ArticlePDF Available

Abstract and Figures

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.
Content may be subject to copyright.
Running Head: ARE MORALLY GOOD ACTIONS EVER FREE?
!
1!
***NOW IN PRESS AT CONSCIOUSNESS AND COGNITION (AS OF MAY 15, 2018)***
Are morally good actions ever free?
Cory J. Clark
Florida State University
Adam Shniderman
University of Michigan
Jamie B. Luguri
University of Chicago
Roy F. Baumeister
Florida State University and University of Queensland
Peter H. Ditto
University of California, Irvine
Please contact Cory J. Clark with any questions concerning this article at: Department of
Psychology, Florida State University. 1107 W. Call Street, Tallahassee, FL. 32306-4301. Tel:
(330) 714-9094, Email: callcoryclark@gmail.com
The work of Cory J. Clark was supported in part by a grant from The Charles Koch Foundation.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
2!
Abstract
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.
Keywords: free will, morality, praise, blame, motivated cognition, affect, punishment,
reward, responsibility
ARE MORALLY GOOD ACTIONS EVER FREE?
!
3!
Are Morally Good Actions Ever Free?
1. Introduction
In October 2015, the Concord, Massachusetts police department began issuing citations
to their residents for behaviors such as crossing the street at marked crosswalks, wearing
seatbelts, and yielding to pedestrians. These citations did not bring fines or jail time but rather
could be redeemed for ice cream at a local café. The peculiarity of these good-behavior citations
gained the community media attention. However, the program was not adopted as a permanent
policy for incentivizing law-abiding behavior within the community. Whereas most governments
have penal systems for punishing rule breakers, very few have systems for rewarding rule
followers or philanthropists.
On a societal level, establishing accountability for harmful actions and outcomes is
highly prioritized above establishing accountability for helpful actions and outcomes. This same
moral valence asymmetry has been found repeatedly in individual judgments of accountability.
People ascribe more responsibility to morally bad actions and outcomes than morally good ones
(e.g., Alicke, 1992; Knobe, 2003; Reeder & Spores, 1983). Expanding on this work, recent
research has demonstrated that “bad is freer than good” (p.26): People attribute more free will to
bad actions and actions with bad outcomes than good actions and actions with good outcomes
(Feldman, Wong, & Baumeister, 2016). These and many similar results seem to suggest that
people are disinclined to assign responsibility for good actions. To our knowledge, however, no
work has compared responsibility judgments for morally good actions to morally neutral ones.
Because rewarding prosocial behavior, like punishing antisocial behavior, is beneficial for group
functioning, we hypothesized that reward and praise motives might also increase ascriptions of
free will. Therefore, we expected that people would ascribe more free will to morally good
ARE MORALLY GOOD ACTIONS EVER FREE?
!
4!
actions than morally neutral ones, but, replicating past work, free will judgments would be
higher for morally bad actions than both morally good and neutral actions.
The present research also sought to test the hypothesis that responses to morally good
actions would be more context-sensitive than responses to bad actions. Outrage, indignation, and
other emotional responses to bad actions produce strong impulses to punish, regardless of
whether the punishment will have any deterrent effect (e.g., Carlsmith, Darley & Robinson,
2002; Crockett, Özdemir, & Fehr, 2014). These impulses motivate individuals to perceive
harmful actions as freely performed, which helps to justify punishing the wrongdoers (e.g., Clark
et al., 2014). In contrast, positive emotional responses to morally positive actions may increase
sensitivity to contextual features of the act, consistent with the generally broadening effect of
positive emotion (e.g., Frederickson, 2001, 2013; Frederickson & Branigan, 2005). When people
witness morally good actions, they may more carefully evaluate whether the behavior is
deserving of praise and whether rewarding the behavior would encourage future good behavior.
We hypothesized that free will judgments for morally bad actions would be driven
predominantly by affective punitive motives, whereas people might more carefully consider a
variety of relevant features when determining whether morally good actions were freely
performed.
1.1 The Good, The Bad, and The Neutral
Psychology has consistently shown that negativity plays a greater role in peoples’ lives
than positivity (Kanouse & Hanson, 1971; Lewicka, Czapinski, & Peeters, 1992; Skowronski &
Carlston, 1989). A sweeping literature review concluded that “bad is stronger than good
(Baumeister, Bratslavsky, Finkenauer, & Vohs, 2001): people give greater weight to negative
events, objects, and personality traits than good ones (Rozin & Royzman, 2001). Regarding
ARE MORALLY GOOD ACTIONS EVER FREE?
!
5!
responsibility-related judgments for morally relevant behavior, myriad studies have
demonstrated that people attribute more responsibility to bad actions and actions with bad
outcomes than to closely matched 1) good actions and actions with good outcomes (e.g., Alicke,
1992; Knobe, 2003; Reeder & Spores, 1983), 2) neutral actions and actions with neutral
outcomes (e.g., Cushman, Knobe, & Sinnott-Armstrong, 2008; Knobe & Fraser, 2008), and 3)
less bad actions and actions with less bad outcomes (e.g., Walster, 1966). Most relevant to the
present work, people attribute more free will to morally bad actions and outcomes than to
morally good (e.g., Everett et al., 2017) and morally neutral ones (Clark et al., 2014).
This body of work seems to suggest that people do not attribute responsibility to morally
good actions, but very little work has compared responsibility judgments for morally positive
actions to morally neutral ones. In one cross-cultural analysis, participants from the U.S. and
India attributed less intentionality to actions with helpful side-effects than neutral side-effects,
seeming to suggest that people do not attribute responsibility to helpful actions (Clark, Bauman,
Kamble, & Knowles, 2016). However, in these same studies, praise motives were strong
predictors of intentionality judgments for actions with helpful side-effects, seeming to suggest
that praise motives do increase attributions of responsibility. To our knowledge, no similar work
has compared responsibility judgments between morally positive and neutral actions where the
moral valence of the action was not a mere side-effect. Thus, it remains unknown whether praise
motives elicit heightened responsibility judgments for morally positive actions relative to
morally neutral ones. The first goal of the present research was to investigate these differences in
the context of free will judgments. We sought to examine whether exposure to morally positive
actions, and subsequent motives to praise, would increase free will judgments. We readily
concede that bad actions and subsequent punitive motives elicit the strongest free will judgments
ARE MORALLY GOOD ACTIONS EVER FREE?
!
6!
— but good actions and praise motives may elicit stronger free will judgments than morally
neutral actions.
1.2 Free Will Judgments
At least since the days of Aristotle, philosophers and other intellectuals have debated the
existence of human free will, a debate revived in recent years by neuroscientists and
psychologists (e.g., Bargh, 2008; Baumeister, Clark, & Luguri, 2015; Libet, 1985; Soon, Brass,
Heinze, & Haