged differently by conservatives and liberals (e.g.,Doug pretends to be ill to avoid being
sent to war”).
To select the five ‘politically neutral’ items on which conservatives and liberals did not
differ in judgments of wrongness, we used the same statistical cut-off point as in Study 4 (r
< .10; p > .40). To select the five ‘conservative wrong’ items, we used a statistical cut-off point
(r > .40; p < .005) to ensure that items were significantly associated with political orientation
such that that conservatives judged the items as significantly more wrong. Based on these
criteria, we selected 5 items from each category to use as the ‘Conservative Wrong’
dependent measures in this study (see Table 2).
The list of full items and correlations with political ideology can be seen in at the OSF.
Note that the items used in both Study 6a and 6b were the same, but we changed some of
the names of the protagonists in Study 6b. In the time between conducting our initial study
and the replication, a new paper was published that looked at the effect of names in vignette
experiments, providing a list of names matched in perceived age, warmth, and competence
(Newman et al., 2018). To eliminate any unintentional influence of the names we had chosen,
in Study 6b we therefore used names from this list.
Table 2: Items by Condition for Study 6b
Conservative Wrong
Politically Neutral
Jennifer has her second abortion in two
years
Evelyn is taking a casual walk around
the block on a snowy day, and she
notices a driver whose car is stuck in
the snow. She keeps walking rather than
stopping to see if she can help.
Everett, Clark, et al. (2020, JPSP:PPID)
21
David becomes addicted to drugs and
drops out of college.
Paul cleans out his closet and finds
several pieces of clothing he no longer
wears. He can dispose of them or drive
five miles to the Salvation Army and
drop them in their drop-off box. He
throws away the clothes.
Ann is out to dinner with some friends
and has some gas pains in her stomach.
She decides to release gas, even though
she knows it will make an awful smell.
Thomas starts smoking a cigarette in a
non-smoking section of a restaurant.
Sarah made an abstinence-until-
marriage vow when she was 16, but
now she’s 18 and in college, and she has
sex with a boyfriend with whom she’s
in love.
Rachel has sex with another man while
her boyfriend is out of town for the
weekend.
John pretends to be ill to avoid being
sent to war
Caroline goes into a college dorm
community bathroom and uses a
random toothbrush (belonging to
someone else) that is lying around. She
puts the toothbrush back and leaves.
Study 6a
Method
Open Science
We report all measures, manipulations, and exclusions. Results, analysis code, and
experiment materials are available for download from the OSF. This study received ethical
approval from the IRB of the University of Oregon (“The Effect of Culture on Attitudes and
Outcomes”; Protocol #10162012.023).
Participants. 301 American participants completed the survey online via MTurk.
Participants were excluded from data analysis if they took the survey more than once (n = 1)
or failed one or both of two simple checks in which they had to select a certain scale-point in
the FAD+, and then indicate at the end of the survey which event they did not answer questions
about (n = 4). This left a final sample of 294 participants (121 female; Mage = 32 years). Sample
size was determined by available funds and a goal of approximately 150 participants per
condition.
Design. Participants were randomly assigned to one of two conditions (Conservative
Wrong; Politically Neutral) in a between-subjects design. Participants were given a list of five
events that pre-testing showed were either differentially perceived to be more morally wrong
to conservatives than to liberals (Conservative Wrong), or five items on which conservatives
and liberals did not differ in their perceptions of wrongness (Politically Neutral). For each of
these five items (see Table 2), participants were asked to rate the degree of wrongness and
free will the actor had for each event using the same measures as Study 5 (how much
Everett, Clark, et al. (2020, JPSP:PPID)
22
responsibility, control, free will, and ability to do otherwise). The order of the events and the
five dependent variables for each event were all randomized, and at the end participants
completed the free will subscale of the FAD+.
Results
First, we again looked at correlations of political conservativism with general free will
belief. As in the previous studies, conservatism was significantly positively correlated with
scores on the free will subscale of the FAD+ (r = .38, p < .001), with conservatives showing
higher endorsement of both the more abstract (r = .33, p < .001) and moral responsibility
focused (r = .37 p < .001) items of the scale.
Next, we checked whether political orientation was indeed associated with greater
judgments of moral wrongness in the ‘conservative wrong’ condition. Despite pre-testing the
political neutral events to be equally morally wrong for liberals and conservatives, we actually
found in the main study that conservatives rated the ‘politically neutral’ events to be more
wrong (r = .24, p = .003), though unsurprisingly this relationship was much weaker than in the
‘conservative wrong’ condition (r = .57, p < .001). Across participants, wrongness ratings were
positively associated with free will attributions for the conservative wrong events (r = .23, p
= .005), but not the politically neutral events (r = .02, p = .77).
We then turned to our key analyses and examined how morality condition (-1 =
politically neutral; 1 = conservative wrong) and participant political orientation (centered)
interacted to predict attributions of free will. We observed a significant interaction of political
ideology and morality condition, b = .05, SE =.02, t = 2.18, p = .030, 95% CI [.00, .09], R2 = 0.21,
semi-partial r = 0.11, whereby as predicted conservatives attributed more free will for the
‘conservative wrong’ items, b = .08, SE =.03, t = 2.33, p = .021, 95% CI [.01, .15], R2 = .04, semi-
partial r = .19, but there was no difference in attributions of free will for the politically neutral
events, b = -.02, SE =.03, t = 0.57, p = .57, 95% CI [-.07, .04], R2 = .00, semi-partial r = -.05.
Finally, we tested our prediction that wrongness judgments would statistically mediate
the effect of political orientation on attributions of free will in the ‘conservative wrong’
condition. We began by conducting a moderated mediation analysis looking at whether the
wrongness condition moderated the mediation of wrongness judgments on the path from
political orientation to attributions of free will. We observed significant moderation, 95% CI
[.03, .10], which we then probed by looking at the mediation of wrongness judgments within
each condition (10,000 resamples). We found that the mediation effect was stronger (and
marginal) in the ‘conservative wrong’ condition, b = .04, se = .02, 95% CI [-.00, 0.09], Z = 1.79,
p =.07, than in the ‘politically neutral’ condition, b = .00, se = .01, 95% CI [-.01, 0.2], Z = 0.44, p
=.66, but note that wrongness judgments did not significantly mediate the effect of political
orientation on free will attributions in either condition at traditional levels of statistical
significance..
Discussion
In Study 6a, we tested the prediction that conservatives’ greater attributions of
free will would only be observed for events that they saw to be more morally wrong, and that
these perceptions of wrongness would mediate the relationship between political orientation
and free will attributions. Interestingly, despite having pre-tested items to be politically neutral
in perceived moral wrongness, conservatives still showed a tendency to find the politically
Everett, Clark, et al. (2020, JPSP:PPID)
23
neutral items more morally wrong, supporting the view that conservatives moralize more than
liberals. As predicted, higher political conservatism only predicted higher free will attributions
for actions that were pre-tested to be perceived as more wrong by conservatives. We also
found some suggestive, but not statistically significant, evidence that moral wrongness
statistically mediated the relationship between more conservative political ideology and
attributions of free will mainly in the ‘conservative wrong’ condition. In Study 6b we sought to
enhance our confidence in these results by conducting a new pre-registered replication with
a larger sample size.
Study 6b
Method
Open Science. Our design, hypotheses, and analysis plan were all pre-registered at
the Open Science Framework. We report all measures, manipulations, and exclusions, and
analysis code, and experiment materials are available for download from the OSF. This study
received ethical approval from the IRB of the University of British Columbia (“Social Impacts
of Emerging Technology”; Protocol #H18-02727).
Participants. 591 American participants completed the survey online via MTurk.
Participants were excluded from data analysis if they took the survey more than once (n = 10)
or failed one or both of two simple checks in which they had to select a certain scale-point in
the FAD+ and then indicate at the end of the survey which event they did not answer questions
about (n = 26). This left a final sample of 555 participants (280 female; Mage = 37 years).
Design. This study was identical to Study 6a, except that we replaced some of the
names of the protagonists to ensure they were matched in perceived age, warmth, and
competence (Newman et al., 2018; see discussion in pre-test).
Results
First, we again looked at correlations of political conservativism with general free will
belief. As in the previous studies, conservatism was significantly positively correlated with
scores on the free will subscale of the FAD+ (r = .33, p < .001), with conservatives showing
higher endorsement of both the more abstract (r = .29, p < .001) and moral responsibility
focused (r = .34 p < .001) items of the scale.
Next, we checked whether political orientation was indeed associated with greater
judgments of moral wrongness in the ‘conservative wrong’ condition. Here, political ideology
was significantly associated with thinking the moral transgressions were more wrong (r = .51,
p < .001), but in the matched condition with items pre-tested to be politically neutral, we
found no correlation of political ideology with wrongness ratings (r = .08, p = .21). Across
participants, wrongness ratings were positively associated with free will attributions, both for
the conservative wrong events (r = .24, p < .001) and the politically neutral events (r = .16, p
= .007).
We then turned to our key analyses and examined how morality condition (-1 =
politically neutral; 1 = conservative wrong) and participant political orientation (centered)
interacted to predict attributions of free will. We observed a significant interaction of political
ideology and morality condition, b = .05, SE =.02, t = 2.71, p = .007, 95% CI [.01, .08], R2 = .11,
semi-partial r = 0.11, whereby as predicted conservative ideology predicted more free will for
Everett, Clark, et al. (2020, JPSP:PPID)
24
the ‘conservative wrong’ items, b = .09, SE =.02, t = 3.76, p < .001, 95% CI [.04, .14], R2 = .05,
semi-partial r = .22, but there was no difference in attributions of free will for the politically
neutral events, b = .00, SE =.02, t = 0.01, p = .99, 95% CI [-.05, .05], R2 = .00, semi-partial r =
-.00.
Finally, we observed significant moderated mediation, 95% CI [.04, .08], which we then
probed by looking at the mediation of wrongness judgments within each condition (10,000
resamples). With this larger sample size, we found that wrongness judgments statistically
significantly mediated the effect of more conservative political ideology on higher free will
attributions in the ‘conservative wrong’ condition, b = .04, se = .02, 95% CI [.01, .06], Z = 2.41,
p =.02, but not in the ‘politically neutral’ condition, b = .05, se = .00, 95% CI [-.00, .01], Z =
1.15, p =.25.
Discussion
In Study 6b, we conducted a pre-registered replication of Study 6a to test our
prediction that conservatives’ greater attributions of free will would only be observed for
events that they saw to be more morally wrong, and that these perceptions of wrongness
would mediate the relationship between political orientation and free will attributions.
Confirming our predictions, political ideology only predicted free will attributions for actions
pre-tested to be perceived as more wrong by conservatives, and the relationship between
ideology and free will attributions was statistically mediated by perceptions of moral
wrongness. Together, the results of Studies 6a and 6b provide further support for our main
hypothesis that the relationship between political ideology and free will beliefs can be at least
partially linked to differential perceptions of moral wrongness and blameworthiness.
Study 7
It seems likely that political differences in attributions of free will (Study 3) are not
reflecting some genuine disagreement about the basis of human freedom, given that these
differences appear only for those events that are also judged as differentially morally wrong
(Studies 4-6). Philosophically, the concept of free will should be domain-general and enduring:
the deterministic laws of the universe are unlikely to be sensitive to the specific moral qualities
of the action. Perhaps psychologically, though, people have heuristics about which actions
require more or less free will (regardless of the philosophical coherence of such a stance). If
this were the case, perhaps conservatives and liberals simply have different heuristics about
which actions are more free, and higher perceptions of freedom cause stronger judgments of
moral wrongness rather than the reverse motivated reasoning pattern we hypothesize (where
moral assessments influence free will assessments). In our final set of studies, therefore, we
wanted to conduct an even more direct test of the specifically motivated basis of political
differences in attributions of free will by looking at whether liberals and conservatives would
differentially attribute free will for the same action that differed only on who performed it so
as to isolate motivations to blame. We do this across four studies, Studies 7a-7d. These are
the only four studies we have conducted testing this interaction (i.e., there are no file drawer
studies).
Study 7a
Method
Everett, Clark, et al. (2020, JPSP:PPID)
25
Open Science and Ethics Statement
We report all measures, manipulations, and exclusions. Results, analysis code, and
experiment materials are available for download at the OSF. This study, along with Studies 7b,
7c, and 7d, received ethical approval from the IRB of the University of British Columbia (“Social
Impacts of Emerging Technology”; Protocol #H18-02727).
Participants. 600 American participants completed the survey online via MTurk.
Participants were excluded from data analysis if they took the survey more than once (n = 1)
or failed one or both of two simple checks in which they had to select a certain scale-point in
the FAD+, and then correctly report at the end of the study what the protagonist in the
vignette had done (n = 7). This left a final sample of 592 participants (321 female; Mage = 38
years).
Pre-Tests. We completed two pre-tests on MTurk to find events to use for Study 7a.
In both pre-tests we only asked participant to rate wrongness and blameworthiness and not
free will to avoid biasing item selection.
In our first pre-test (n = 249), we used six conventional moral violations and simply
varied whether the actor was characteristically left- or right-wing (see materials on the OSF
for the wording of all events and the corresponding results). For example, in one event we
described to participants someone “who regularly smokes cigarettes indoors around his small
child”, and varied whether this person was “an activist who frequently protests against Wall
Street” or “a wealthy Wall Street banker”. In another event we described someone who was
“recently arrested for driving while approximately two drinks over the legal driving limit”, and
varied whether this person conducts research for a think tank and investigates how pervasive
racism might explain disparate outcomes between racial groups” or “conducts research for a
think tank and investigates genetic differences between races that might explain disparate
outcomes between racial groups”. Across these events, we failed to find significant interactions
between participants’ own political beliefs and the apparent political beliefs of the target:
participants tended to rate the moral violation as equally wrong regardless of whether it was
a left or right-wing person doing it. In other words, for these general behaviors, liberals and
conservatives displayed no favoritism toward their own political ingroup members in terms of
their moral evaluations. While pleasant from a normative perspective, this rendered these
events unfit for our purposes in the main study.
To this end, we conducted a second pre-test (n = 300) in which we gave participants
longer descriptions of moral violations that an actor performed in the context of achieving
certain political goals (see materials on the OSF for the wording of all events and the
corresponding results). This was done for two reasons: first, by giving more detailed
information we hoped to make it less obvious to participants that we expected them to use
the actor’s political affiliation as a cue to wrongness; and second, to leverage partisan
intergroup cognition that would plausibly make the moral violations more acceptable if it was
done against “them”, for “us”. One item concerned using charity funds to buy a more expensive
suit for oneself in the hopes this will impress donors for one’s charity; one concerned
blackmailing a town mayor about his use of prostitutes in order to get him to publicly support
one’s social movement; and one concerned a student who gets violent at a protest. As before,
in this pre-test we measured only participants ratings of wrongness and blameworthiness, not
Everett, Clark, et al. (2020, JPSP:PPID)
26
free will, in order to avoid biasing our item selection. There was no significant interaction
between participant political orientation and the political leanings of the target on ratings of
wrongness for the first item, but there was for the second and third items, so these were the
ones used in Study 7a.
Design. All participants read a short story about an immoral action performed by
“Noah”. In a between-subjects design, participants were randomly assigned to learn that Noah
was either politically conservative or liberal and performed the action in the context of
achieving these political ends. To increase confidence in the generalizability of our results, as
an additional between-subjects factor we used two different events (violence at a protest and
blackmail). As noted above, pre-testing revealed that in both events, the same action was
perceived as differently wrong depending on the participant’s political beliefs and the beliefs
of the actor.
In the first story, participants read that Noah was a college student who gets caught
up in a student protest and “In the heat of the moment, throws a glass bottle at the other
protestor though luckily, it didn’t hit them”. For half of our participants, Noah was described
as supporting “the Antifa movement a left-wing protest movement that often gets into
violent clashes with more right-wing protesters”. The other half of participants read instead
that Noah supported “the Patriot Movement – a right-wing protest movement that often gets
into violent clashes with more left-wing protesters.” Full text can be seen at the OSF.
In the second story, participants read that Noah was the local chapter head of a
political group who blackmails the mayor of his town in order to obtain support for his
movement. For half of our participants, Noah was described was leading the local Black Lives
Matter” movement, and blackmails the mayor into “calling for stricter punishment for police
who kill black people”. For the other half, Noah was described as leading the local “Blue Lives
Matter” and instead blackmails the mayor into “calling for stricter punishment for people that
kill law enforcement officers”.
After reading the story, participants were asked to rate the degree of free will,
wrongness, and blameworthiness associated with each described event (as in Studies 3-6). The
order of the five dependent variables was randomized, and at the end participants completed
the free will subscale of the FAD+.
As anticipated, results were the same for both stories and so we report our main
analyses combining across the two. Full results for each individually can be seen at the OSF.
Results
First, we looked at correlations of political conservativism with general free will belief.
As in the previous studies, conservatism was significantly positively correlated with scores on
the free will subscale of the FAD+ (r = .25, p < .001), with conservatives showing higher
endorsement of both the more abstract (r = .23, p < .001) and moral responsibility focused (r
= .23 p < .001) items of the scale.
Second, we confirmed that participants did perceive the moral violation to be more
wrong when the agent and the agent’s actions opposed their own political group than when
the agent and the agent’s actions supported their own political group. A regression analysis
revealed a significant interaction between the agent’s political position (-1 right-wing target,
+1 left-wing target) and participant’s own political orientation (centered) on ratings of moral
Everett, Clark, et al. (2020, JPSP:PPID)
27
wrongness, b = .21, SE =.03, t = 6.37, p < .001, 95% CI [.15, .28], R2 = 0.07, semi-partial r = 0.25.
Whereas participant conservatism predicted thinking the left-wing agent performed a more
morally wrong action, b = .22, SE =.05, t = 4.60, p < .001, 95% CI [.12, .31], R2 = 0.07, semi-
partial r = 0.26, political liberalism predicted thinking the right-wing agent performed a more
morally wrong action, b = -.20, SE = .05, t = -4.41, p < .001, 95% CI [-.29, -.11], R2 = 0.06, semi-
partial r = -0.25.
Finally, we turned to our main analysis of whether participants attributed differential
free will for the same action depending on their own political beliefs (centered) and the
political beliefs of the actor (-1 right-wing target, +1 left-wing target). The interaction did not
reach statistical significance, b = .03, SE =.02, t = 1.70, p = .089, 95% CI [-.01, .07], R2 = 0.01,
semi-partial r = 0.07, and simple effects indicated that political conservatism predicted lower
attributions of free will towards the right-wing target, b = -.07, SE =.03, t = -2.38, p = .018, 95%
CI [-.12, -.01] R2 = 0.02, semi-partial r = -0.14, but not the left-wing target b = .00, SE =.03, t
= 0.12, p = .91, 95% CI [-.05, .06] R2 = 0.00, semi-partial r = 0.01. Probing this further by looking
at attributions of free will by self-identified Republican (n = 160) and Democrat (n = 249)
participants, we found a significant interaction of participant political affiliation and the agent’s
political beliefs on attributions of free will, F(1,405) = 3.99, p = .047, p = .01, partial η2 = .01.
Simple effects results revealed that whereas Democrats attributed marginally more free will
to the right-wing than left-wing agent, t(246.83) = -1.80, p = .074, d = 0.23, Republicans,
however, did not significantly differ in the degree of free will they attributed to both agents,
t(158) = 1.12, p = .266, d = 0.18
Discussion
In Study 7a, we took a different approach to exploring the motivated basis of
free will belief by looking at whether participants’ political orientation might lead to differential
attributions of free will for the same action depending on who performed it. Using two
different moral violations that yielded the same results, we found tentative but weak -
evidence in support of our predictions: there was the suggestion of differential free will
attributions for the same action depending on participants’ own political beliefs, though this
was not significant.
Though these results appear to only weakly align with the findings of the
previous studies, ceiling effects in Study 7a’s measures of both moral violation and free will
may have limited our ability to properly test our hypothesis. Regardless of the actor’s political
beliefs, both Republican and Democrat participants thought the moral violation very wrong
(all means > 5.8 on a 1-7 scale) and more problematically - attributed high free will to both
actors (all means > 6.19 on a 1-7 scale). Previous work has shown that because people like to
see themselves as fair and objective judges, motivated reasoning is most pronounced in
situations where plausibility constraints are loose and ambiguous (Ditto & Boardman, 1995;
Ditto & Lopez, 1992; Dunning et al., 1995). The events described clear moral violations, and
with such high free attributions across the board, there might have been little room for
motivated blame to selectively increase free will attributions for the politically-opposing agent.
Given this, in Study 7b we sought to run the study again, but with events that were more
morally ambiguous.
Study 7b
Everett, Clark, et al. (2020, JPSP:PPID)
28
Method
Open Science
We report all measures, manipulations, and exclusions. Results, analysis code, and
experiment materials are available for download from the OSF.
Participants. 600 American participants completed the survey online via MTurk.
Participants were excluded from data analysis if they took the survey more than once (n = 0)
or failed one or both of two simple checks in which they had to select a certain scale-point in
the FAD+, and then correctly report at the end of the study what the protagonist in the
vignette had done (n = 34). This left a final sample of 566 participants (274 female; Mage = 38
years).
Design. This study had the same design as Study 7a, where again participants read a
story describing a moral violation and in a between-subjects design we manipulated whether
the actor was politically right-wing or left-wing and performed the action in the context of
achieving those political ends. Again, to increase confidence in the generalizability of our
results, as an additional between-subjects factor, we used two different events, but sought to
use scenarios more ambiguously immoral than in Study 7a (staying silent about disruption at
a protest; not retracting a magazine expose based on an unreliable source). Pre-testing (see
details at OSF) revealed that in both events, the same action was perceived as differently wrong
depending on the participant’s political beliefs and the beliefs of the actor (and that overall,
the items were rated as less wrong than the items in Study 7a).
In the first story, participants read about Anna”: an editor and journalist who publishes
a widely-discussed and influential expose about a sex scandal amongst the highest members
of a political party. Participants were told that “A few months after this expose was released,
Anna learns about serious problems with her source which makes her doubt whether they
are telling the truth. The source is adamant they are telling the truth, but Anna has found that
they have made false accusations in the past. Usually Anna would be required to publish a
retraction, but because the evidence of lying is not conclusive, she does not issue a correction.”.
For half of our participants, Anna was described as being the editor of a “prominent left-wing
online journal known for regularly and strongly attacking the Republican party”, who publishes
an expose about a sex scandal in the Republican party. For the other half, Anna was described
as being the editor of a “prominent right-wing online journal” who publishes an expose about
a sex scandal in the Democratic party.
In the second story, participants read about Noah, a politically active college student
who is peacefully protesting a visit by a controversial speaker. During the protest, Noah
observes a fellow student “cut the sound for the speaker, meaning that no-one can hear the
talk and it has to be called off”. Participants were told that the police want to find and
prosecute the activist “to make an example of them, drawing attention to what they see as a
rising problem of right/left-wing activists”, but “When the police come and question Noah, he
says that he did not see who cut the sound”. Half of our participants were told that both Noah
and the activist who cut the sound were members of the “Young Republicans Society” and
were protesting a controversial left-wing speaker. The other half were told that Noah and the
activist were members of the “Young Democrats Society” and were protesting a right-wing
speaker (see full wording on the OSF).
Everett, Clark, et al. (2020, JPSP:PPID)
29
The dependent measures were the same as in Study 7a and because results were again
the same for both stories, we report our main analyses combining across the two (see OSF
for results for event separately).
Results
First, we looked at correlations of political conservativism with general free will belief.
As in the previous studies, conservatism was significantly positively correlated with scores on
the free will subscale of the FAD+ (r = .16, p < .001), with conservatives showing higher
endorsement of both the more abstract (r = .29, p < .001) and moral responsibility focused (r
= .26 p < .001) items of the scale.
Second, we confirmed that participants did perceive the moral violation to be more
wrong when the target opposed rather than supported their own political interests. A
regression analysis revealed a significant interaction between the agent’s political position (-1
right-wing target, +1 left-wing target) and participant’s own political orientation (centered) on
ratings of moral wrongness, b = .27, SE =.04, t = 6.82, p < .001, 95% CI [.19, .35], R2 = 0.13,
semi-partial r = 0.27. Whereas participant conservatism predicted thinking the left-wing agent
performed a more morally wrong action, b = .50, SE =.05, t = 9.31, p < .001, 95% CI [.39, .60],
R2 = 0.23, semi-partial r = 0.48, despite our pre-testing, we found no effect of political
orientation on wrongness ratings of the conservative target, b = -.04, SE = .06, t = -0.72, p
= .48, 95% CI [-.16, -.07] R2 = 0.00, semi-partial r = -0.04.
Finally, we turned to our main analysis of whether participants attributed differential
free will for the same action depending on their own political beliefs (centered) and the
political beliefs of the actor (-1 right-wing target, +1 left-wing target). We observed a significant
interaction, b = .09, SE =.02, t = 3.85, p < .001, 95% CI [.05, .14], R2 = 0.03, semi-partial r =
0.16, where simple effects indicated that political conservatism predicted greater attributions
of free will towards the left-wing target, b = .13, SE =.03, t = 3.90, p < .001, 95% CI [.06, .19],
R2 = 0.05, semi-partial r = 0.23, but there was no significant effect of participant political
orientation on attributions of free will towards the right-wing target (though the direction of
results was as expected), b = -.06, SE =.04, t = 0.11, p = .11, 95% CI [-.13, .01], R2 = 0.01, semi-
partial r = -0.10 (see Figure 3). Again probing this further by looking at attributions of free will
by self-identified Republican (n = 140) and Democrat (n = 235) participants, we found a
significant interaction of participant political affiliation and the agent’s political beliefs on
attributions of free will, F(1,371) = 7.46, p = .007, partial η2 = .02. Simple effects results revealed
that Democrats attributed significantly more free will to the right-wing than left-wing agent,
t(232.98) = 2.17, p = .031, d = 0.28, and there was a trend for Republican participants to
attribute more free will to the left-wing agent than the right wing one, though this was only
marginally significant, t(128.05) = -1.80 p = .074, d = 0.30
Everett, Clark, et al. (2020, JPSP:PPID)
30
Figure 3. Attributions of free will as a function of participant political ideology and the
political beliefs of the actor (Study 7b).
Discussion
In Study 7b we found further evidence for the motivated basis of free will belief by
showing that participants attribute different levels of free will for the same action depending
on whether someone shares or disagrees with their own political stance. Note, however, that
our claim is not that political conservatives and liberals will always differentially attribute free
will for the same event depending on who does it, just as we are not claiming that political
conservatives and liberals will always make in-group favouring biased moral judgments (indeed,
our first pre-test for Study 7a shows that this is not the case). Instead, we arguing that where
there is a motivation to differentially blame the target that emerges in biased ratings of moral
wrongness, there will typically be a motivation to differentially attribute free will too. To
confirm our results, in Study 7c we opted to conduct our second pre-registered replication
of this series.
Study 7c
Method
Open Science
Our design, hypotheses, and analysis plan were all pre-registered at the Open Science
Framework. We report all measures, manipulations, and exclusions, and analysis code, and
experiment materials are available for download from the OSF.
Everett, Clark, et al. (2020, JPSP:PPID)
31
Participants. 595 American participants completed the survey online via MTurk.
Participants were excluded from data analysis if they took the survey more than once (n = 0)
or failed one or both of two simple checks in which they had to select a certain scale-point in
the FAD+, and then correctly report at the end of the study what the protagonist in the
vignette had done (n = 58). This left a final sample of 537 participants (255 female; Mage = 38
years).
Design. This study had the same design as Study 7b.
Results
First, we looked at correlations of political conservativism with general free will belief.
As in the previous studies, conservatism was significantly positively correlated with scores on
the free will subscale of the FAD+ (r = .28, p < .001), with conservatives showing higher
endorsement of both the more abstract (r = .26, p < .001) and moral responsibility focused (r
= .25 p < .001) items of the scale.
Second, we confirmed that participants did perceive the moral violation to be more
wrong when the agent opposed rather than aligned with their own political interests. A
regression analysis revealed a significant interaction between the agent’s political position (-1
right-wing target, +1 left-wing target) and participant’s own political orientation (centered) on
ratings of moral wrongness, b = .25, SE =.04, t = 5.86, p < .001, 95% CI [.17, .33], R2 = 0.09,
semi-partial r = 0.24. Whereas participant conservatism predicted thinking the left-wing agent
performed a more morally wrong action, b = .41, SE =.06, t = 6.87, p < .001, 95% CI [.30, .53],
R2 = 0.15, semi-partial r = 0.39 despite our pre-testing (but as in Study 7b), we found no effect
of political orientation on wrongness ratings of the conservative target, b = -.08, SE = .06, t =
-1.39, p = .17, 95% CI [-.02, -.03], R2 = 0.01, semi-partial r = -0.08.
Finally, we turned to our main analysis of whether participants attributed differential
free will for the same action depending on their own political beliefs (centered) and the
political beliefs of the actor (-1 right-wing target, +1 left-wing target). Despite the fact that this
was a direct registered replication on Study 7b in which we found a significant interaction, the
interaction here was non-significant, b = .01, SE =.03, t = 0.32, p = .75, 95% CI [-.06, .08], R2 =
0.00, semi-partial r = 0.03. Surprisingly, political conservatism did not predict greater
attributions of free will towards either the left-wing target, b = .01, SE =.03, t = 0.32, p = .75,
95% CI [-.06, .08], R2 = 0.00, semi-partial r = 0.02 , or the right-wing target, b = -.02, SE =.04, t
= -0.58, p = .56, 95% CI [-.09, .05], R2 = 0.00, semi-partial r = -0.04. Similarly, when looking at
attributions of free will by self-identified Republican (n = 131) and Democrat (n = 232)
participants, we found no interaction of participant political affiliation and the agent’s political
beliefs on attributions of free will, F(1,359) = 0.59, p = .44, partial η2 = .00.
Discussion
In Study 7c we conducted our second pre-registered replication in this package
of studies. We sought to replicate our finding from Study 7b that participants attribute different
levels of free will for the same action depending on whether someone shares or disagrees
with their own political stance. Surprisingly, our key interaction was not significant despite
using the same materials, sample recruitment, participant exclusion rules, and data analysis.
How to explain this? One possibility is that on the 1-7 scale, we had high free will
attributions across the board (>5.5), potentially leading to ceiling effects. Another possibility is
Everett, Clark, et al. (2020, JPSP:PPID)
32
the time we ran the study. The data for Study 7c were collected between 12am and 6am PST
(3am to 9am EST). With our restriction that participants needed to be from the U.S. to
complete the study, this meant that participants were completing the study in the middle of
the night. Despite employing the same attention and comprehension checks used across the
studies in this paper, it is possible that participants were still not attending as thoroughly as
they would have been if they hadn’t been doing it in the middle of the night. Perhaps the null
results in Study 7c were just a fluke, potentially influenced by ceiling effects and/or the unusual
time of data collection. Or perhaps it was the significant effect in Study 7b that was the fluke,
and that there was no real effect to find at all. In Study 7d we sought conduct one final pre-
registered replication to settle the matter, making sure to collect the data during the day while
also using a larger sample size and increasing the scale variance from 1-7 to 0-100.
Study 7d
Method
Open Science
Our design, hypotheses, and analysis plan were all pre-registered at the Open Science
Framework. As for all studies in this paper, we report all measures, manipulations, and
exclusions, and analysis code, and experiment materials are available for download from the
OSF.
Participants. We recruited 900 American participants via MTurk, 884 of whom
completed the survey. Participants were excluded from data analysis if they took the survey
more than once (n = 1) or failed one or both of two simple checks in which they had to select
a certain scale-point in the FAD+, and then correctly report at the end of the study what the
protagonist in the vignette had done (n = 68). This left a final sample of 815 participants (478
female; Mage = 37 years).
Design. This study had the same design as Study 7b and 7c, with the exception that,
in Study 7d, the dependent measures were on a 0-100 scale instead of 1-7 scale, and we also
included an additional question about how severely participants thought the agent should be
punished for their action.
Results
First, we looked at correlations of political conservativism with general free will belief.
As in the previous studies, conservatism was significantly positively correlated with scores on
the free will subscale of the FAD+ (r = .34, p < .001), with conservatives showing higher
endorsement of both the more abstract (r = .30, p < .001) and moral responsibility focused (r
= .32 p < .001) items of the scale.
Second, we confirmed that participants did perceive the moral violation as more wrong
when the agent aligned withrather than opposedtheir own political interests. A regression
analysis revealed a significant interaction between the agent’s political position (-1 right-wing
target, +1 left-wing target) and participant’s own political orientation (centered) on ratings of
moral wrongness, b = 5.33, SE =.58, t = 9.12, p < .001, 95% CI [.4.19, 6.48], R2 = 0.13, semi-
partial r = 0.30. Participant conservatism positively predicted thinking the left-wing agent
performed a more morally wrong action, b = 7.82, SE =.86, t = 9.07, p < .001, 95% CI [6.12,
9.51], R2 = 0.17, semi-partial r = 0.41, and negatively predicted wrongness ratings for the same
Everett, Clark, et al. (2020, JPSP:PPID)
33
action when performed by a right-wing agent, b = -2.85, SE = .79, t = -3.60, p < .001, 95% CI [-
4.41, -1.29], R2 = 0.03, semi-partial r = -0.17.
Finally, we turned to our main analysis of whether participants attributed differential
free will for the same action depending on their own political beliefs (centered) and the
political beliefs and interests of the actor (-1 right-wing target, +1 left-wing target). We
observed a significant interaction, b = 1.45, SE =.035, t = 4.15, p < .001, 95% CI [.77, 2.14], R2
= 0.03, semi-partial r = 0.14. Simple effects indicated that political conservatism negatively
predicted attributions of free will towards the right-wing target, b = -2.03, SE =.48, t = -4.27, p
< .001, 95% CI [-2.97, -1.10], R2 = 0.04, semi-partial r = -0.21. The effect of participant
conservatism on free will attributions for the left-wing agent was not significant, though the
direction of results was as expected: b = 0.88, SE =.52, t = 1.70, p = .090, 95% CI [-.14, 1.89],
R2 = 0.01, semi-partial r = 0.08 (see Figure 4). Again probing this further by looking at
attributions of free will by self-identified Republican (n = 180) and Democrat (n = 354)
participants, we found a significant interaction of participant political affiliation and the agent’s
political beliefs on attributions of free will, F(1,530) = 6.80, p = .009, partial η2 = .01. Simple
effects results revealed that Republicans attributed significantly more free will to the left-wing
agent, t(176.46) = 2.35, p = .020, d = 0.35, but Democrats did not differ in attributions
depending on the agent, t(345.74) = -1.18, p = .24, d = -0.13.
Figure 4. Attributions of free will as a function of participant political ideology and the
political beliefs of the actor (Study 7d).
Discussion
Everett, Clark, et al. (2020, JPSP:PPID)
34
In Study 7d we conducted a final pre-registered replication in attempt to clarify
inconsistencies in our findings in Study 7a-c. With a larger sample size and increasing the scale
variance from 1-7 to 0-100, we replicated the significant interaction found in Study 7b:
participants attributed different levels of free will for the same action depending on whether
the target shared or disagreed with their own political stance.
Study 7e: Meta-analysis of Studies 7a-7d
Because the effect sizes for the interaction between ideology and the experimental
manipulation varied somewhat across Studies 7a-7d, we concluded with mini-meta-analyses of
the interaction effect for these four studies. As noted in the introduction, we have reported
in the manuscript all the studies we have run testing this interaction.
Method
We used procedures outlined by Goh and colleagues (2016) for conducting
mini-meta-analyses on the four interaction effects between ideology and the experimental
manipulation in Studies 7a-7d. Because the methods used across studies were very similar (in
which case fixed effects, which weights by sample size, might be preferred), but not quite
identical (in which case random effects, which treats all effects equally, might be preferred), we
report both fixed and random effects, though note the random effects approach is very
conservative with only four effect sizes (Goh et al., 2016). We used semipartial rs as estimates
of the effect size for the interaction terms. For fixed effects, the four rs were Fisher’s Z
transformed to rzs, which were then weighted and averaged using the following formula:
Weighted r̄z = Σ ([N-3] rz) / Σ (N-3). The weighted r̄z was then converted back to a Pearson’s
r correlation for presentation. To determine statistical significance, we utilized the Stouffer’s Z
test, in which the p values for each interaction effect were converted to Zs, combined using
the following formula: Zcombined = Σ Z / sqrt(k), and then converted back to ps for presentation.
For random effects, we conducted a single sample t-test of the semipartial rs.
Results
For both random effects (r = .101, p = .048) and fixed effects (r = .106, p
< .0001), there were small but statistically significant effects of the interaction between
ideology and the experimental manipulation on free will attributions.
Discussion
Though we found slightly different results across Studies 7a-7d, when analyzed
together in a meta-analysis, these studies provided evidence for the existence of the
hypothesized interaction. Thus, free will attributions for identical actions, vary as a function of
whether those actions oppose or align with one’s own political interests. In our view, these
results provide the most compelling support for our contention that differences in perceptions
of free will between liberals and conservatives reflect varying motivations to blame rather
than principled and consistent beliefs about human freedom and control.
General Discussion
Personal responsibility and autonomy feature heavily in conservative ideologies
(Reagan, 1968; Thatcher, 1981), and recent evidence has suggested that conservatives exhibit
greater belief in free will than liberals (Carey & Paulhus, 2013). In this paper we sought to test
Everett, Clark, et al. (2020, JPSP:PPID)
35
the hypothesis that political differences in free will belief do not reflect some genuine
principled disagreement about the metaphysical nature of human freedom, but rather are
largely explicable through motivated reasoning. Previous work has shown that, in general, free
will beliefs are motivated by desires to punish others and to justify holding them morally
responsible (Clark et al., 2014, 2017, 2018, 2019), and in this paper we tested whether the
reason that conservatives tend to attribute more free will is because they have a stronger
tendency to moralize, perceiving a wider spectrum of transgressions for which moral
responsibility must be assigned and moral blame attributed.
Overview of Findings
In Study 1, we directly tested our background assumption that political conservatives
have a stronger tendency to moralize (though this does not, of course, mean that conservatives
always see things as more wrong than liberals: see Study 5). Meta-analysing five new studies
drawn from a variety of populations (total n = 308,499), we show that political ideology was
consistently associated with moralization. Political conservatism was associated with greater
wrongness judgments in the Moral Foundations Questionnaire (Study 1a); with perceiving
minor everyday moral violations to be more wrong (Study 1b); with judging a variety of traits
as more necessary for someone to be a morally good person (Study 1c); with perceiving
someone as a bad person based on an image of their face (Study 1d); and as rating personality
traits as more morally bad (Study 1e).
In Study 2, we turned to investigate our main question of political differences in free
will belief. Using a large sample of yourmorals.org data (n = 14,707) we looked at the relationship
among political ideology, free will belief, and moralization. Replicating previous work (Carey &
Paulhus, 2013), we found that political conservatism was indeed associated with a greater belief
in abstract free will. As would be expected if this relationship arises partially from moralization,
we found that beliefs that people are morally responsible for their bad behaviors statistically
mediated the relationship between more conservative ideology and stronger beliefs in free
will.
In Study 3, we turned away from reports of general, abstract belief in free will to look
at attributions of free will for specific events. More specifically, we wanted to look at
attributions of free will for both positive and negative events. We found that political ideology
predicted free will attributions overall, and there was an indication that this relationship was
stronger for negative than positive events. Conceptually, the capacity for free will should hold
whether one experiences good or bad outcomes, and so if ideology is genuinely related to an
abstract belief in free will, there should have been no difference depending on the valence of
the outcomes. The fact that conservatism predicted higher free will attributions mainly for
negative events is consistent with the claim that free will attributions are not, or at least not
solely, reflecting some dispositional variance in a belief in human autonomy, but a more basic,
social psychological phenomenon likely one relating to blame.
In Study 4, we considered that if political conservatives are more likely to attribute
free will because they see more things as morally wrong, this association should not be
observed when looking at attributions of free will for specific events that are perceived as
equally (im)moral for liberals and conservatives. Using a range of both moral and non-moral,
and positive and negative events that were pre-tested to be equally matched in morality and
Everett, Clark, et al. (2020, JPSP:PPID)
36
valence amongst political liberals and conservatives, we found no relationship between political
ideology and free will attributions. Again, these results were consistent with our theory that
differences in conservatives’ and liberals’ perceptions of free will may be partially due to
differences in moralization, rather than representing any generalized, abstract belief that human
behaviors are freely chosen.
In Study 5, we tested whether - if the typical tendency for conservatives to endorse
stronger beliefs in free will is due to blame motives - this association would be reversed for
those less frequent events that conservatives perceive as less morally wrong than their liberal
counterparts. Whereas conservatives again reported higher general, abstract belief in free will
as measured by the FAD+, when using events that liberals saw as more morally wrong than
conservatives, it was liberals that attributed more free will.
In Study 6, we tested the prediction that conservatives’ greater attributions of free will
would only be observed for events that they saw to be more morally wrong, and that these
perceptions of wrongness would mediate the relationship between political orientation and
free will attributions. In two studies an initial one (Study 6a), and then a pre-registered
replication with a larger sample size (Study 6b) we confirmed our predictions, showing that
political ideology only predicted free will attributions for actions that pre-tested to be
perceived as more wrong by conservatives, and this relationship was statistically mediated by
perceptions of moral wrongness. Again, this supports our contention that political differences
in free will beliefs are linked to differential perceptions of moral wrongness and
blameworthiness.
In Study 7, we took a different approach. Instead of manipulating the moral content of
the event and then looking at how political ideology is associated with free attributions, we
looked at whether liberals and conservatives would differentially attribute free will for the
same action depending on who performed it. In Study 7a, we used two different moral violations
(violence at a protest; blackmail) and told participants that the action was performed by a left-
wing or right-wing agent. Across both events we found tentative but weak - evidence in
support of our predictions: there was the suggestion of differential free will attributions for
the same action depending on participants’ own political beliefs, though this was not significant.
Recognizing that this could have been because the events described clear and intentional moral
violations, whereas motivated reasoning is most pronounced in situations where plausibility
constraints are loose and ambiguous, in Study 7b we used the same design but with events
that were more morally ambiguous (staying silent about disruption at a protest; not retracting
a magazine expose based on an unreliable source). Confirming predictions, we found that
participants attributed different levels of free will for the same action depending on whether
the actor shared or disagreed with their own political stance. To confirm our results, in Study
7c we conducted a pre-registered replication. Surprisingly, our key interaction was not
significant despite using the same materials, sample recruitment, participant exclusion rules,
and data analysis. Because this could have been caused by technical problems, two months
later we ran a second pre-registered replication for Study 7d. With a larger sample size and
an increase in the scale variance from 1-7 to 0-100, we replicated the significant interaction as
predicted, showing that participants attributed different levels of free will for the same action
depending on whether the target shared or disagreed with their own political stance. Finally,
Everett, Clark, et al. (2020, JPSP:PPID)
37
because the effect sizes for the interaction between ideology and the experimental
manipulation varied across Studies 7a-d, we concluded with mini-meta-analyses of the
interaction effect for these four studies. These showed small but statistically significant effects
of the interaction between ideology and the experimental manipulation on free will
attributions.
Together, these 14 studies paint a picture whereby conservatives’ comparatively strong
free will beliefs are linked to a desire to hold others accountable for transgressions, and not
merely reflective of a generalized and abstract metaphysical belief concerning the nature of
human agency. These results are consistent with our theory that the relationship between
political ideology and free will beliefs can be at least partially explained as a manifestation of
motivated cognition: People endorse the idea of free will in order to justify their desire to
blame others for moral wrongdoing (Clark et al., 2014, 2017, 2019) with conservatives
reporting higher free will beliefs in part because they find a wider spectrum of issues to be
more morally wrong. Thus, to understand apparent differences in free will belief we need not
appeal to some special kind of “Republican Brain” (Mooney, 2012) that differs from that of
liberals on metaphysical beliefs about human autonomy. Instead where we do see differences
in free will beliefs, they are derived from more basic social cognitive processes that are shared
regardless of the political party one votes for, and more basic differences in moralization.
Limitations
There are, as with any project, certain limitations to our analysis here. First, it is
important to note that whereas we have focused on how moralization can help explain the
relationship between political ideology and free will beliefs, we do believe that the causal
relationships between our variables of interest (political ideology, free will beliefs, and
moralization) are likely complex and involve feedback loops. There is already debate in the
field as to whether it makes more sense to conceptualise political orientation as being caused
by, or following from, different moral intuitions. Whereas some scholars have argued that
moralization tendencies and a broader moral domain can explain why people are attracted to
particular political ideologies (e.g., Graham et al., 2009; Haidt, 2012), others have argued for
the opposite casual direction whereby individual differences associated with ideological beliefs
relating to system justification and social dominance orientation are what explains differences
in the moral domain (e.g., Hatemi et al., 2019; Kugler et al., 2014; K. B. Smith et al., 2017).
Similarly, we have treated political ideology as a predictor variable and free will attributions as
an outcome variable, but differences in perceived freedom likely lead to particular ideological
views as well. We have little doubt that all three of our variables influence and reinforce one
another. What we aimed to test here is whether experimentally manipulating differences in
blame desires between ideological groups can generate (or eliminate) differences in free will
attributions. We found consistent support for this, and thus, we think that this causal pattern
does exist, but it does not preclude other causal patterns from also existing. It is thus possible
that conservatives have proclivities for free will beliefs above and beyond moralization
tendencies, and an individual's moralizing could lead to a more conservative ideology that
further exaggerates their moralizing-related affinity for free-will beliefs. Investigating the
possible feedback loops will be an interesting direction for future research. That said, we do
believe that the model we have focused on here (political orientation predicts free will
Everett, Clark, et al. (2020, JPSP:PPID)
38
attributions partially through increased moralization) is likely to be a particularly promising
approach. Whereas there is a large body of evidence showing substantial stability and
consistency in ideological beliefs (e.g., Jost, 2006; Jost et al., 2008), along with significant
heritability (Hatemi et al., 2014) and distinct neurocognitive correlates (Amodio et al., 2007),
we are aware of no such evidence establishing the stability, consistency, and heritability of free
will beliefs. Indeed, our own results here show that while conservatives are consistently likely
to report a higher abstract belief in free will, their specific attributions of free will are deeply
context-dependent and susceptible to motivated cognition.
Second, there is room for debate about how precisely free will attributions should be
conceptualised, and especially whether particpants’ ratings of responsibility should be included
in our measure of overall free will attributions. In line with previous empirical work in this
area, we have used the term “free will” to refer to an autonomous choice of action that a
person performs in the absence of substantial internal and external constraints (Baumeister
& Monroe, 2014; Paulhus & Carey, 2011), where this ability to choose renders one morally
responsible for their actions (Nichols, 2007; Nichols & Knobe, 2007). In short, scholars and
laypeople alike appear concerned with free will primarily because they are concerned with
responsibility. And we are interested in motivated attributions of free will precisely because
attributions of free will create a sense of responsibility, which is relevant to many political
disagreements along both economic and social dimensions. It is for these reasons that our
pre-registered composite measure of free will attributions included a question about how
responsible participants judged the actor to be, in addition to questions about the actor’s
control, free will, and ability to choose otherwise. For scholars who would restrict conceptions
of free will to those related to choice, control, and freedom (and not responsibility), we cross-
checked our main results without the responsibility item and this did not impact the statistical
significance of any of our main results. Thus the pattern of results here applies to the lay
conception of free will regardless of whether that includes responsibility or not.
Third, our analysis here examines how manipulating desires to blame increases or
decreases ideological differences in tendencies to attribute free will to individual actors, yet
they are only suggestive of potential influences on persistent individual differences in free will
beliefs. We provided evidence that (1) conservatism is associated with higher beliefs in free
will, (2) conservatism is associated with moralizing, and (3) increasing desires to blame
increases attributions of free will to individual actors. It is therefore theoretically and
empirically plausible that conservatives, experiencing stronger blame desires in their everyday
life, might come to believe more in free will for this reason, but the present work cannot
confirm this definitively. We hope the present results might inspire future work to seek ways
of exploring the causes of persistent individual differences among liberals and conservatives in
their free will beliefs, and particularly to explore the impact of persistent desires to blame.
Fourth, our research here involved attributions of free will for hypothetical (albeit
realistic) and not actual events. This was a deliberate decision to allow us to cleanly manipulate
perceived moral wrongness and control for the information that participants received. Though
we judged this as preferable to studying current real-world situations in which participants
have different levels of knowledge and investment, it could be interesting for future work to
Everett, Clark, et al. (2020, JPSP:PPID)
39
consider the relationship between political ideology, moralization, and attributions of free will
in real-world contexts.
Finally, it should be noted that our experimental research (Studies 3-7) has
focused on data from U.S. participants via MTurk. A key strength of MTurk is that it yields data
that are more representative than those from traditional student samples especially on the
dimensions of age and political ideology (Buhrmester et al., 2011). However, the fact remains
that our samples typically remain more Western, Educated, Industrialized, Rich, and
Democratic (Henrich et al., 2010) than most of the world. For this reason, it would be
interesting for future cross-cultural work to consider the relationship between political
ideology, moralization, and free will in different cultures.
Implications and Directions for Future Research
The findings reported in this paper have theoretical implications for both the
psychology of free will belief and political psychology. First, these findings provide further and
more direct support for previous work conducted on free will belief as motivated social
cognition (Clark, Baumeister, et al., 2017; Clark et al., 2014; Clark, Shniderman, et al., 2018;
Clark, Winegard, & Baumeister, 2019; Vonasch et al., 2017). The work reported here
demonstrates that belief in free will is linked to a desire to hold people accountable for their
moral wrongdoing, and that free will attributions vary as a function of the valence of the action,
how moral or immoral it is perceived to be, and even who the target is. Pragmatically, this
highlights the fact that when exploring free will attributions, it is essential to consider and
control for the valence and perceived morality of the event. Second, these findings provide
further support for the idea that attributions of free will are malleable and context-dependent,
not stable across situations and context (Bargh & Earp, 2009). Third, these results help to shed
light on the pervasive political discourse concerning responsibility and assignment of blame.
The findings suggest that the emphasis within conservative political ideology on personal
responsibility is directly linked to perceptions of immorality. For example, as illegal drug use is
perceived to be more morally wrong, so too are drug users seen as more responsible and in
control of their situation; and as being unemployed and receiving social welfare is seen to be
more morally wrong, so too are people in such situations seen as being more responsible for
and in control of their lot. The reason that political debates concerning responsibility and
deservingness have been an enduring feature of political discourse throughout history is likely
to stem, at least in part, from the powerful and often conflicting moral intuitions driving such
judgments.
Following from this view, our findings have interesting theoretical and practical
implications concerning the structure of policies related to behaviors that conservatives and
liberals perceive as being differentially blameworthy. Consider, for example, attitudes towards
unemployment benefits. The results presented here suggest that conservative policies
opposing greater benefit payments are likely to be linked to the perception that hard work is
a moral principle and that not working is therefore blameworthy. Indeed, such an analysis is
consistent with statements from prominent conservative leaders such as Margaret Thatcher
(1981), who described her policies as “based not on some economics theory, but on things I
and millions like me were brought up with [such as] an honest day’s work for an honest day’s
pay.” Political policies, it seems, are intimately tied to perceptions of morality of the actions
Everett, Clark, et al. (2020, JPSP:PPID)
40
concerned. This may help account for the seeming intractability of political conflict over
economic policy: it is likely easier to reconcile deliberative, fact-based disagreement about the
specific outcomes of policies, than it is to reconcile affect-based intuitions about moral
responsibility. Put simply, while disagreement about economic outcomes can be resolved with
better data, it is difficult to see an easy way to reconcile disagreements about who is morally
responsible, and for what. We have no illusions that this will be easy, but our findings do suggest
that political consensus on hot-topic issues such as welfare and benefits, when it can be
reached, is likely to occur not through extended discussion of the economic features of the
policies, but rather in achieving common moral ground.
Everett, Clark, et al. (2020, JPSP:PPID)
41
References
Alicke, M. (1992). Culpable causation. Journal of Personality and Social Psychology, 63(3), 368
378. https://doi.org/10.1037/0022-3514.63.3.368
Alicke, M. (2000). Culpable control and the psychology of blame. Psychological Bulletin,
126(4), 556574. https://doi.org/10.1037/0033-2909.126.4.556
Alicke, M., Buckingham, J., Zell, E., & Davis, T. (2008). Culpable control and counterfactual
reasoning in the psychology of blame. Personality and Social Psychology Bulletin, 34(10),
13711381. https://doi.org/10.1177/0146167208321594
Alicke, M., Rose, D., & Bloom, D. (2011). Causation, norm violation, and culpable control.
The Journal of Philosophy, 108(12), 670696. https://doi.org/10.5840/jphil20111081238
Amodio, D. M., Jost, J. T., Master, S. L., & Yee, C. M. (2007). Neurocognitive correlates of
liberalism and conservatism. Nature Neuroscience, 10(10), 12461247.
https://doi.org/10.1038/nn1979
Aquino, K., & Reed II, A. (2002). The self-importance of moral identity. Journal of Personality
and Social Psychology, 83(6), 14231440. https://doi.org/10.1037/0022-3514.83.6.1423
Bargh, J. A., & Earp, B. (2009). The will is caused, not “free.” Dialogue: Newsletter of the
Society for Personality and Social Psychology, 24, 1315.
Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001). Bad is stronger than
good. Review of General Psychology, 5(4), 323370. https://doi.org/10.1037/1089-
2680.5.4.323
Baumeister, R. F., & Monroe, A. E. (2014). Recent research on free will: Conceptualizations,
beliefs, and processes. Advances in Experimental Social Psychology, 50, 152.
Baumeister, R. F., & Newman, L. S. (1994). Self-regulation of cognitive inference and decision
processes. Personality and Social Psychology Bulletin, 20(1), 319.
https://doi.org/10.1177/0146167294201001
Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon’s Mechanical Turk: A new
source of inexpensive, yet high-quality, data? Perspectives on Psychological Science, 6(1),
35. https://doi.org/10.1177/1745691610393980
Cameron, D. (2010, May 11). David Cameron’s speech in full. The Guardian.
https://www.theguardian.com/politics/2010/may/11/david-cameron-speech-full-text
Carey, J. M., & Paulhus, D. L. (2013). Worldview implications of believing in free will and/or
determinism: Politics, morality, and punitiveness. Journal of Personality, 81(2), 130
141. https://doi.org/10.1111/j.1467-6494.2012.00799.x
Carroll, J. S., Perkowitz, W. T., Lurigio, A. J., & Weaver, F. M. (1987). Sentencing goals,
causal attributions, ideology, and personality. Journal of Personality, 52(1), 107118.
Cawley, M. J., Martin, J. E., & Johnson, J. A. (2000). A virtues approach to personality.
Personality and Individual Differences, 28(5), 9971013. https://doi.org/10.1016/S0191-
8869(99)00207-X
Clark, C. J., Baumeister, R. F., & Ditto, P. H. (2017). Making punishment palatable: Belief in
free will alleviates punitive distress. Consciousness and Cognition, 51, 193211.
https://doi.org/10.1016/j.concog.2017.03.010
Everett, Clark, et al. (2020, JPSP:PPID)
42
Clark, C. J., Chen, E. E., & Ditto, P. H. (2015). Moral coherence processes: Constructing
culpability and consequences. Current Opinion in Psychology, 6, 123128.
https://doi.org/10.1016/j.copsyc.2015.07.016
Clark, C. J., Luguri, J. B., Ditto, P. H., Knobe, J., Shariff, A. F., & Baumeister, R. F. (2014). Free
to punish: A motivated account of free will belief. Journal of Personality and Social
Psychology, 106(4), 501. https://doi.org/10.1037/a0035880
Clark, C. J., Shniderman, A., Baumeister, R. F., Luguri, J. B., & Ditto, P. H. (2018). Are
morally good actions ever free? Consciousness and Cognition, 63, 161182.
Clark, C. J., Winegard, B. M., & Baumeister, R. F. (2019). Forget the Folk: Moral
Responsibility Preservation Motives and Other Conditions for Compatibilism.
Frontiers in Psychology, 10. https://doi.org/10.3389/fpsyg.2019.00215
Clark, C. J., Winegard, B. M., & Shariff, A. F. (2019). Motivated free will beliefs: The theory,
new (preregistered) studies, and three meta-analyses. Manuscript Submitted for
Publication.
Costa, P. T., & McCrae, R. R. (1992a). Four ways five factors are basic. Personality and
Individual Differences, 13(6), 653665. https://doi.org/10.1016/0191-8869(92)90236-I
Costa, P. T., & McCrae, R. R. (1992b). The Five-Factor Model of Personality and Its
Relevance to Personality Disorders. Journal of Personality Disorders, 6(4), 343359.
https://doi.org/10.1521/pedi.1992.6.4.343
Crandall, C. S. (1994). Prejudice against fat people: Ideology and self-interest. Journal of
Personality and Social Psychology, 66(5), 882894. https://doi.org/10.1037/0022-
3514.66.5.882
Cushman, F., Knobe, J., & Sinnott-Armstrong, W. (2008). Moral appraisals affect
doing/allowing judgments. Cognition, 108(1), 281289.
https://doi.org/10.1016/j.cognition.2008.02.005
Ditto, P. H., & Boardman, A. F. (1995). Perceived accuracy of favorable and unfavorable
psychological feedback. Basic and Applied Social Psychology, 16(12), 137157.
https://doi.org/10.1207/s15324834basp1601&2_9
Ditto, P. H., & Lopez, D. F. (1992). Motivated skepticism: Use of differential decision criteria
for preferred and nonpreferred conclusions. Journal of Personality and Social
Psychology, 63(4), 568584. https://doi.org/10.1037/0022-3514.63.4.568
Ditto, P. H., Pizarro, D. A., & Tannenbaum, D. (2009). Motivated moral reasoning. Psychology
of Learning and Motivation, 50, 307338. https://doi.org/10.1016/S0079-
7421(08)00410-6
Dunning, D., Leuenberger, A., & Sherman, D. A. (1995). A new look at motivated inference:
Are self-serving theories of success a product of motivational forces? Journal of
Personality and Social Psychology, 69(1), 5868. https://doi.org/10.1037/0022-
3514.69.1.58
Eidelman, S., Crandall, C. S., Goodman, J. A., & Blanchar, J. C. (2012). Low-effort thought
promotes political conservatism. Personality and Social Psychology Bulletin, 38(6), 808
820. https://doi.org/10.1177/0146167212439213
Everett, Clark, et al. (2020, JPSP:PPID)
43
Feather, N. T. (1985). Attitudes, values, and attributions: Explanations of unemployment.
Journal of Personality and Social Psychology, 48(4), 876889.
https://doi.org/10.1037/0022-3514.48.4.876
Feldman, G., Wong, K. F. E., & Baumeister, R. F. (2016). Bad is freer than good: Positive
negative asymmetry in attributions of free will. Consciousness and Cognition, 42, 2640.
https://doi.org/10.1016/j.concog.2016.03.005
Frimer, J. A., Tell, C. E., & Motyl, M. (2017). Sacralizing liberals and fair-minded
conservatives: Ideological symmetry in the moral motives in the culture war. Analyses
of Social Issues and Public Policy, 17(1), 3359. https://doi.org/10.1111/asap.12127
Goh, J. X., Hall, J. A., & Rosenthal, R. (2016). Mini meta-analysis of your own studies: Some
arguments on why and a primer on how. Social and Personality Psychology Compass,
10(10), 535549. https://doi.org/10.1111/spc3.12267
Graham, J., & Haidt, J. (2012). Sacred values and evil adversaries: A moral foundations
approach. In The social psychology of morality: Exploring the causes of good and evil (pp.
1131). American Psychological Association. https://doi.org/10.1037/13091-001
Graham, J., Haidt, J., Koleva, S., Motyl, M., Iyer, R., Wojcik, S. P., & Ditto, P. H. (2013). Moral
foundations theory: The pragmatic validity of moral pluralism. Advances in
Experimental Social Psychology, 47, 55130.
Graham, J., Haidt, J., & Nosek, B. A. (2009). Liberals and conservatives rely on different sets
of moral foundations. Journal of Personality and Social Psychology, 96(5), 1029.
https://doi.org/10.1037/a0015141
Graham, J., Nosek, B. A., Haidt, J., Iyer, R., Koleva, S., & Ditto, P. H. (2011). Mapping the
moral domain. Journal of Personality and Social Psychology, 101(2), 366385.
https://doi.org/10.1037/a0021847
Haidt, J. (2001). The emotional dog and its rational tail: A social intuitionist approach to
moral judgment. Psychological Review, 108(4), 814834. https://doi.org/10.1037/0033-
295X.108.4.814
Haidt, J. (2012). The righteous mind: Why good people are divided by politics and religion. Vintage.
Hamlin, J. K., & Baron, A. S. (2014). Agency attribution in infancy: Evidence for a negativity
bias. PLOS ONE, 9(5), e96112. https://doi.org/10.1371/journal.pone.0096112
Harris, L. T., & Fiske, S. T. (2006). Dehumanizing the lowest of the low: Neuroimaging
responses to extreme out-groups. Psychological Science, 17(10), 847853.
https://doi.org/10.1111/j.1467-9280.2006.01793.x
Hatemi, P. K., Crabtree, C., & Smith, K. B. (2019). Ideology justifies morality: Political beliefs
predict moral foundations. American Journal of Political Science, 63(4), 788806.
https://doi.org/10.1111/ajps.12448
Hatemi, P. K., Medland, S. E., Klemmensen, R., Oskarrson, S., Littvay, L., Dawes, C.,
Verhulst, B., McDermott, R., Nørgaard, A. S., Klofstad, C., Christensen, K.,
Johannesson, M., Magnusson, P. K. E., Eaves, L. J., & Martin, N. G. (2014). Genetic
influences on political ideologies: Twin analyses of 19 measures of political ideologies
from five democracies and genome-wide findings from three populations. Behavior
Genetics, 44(3), 282294. https://doi.org/10.1007/s10519-014-9648-8
Everett, Clark, et al. (2020, JPSP:PPID)
44
Hedges, L., & Vevea, J. (1998). Fixed- and random-effects models in meta-analysis.
Psychological Methods, 3(4), 486504.
Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world?
Behavioral and Brain Sciences, 33(23), 6183.
https://doi.org/10.1017/S0140525X0999152X
Jost, J. T. (2006). The end of the end of ideology. American Psychologist, 61(7), 651670.
https://doi.org/10.1037/0003-066X.61.7.651
Jost, J. T., Federico, C. M., & Napier, J. L. (2009). Political ideology: Its structure, functions,
and elective affinities. Annual Review of Psychology, 60(1), 307337.
https://doi.org/10.1146/annurev.psych.60.110707.163600
Jost, J. T., Nosek, B. A., & Gosling, S. D. (2008). Ideology: Its resurgence in social,
personality, and political psychology. Perspectives on Psychological Science, 3(2), 126
136. https://doi.org/10.1111/j.1745-6916.2008.00070.x
Knobe, J. (2003). Intentional action and side effects in ordinary language. Analysis, 63(279),
190194. https://doi.org/10.1111/1467-8284.00419
Knobe, J., & Fraser, B. (2008). Causal judgment and moral judgment: Two experiments. In
W. Sinnott-Armstrong (Ed.), Moral Psychology. MIT Press.
Kugler, M., Jost, J. T., & Noorbaloochi, S. (2014). Another look at moral foundations theory:
Do authoritarianism and social dominance orientation explain liberal-conservative
differences in “moral” intuitions? Social Justice Research, 27(4), 413431.
https://doi.org/10.1007/s11211-014-0223-5
Lakens, D. (2014). Performing high-powered studies efficiently with sequential analyses.
European Journal of Social Psychology, 44(7), 701710.
Lapsley, D. K., & Lasky, B. (2001). Prototypic Moral Character. Identity, 1(4), 345363.
https://doi.org/10.1207/S1532706XID0104_03
Leslie, A. M., Knobe, J., & Cohen, A. (2006). Acting intentionally and the side-effect effect:
Theory of mind and moral judgment. Psychological Science, 17(5), 421427.
https://doi.org/10.1111/j.1467-9280.2006.01722.x
Lovett, B. J., Jordan, A. H., & Wiltermuth, S. S. (2012). Individual differences in the
moralization of everyday life. Ethics & Behavior, 22(4), 248257.
https://doi.org/10.1080/10508422.2012.659132
McAdams, D. P., Albaugh, M., Farber, E., Daniels, J., Logan, R. L., & Olson, B. (2008). Family
metaphors and moral intuitions: How conservatives and liberals narrate their lives.
Journal of Personality and Social Psychology, 95(4), 978990.
https://doi.org/10.1037/a0012650
Monroe, A. E., & Ysidron, D. (2019). Not so motivated after all? Three replication attempts
and a theoretical challenge to a morally-motivated belief in free will. Manuscript
Submitted for Publication.
Mooney, C. (2012). The Republican brain: The science of why they deny scienceand reality. John
Wiley & Sons.
Newman, L. S., Tan, M., Caldwell, T. L., Duff, K. J., & Winer, E. S. (2018). Name norms: A
guide to casting your next experiment. Personality and Social Psychology Bulletin,
0146167218769858. https://doi.org/10.1177/0146167218769858
Everett, Clark, et al. (2020, JPSP:PPID)
45
Nichols, S. (2007). The rise of compatibilism: A case study in the quantitative history of
philosophy. Midwest Studies In Philosophy, 31(1), 260270.
https://doi.org/10.1111/j.1475-4975.2007.00152.x
Nichols, S., & Knobe, J. (2007). Moral responsibility and determinism: The cognitive science
of folk intuitions. Noûs, 41(4), 663685. https://doi.org/10.1111/j.1468-
0068.2007.00666.x
Padilla-Walker, L. M., & Jensen, L. A. (2016). Validation of the long- and short-form of the
Ethical Values Assessment (EVA): A questionnaire measuring the three ethics
approach to moral psychology. International Journal of Behavioral Development, 40(2),
181192. https://doi.org/10.1177/0165025415587534
Paolacci, G., & Chandler, J. (2014). Inside the Turk: Understanding Mechanical Turk as a
Participant Pool. Current Directions in Psychological Science, 23(3), 184188.
https://doi.org/10.1177/0963721414531598
Paulhus, D. L., & Carey, J. M. (2011). The FADPlus: Measuring Lay Beliefs Regarding Free
Will and Related Constructs. Journal of Personality Assessment, 93(1), 96104.
https://doi.org/10.1080/00223891.2010.528483
Paulhus, D. L., & Margesson, A. (1994). Free Will and Scientific Determinism (FAD-4) scale.
Unpublished instrument, University of British Columbia, Vancouver, BC, Canada.
Pettit, D., & Knobe, J. (2009). The pervasive impact of moral judgment. Mind & Language,
24(5), 586604. https://doi.org/10.1111/j.1468-0017.2009.01375.x
Phillips, J., & Knobe, J. (2009). Moral judgments and intuitions about freedom. Psychological
Inquiry, 20(1), 3036. https://doi.org/10.1080/10478400902744279
Reagan, R. (1968, July 31). Excerpts of a Speech by Governor Ronald Reagan. Republican
National Convention, Miama, FL.
https://en.wikiquote.org/wiki/Ronald_Reagan#1960s
Sagarin, B. J., Ambler, J. K., & Lee, E. M. (2014). An Ethical Approach to Peeking at Data:
Perspectives on Psychological Science. https://doi.org/10.1177/1745691614528214
Sargent, M. J. (2004). Less Thought, More Punishment: Need for Cognition Predicts Support
for Punitive Responses to Crime. Personality and Social Psychology Bulletin, 30(11),
14851493. https://doi.org/10.1177/0146167204264481
Skitka, L. J., Mullen, E., Griffin, T., Hutchinson, S., & Chamberlin, B. (2002). Dispositions,
scripts, or motivated correction?: Understanding ideological differences in
explanations for social problems. Journal of Personality and Social Psychology, 83(2),
470487. https://doi.org/10.1037/0022-3514.83.2.470
Skitka, L. J., & Tetlock, P. E. (1992). Allocating scarce resources: A contingency model of
distributive justice. Journal of Experimental Social Psychology, 28(6), 491522.
https://doi.org/10.1016/0022-1031(92)90043-J
Skitka, L. J., & Tetlock, P. E. (1993). Providing public assistance: Cognitive and motivational
processes underlying liberal and conservative policy preferences. Journal of Personality
and Social Psychology, 65(6), 1205. https://doi.org/10.1037/0022-3514.65.6.1205
Smith, K. B., Alford, J. R., Hibbing, J. R., Martin, N. G., & Hatemi, P. K. (2017). Intuitive ethics
and political orientations: Testing moral foundations as a theory of political ideology.
American Journal of Political Science, 61(2), 424437. https://doi.org/10.1111/ajps.12255
Everett, Clark, et al. (2020, JPSP:PPID)
46
Smith, K. D., Smith, S. T., & Christopher, J. C. (2007). What defines the good person?
Cross-cultural comparisons of experts’ models with lay prototypes. Journal of Cross-
Cultural Psychology, 38(3), 333360. https://doi.org/10.1177/0022022107300279
Thatcher, M. (1981, September 20). My policies are based not on some economics theory...
The News of The World. https://en.wikiquote.org/wiki/Margaret_Thatcher
Vonasch, A. J., Clark, C. J., Lau, S., Vohs, K. D., & Baumeister, R. F. (2017). Ordinary people
associate addiction with loss of free will. Addictive Behaviors Reports, 5, 5666.
https://doi.org/10.1016/j.abrep.2017.01.002
Walker, L. J., & Pitts, R. C. (1998). Naturalistic conceptions of moral maturity. Developmental
Psychology, 34(3), 403419.
Zucker, G. S., & Weiner, B. (1993). Conservatism and perceptions of poverty: An
attributional analysis. Journal of Applied Social Psychology, 23(12), 925943.
https://doi.org/10.1111/j.1559-1816.1993.tb01014.x
... On the other hand, Carey and Paulhus (2013) found that greater belief in free will was positively correlated with the endorsement of retributive punishment, belief in a just world, conservatism, religiousness, and moralistic judgment. Analysis of survey data (Martin et al., 2017) and a recent meta-review (Everett et al., 2021) corroborated these findings. Free will belief was also found to be related to blaming individuals for their mental illness or obesity (Chandrashekar, 2020) and to lower tolerance of sexual minorities (Brewer, 2013). ...
... Stigma towards substance users could be biased due to differences in the familiarity of men and women with substance use. Moreover, the literature suggests an association between beliefs in free will and individuals' cultural backgrounds, political orientations, and religious beliefs (e.g., Caspar et al., 2017;Everett et al., 2021;Hook & Markus, 2020). These could play a role in the general mechanism of the relation of belief in free will and stigma for various stigmatized identities and conditions. ...
Article
Full-text available
Belief in free will reflects the view that behaviors can be controlled via willpower. If a condition is thought to result from a lack of willpower, individuals with that condition may be subject to higher levels of stigma due to greater attribution of blame. In a study conducted online, we manipulated belief in free will experimentally by giving 200 Turkish participants a list of sentences about free will and asking them to explain why these statements are either true or false (depending on their randomly assigned condition). The manipulation did change significantly belief in free will and substance use stigma, but not suicide stigma. Furthermore, belief in free will mediated the effect of the manipulation on stigma related to both substance use and suicide. Efforts to reduce stigma related to substance use and suicide may include strategies to change beliefs about the role of free will in these conditions.
... Conversely, an individual could conclude that a consumer of ethical goods committing ethical transgression is indeed a worse person due to the inconsistency between one's projected image and actual behaviors. Such a reaction would be aligned with the need of more conservatives individuals to maintain a sense of certainty about the world, which is associated with a greater tendency to assign personal blame for bad behaviors (Everett et al., 2021) or punish those who deviate from group stereotypes (Stern, West, & Rule, 2015). While there is nothing necessarily "conservative" about conventional products and thus no positive disposition to such goods, the cognitive rigidity associated with conservatism would be consistent with more negative reactions and more difficulty in categorizing others in the face of discrepant information. ...
... CI 95 [0.05, 0.21]) and a main effect of political orientation (β = 0.28, b = 0.24, SE = 0.04, t(392) = 5.92, p < .001, CI 95 [0.16, 0.32]) such that conservatives were more likely to assign personal blame (as also found in Everett et al., 2021). That is, participants who rated at −2.87 or below on the political orientation scale (29th percentile; more liberal) were relatively more likely to attribute the event to external factors when the person drove a Toyota Prius user (ŷ −2.87 = 0.21) compared with a Camry (ŷ −2.87 = 0.91). ...
Article
Full-text available
In today's marketplace, users (e.g., purchasers, influencers) are increasingly the "face" of brands to potential consumers, increasing the risk for brands should these users act poorly. Across seven studies, we document that political orientation moderates the desire for punishment toward users of ethical (vs. conventional) brands who commit moral transgressions. In response to identical marketplace transgressions, we observe that liberals punish ethical brand users less than conventional brand users. In contrast, conservatives punish the same users of ethical brands more than conventional brand users. We document that this bias stems from how people interpret the inconsistency between the ethical branding and the act of transgression, rather than from a group‐identity effect, showing how it does not arise in the absence of inconsistent information or when consumers are not able to integrate the inconsistent information to their judgments. We also investigate an avenue by which firms can reframe their ethical branding to reduce this politically motivated bias. We discuss this work's implications for moral judgments, marketplace attribute formation, and the branding of ethical goods in a politically divided world.
... Similarly, people attributed more intentionality and causality to actors whose behavior produced harmful consequences than actors whose identical behaviors produced helpful consequences (e.g., Knobe, 2003Knobe, , 2006Knobe & Fraser, 2008;Leslie et al., 2006). In general, people seem to attribute more responsibility (and related attributes such as control and freedom) to harmful actions and actions with harmful outcomes than to closely matched helpful actions and actions with helpful outcomes (e.g., Clark et al., 2016;Reeder & Spores, 1983), neutral actions and actions with neutral outcomes (e.g., Cushman et al., 2008), and less harmful actions and actions with less harmful outcomes (e.g., Everett et al., 2020;Walster, 1966). ...
... Like motivated attributions of free will (e.g., Clark et al., , 2018Everett et al., 2020), the Blame Efficiency Hypothesis may explain motivated free will beliefs: people tend to report stronger beliefs in free will and resist science that challenges free will after exposure to others' harmful actions (e.g., Clark et al., 2014Clark et al., , 2021. The existence of free will may be the ultimate ambiguous moral judgment. ...
Chapter
Full-text available
The Blame Efficiency Hypothesis applies insights from evolutionary psychology to resolve the apparent conflict between rationalist and intuitionist perspectives on moral judgment. First, people reserve moral condemnation for actors and actions that are likely to be deterred by moral condemnation. This includes intended and controllable actions (consistent with rationalist perspectives that highlight how considerations of intentionality and causal control influence blame judgments) because blame cannot deter intentions that never existed in the first place, nor can it alter unalterable actions. And this includes actors who care about their moral reputations (i.e., cognitively competent and mentally healthy humans), because moral condemnation cannot deter those with no regard for their moral reputations. Second, moral judgment is subject to error management principles. Failing to condemn a morally responsible harmdoer signals exploitability, and so it would be more costly to erroneously not blame a harmdoer who could have been deterred by blame than to erroneously blame a harmdoer who could not have been deterred by blame (called The Don Corleone Principle). In contrast, no obvious cost asymmetry exists for overestimations or underestimations of moral responsibility for helpful or neutral actions. Therefore, in ambiguous cases, people should err on the side of assuming harmdoers are morally responsible (consistent with intuitionist perspectives which highlight how punitive desires influence judgments of intentionality and moral responsibility). Blaming efficiently means that people attribute moral responsibility to those who appear capable of controlling their behavior in future situations and to those who appear controllable—likely to change their behavior in response to harsh moral judgment, with a small bias toward ascribing responsibility in ambiguous cases of harm to minimize the costly error of under-blaming.
... As for the relationship between free will beliefs and explicitly moral beliefs and behavior, researchers have reported that free will beliefs are associated with decreased aggression and increased helpfulness towards strangers (Baumeister, Masicampo, & DeWall, 2009), causal cognition (Genschow & Brass, 2017;Genschow, Rigoni, & Brass, 2019), moral attitudes (Paulhus & Carey, 2011), subjective well-being (Li, Wang, Zhao, Kong, & Li, 2016;Moynihan et al., 2017;Feldman & Chandrashekar, 2018), moralization and political ideology (Everett et al. 2021); attitudes toward authoritarianism (Carey & Paulhus, 2013;Costello, Bowes, & Lilienfeld, 2020;Nadelhoffer & Goya-Tocchetto, 2013), heightened moral judgments and attributions of blame and punishment (Carey & Paulhus, 2013;Clark, Luguri, Ditto, Knobe, Shariff, & Baumeister, 2014;Clark, Baumeister, & Ditto, 2017;Krueger, Hoffman, Walter, & Grafman, 2013;Martin, Rigoni, & Vohs, 2017;Shariff et al., 2014), and views about inequality (Mercier, Wiwad, Piff, Aknin, Robinson, & Shariff, 2020). 3 In short, the gathering data highlight that what people think about free will and moral responsibility is both theoretically interesting and practically important. ...
... In addition, the additional assessment of misleadingness allowed us to examine whether any possible differences are specifically bound to judgments of lying or whether they apply to judgments of deceptiveness more generally. The morality question, on the other hand, was included so that participants would be able to express their moral evaluation of each of the deceptions, in order to prevent the lying and misleading questions from being inflated by a desire to blame (an effect which has been reported in other judgments related to the moral domain; e.g., Everett et al. 2021). While the design originally resulted in a total of 16 scenarios, 2 scenarios (the PCI in the content domain marriage and the presupposition in the content domain police) were removed from all analyses reported below. ...
Article
Full-text available
The present study examines cross-cultural differences in people’s concept of lying with regard to the question of whether lying requires an agent to say something they believe to be false. While prominent philosophical views maintain that lying entails that a person explicitly expresses a believed-false claim, recent research suggests that people’s concept of lying might also include certain kinds of deception that are communicated more indirectly. An important drawback of previous empirical work on this topic is that only few studies have investigated people’s concept of lying in non-Western samples. In the present study, we compare people’s intuitions about lying with indirect deceptions (i.e., presuppositions, conversational implicatures, and non-verbal actions) in a sample of N = 255 participants from Russia and N = 300 participants from the United Kingdom. Our findings show a strong degree of similarity between lie ratings of participants from Russia and the United Kingdom, with both samples holding it possible for agents to lie with deceptive statements and actions that do not involve the agent saying something they believe to be false.
... In particular, we examine how beliefs about changes in financial well-being are associated with 221 rated importance of different goals that a government may pursue (Study 2), the relative 222 characterized in part by a need to "avoid change, disruption, and ambiguity … and to explain, 247 order, and justify inequality among groups and individuals." Conservatives and liberals also 248 differ in their lay beliefs about free will; conservatives tend to believe that people have more 249 autonomous control over their behavior (Carey & Paulhus, 2013;Everett et al., 2020). These 250 differences may be a reason why conservatives tend to favor internal causal attributions for 251 outcomes in life. ...
Article
Full-text available
People differ in their lay theories about how and why the financial well-being of individuals changes over time or varies between individuals. We introduce a measure of Causal Attributions of Financial Uncertainty—the CAFU scale—and find that such attributions can be reliably described along three distinct dimensions, respectively capturing the extent to which changes in financial well-being are perceived to be: (1) knowable and within individuals’ control due to individual factors such as effort (“Rewarding”); (2) knowable and outside of individuals’ control due to systemic factors such as favoritism and discrimination (“Rigged”); and (3) inherently unpredictable and determined by chance events (“Random”). In a sample representative of the U.S. population on various demographic characteristics (N = 1,102), we find that differences in these beliefs are associated with political ideology, revealing a predicted pattern: conservatives scored higher on the Rewarding subscale and liberals scored higher on the Rigged and Random subscales, even when controlling for key demographics. Moreover, we find that these three dimensions predict responses to different policy messages when controlling for political ideology. In three preregistered experiments (combined N = 2,560), we observe increased support for various social welfare policies when we highlighted aspects of these policies that are compatible with people’s beliefs about financial well-being. Likewise, we observe increased support for political candidates when they expressed their positions in a way that is compatible with people’s beliefs. Thus, this work can help better understand drivers of political attitudes and guide in crafting more persuasive policy messages.
Article
Full-text available
The primary aim of this essay is to consider whether free will skeptics’ assertions of moral claims pertaining to human agents and their actions are consistent with the rejection of the kind of moral responsibility—namely, that which requires free will—that is espoused by virtually all of them. In order to address this issue, I will examine one of the most well-known and detailed defenses of morality by a free will skeptic; namely, that provided by Derk Pereboom in his book Living without Free Will. I conclude that free will skeptics should dispense with making any moral assertions with regard to human agents or their actions since doing so does little more than muddy up the conceptual waters by recasting moral terms in ways that are at odds with their traditional or folk usage.
Article
Within the criminal justice system, one of the most prominent justifications for legal punishment is retributivism. The retributive justification of legal punishment maintains that wrongdoers are morally responsible for their actions and deserve to be punished in proportion to their wrongdoing. This book argues against retributivism and develops a viable alternative that is both ethically defensible and practical. Introducing six distinct reasons for rejecting retributivism, Gregg D. Caruso contends that it is unclear that agents possess the kind of free will and moral responsibility needed to justify this view of punishment. While a number of alternatives to retributivism exist - including consequentialist deterrence, educational, and communicative theories - they have ethical problems of their own. Moving beyond existing theories, Caruso presents a new non-retributive approach called the public health-quarantine model. In stark contrast to retributivism, the public health-quarantine model provides a more human, holistic, and effective approach to dealing with criminal behavior.
Article
Full-text available
Published on Psychology Today: https://www.psychologytoday.com/us/blog/the-antisocial-psychologist/202106/pro-blame-bias-the-don-corleone-principle
Chapter
Full-text available
The social and behavioral sciences have taken a substantial reputational hit over the past decade. Some highly publicized findings have failed to replicate—and those that do replicate often do so with much smaller effect sizes (Camerer et al., 2018; Nosek et al., 2021). Plus some highly touted “science-based” interventions have failed to produce promised positive social change—even when massive efforts are dedicated to making them work (Singal, 2021). This chapter will lay out our two-tiered hypothesis: (a) the ideological homogeneity of the social sciences has entrenched certain scientific orthodoxies and taboos; (b) these orthodoxies and taboos have protected weak ideas from rigorous scrutiny and contributed to the replication crisis. We also explain how open science practices, although a big step in the right direction, leave many researcher degrees of freedom on the table that can bias methodological decisions and research conclusions. We argue that adversarial collaborations are the next necessary science reform for addressing lingering weaknesses in social scientific norms and can further minimize false positives, expedite scientific corrections, stimulate progress for stalemated scientific debates, and ultimately improve the quality of social scientific outputs.
Article
Full-text available
Clark and colleagues (2014) proposed a theory of motivated free will beliefs, according to which at least part of free will beliefs and attributions are caused by a desire to hold moral transgressors responsible. Recently, this theory has been challenged. In the following article, we examine the evidence and conclude that, although not dispositive, much of the evidence seems to support the motivated account. For example, in 14 new (7 preregistered) studies (n=4,014), results consistently supported the motivated theory; and these findings consistently replicated in studies (k=8) that tested an alternative (counternormative) hypothesis. In addition, three meta-analyses of the existing data (including eight vignette types and eight free will judgment types) found support for motivated free will attributions (k=22; n=7,619; r=.25, p<.001) and beliefs (k=27; n=8,100; r=.13, p<.001), which remained robust after removing all potential confounds (k=26; n=7,953; r=.12, p<.001). However, the size of these effects varied by vignette type and free will belief measurement. We discuss these variations and the implications for different theories of free will beliefs and attributions. And we end by discussing the relevance of these findings for past and future research and the significance of these findings for human responsibility.
Article
Full-text available
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.
Article
Full-text available
Psychologists often test hypotheses by constructing vignettes depicting people engaging in behavior and displaying characteristics designed to operationalize specific variables. People described in these vignettes are typically given names, but names have a variety of connotations that could lead to unwanted variance between conditions of an experiment and in other ways have implications for the results of a study. An up-to-date source of information to help guide the selection of names would be useful for researchers. Participants from four different regions of the United States rated a large sample of names in terms of perceived age, warmth, and competence. The full set of names is presented, along with some simple demonstration studies concretely illustrating the implications of name choice.
Article
Full-text available
For years, experimental philosophers have attempted to discern whether laypeople find free will compatible with a scientifically deterministic understanding of the universe, yet no consensus has emerged. The present work provides one potential explanation for these discrepant findings: People are strongly motivated to preserve free will and moral responsibility, and thus do not have stable, logically rigorous notions of free will. Seven studies support this hypothesis by demonstrating that a variety of logically irrelevant (but motivationally relevant) features influence compatibilist judgments. In Study 1, participants who were asked to consider the possibility that our universe is deterministic were more compatibilist than those not asked to consider this possibility, suggesting that determinism poses a threat to moral responsibility, which increases compatibilist responding (thus reducing the threat). In Study 2, participants who considered concrete instances of moral behavior found compatibilist free will more sufficient for moral responsibility than participants who were asked about moral responsibility more generally. In Study 3a, the order in which participants read free will and determinism descriptions influenced their compatibilist judgments—and only when the descriptions had moral significance: Participants were more likely to report that determinism was compatible with free will than that free will was compatible with determinism. In Study 3b, participants who read the free will description first (the more compatibilist group) were particularly likely to confess that their beliefs in free will and moral responsibility and their disbelief in determinism influenced their conclusion. In Study 4, participants reduced their compatibilist beliefs after reading a passage that argued that moral responsibility could be preserved even in the absence of free will. Participants also reported that immaterial souls were compatible with scientific determinism, most strongly among immaterial soul believers (Study 5), and evaluated information about the capacities of primates in a biased manner favoring the existence of human free will (Study 6). These results suggest that people do not have one intuition about whether free will is compatible with determinism. Instead, people report that free will is compatible with determinism when desiring to uphold moral responsibility. Recommendations for future work are discussed.
Article
Free will is often appraised as a necessary input to for holding others morally or legally responsible for misdeeds. Recently, however, Clark and colleagues (2014) argued for the opposite causal relationship. They assert that moral judgments and the desire to punish motivate people's belief in free will. Three replication experiments (Studies 1-2b) attempt to reproduce these findings. Additionally, a novel experiment (Study 3) tests a theoretical challenge derived from attribution theory, which suggests that immoral behaviors do not uniquely influence free will judgments. Instead, our nonviolation model argues that norm deviations of any kind-good, bad, or strange-cause people to attribute more free will to agents. Across replication experiments we found no consistent evidence for the claim that witnessing immoral behavior causes people to increase their general belief in free will. By contrast, we replicated the finding that people attribute more free will to agents who behave immorally compared to a neutral control (Studies 2a and 3). Finally, our novel experiment demonstrated broad support for our norm-violation account, suggesting that people's willingness to attribute free will to others is malleable, but not because people are motivated to blame. Instead, this experiment shows that attributions of free will are best explained by people's expectations for norm adherence, and when these expectations are violated, people infer that an agent expressed their free will to do so. (PsycInfo Database Record (c) 2020 APA, all rights reserved).
Article
Moral Foundations Theory (MFT) is employed as a causal explanation of ideology that posits political attitudes are products of moral intuitions. Prior theoretical models, however, suggest the opposite causal path, that is, that moral judgments are driven by political beliefs. In both instances, however, extant research has assumed rather than explicitly tested for causality. So do moral intuitions drive political beliefs or do political beliefs drive moral intuitions? We empirically address this question using data from two panel studies and one nationally representative study, and find consistent evidence supporting the hypothesis that ideology predicts moral intuitions. The findings have significant implications for MFT as a theory of ideology, and also about the consequences of political beliefs for shaping how individuals rationalize what is right and what is wrong.
Article
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
Research has shown that people ascribe more responsibility to morally bad actions than both morally good and neutral ones, suggesting that people do not attribute responsibility to morally good actions. The present work demonstrates that this is not so: People ascribe more free will to morally good than neutral actions (Studies 1a-1b, Mini Meta). Studies 2a-2b distinguished the underlying motives for ascribing freedom to morally good and bad actions. Free will ascriptions for immoral actions were driven predominantly by affective responses (i.e., punitive desires, moral outrage, and perceived severity of the crime). Free will judgments for morally good actions were similarly driven by affective responses (i.e., reward desires, moral uplift, and perceived generosity), but also more pragmatic considerations (perceived utility of reward, counternormativity of the action, and required willpower). Morally good actions may be more carefully considered, leading to generally weaker, but more contextually sensitive free will judgments.
Article
See updated full text here: https://www.researchgate.net/publication/319944911_Forget_the_Folk_Moral_Responsibility_Preservation_Motives_and_Other_Conditions_for_Compatibilism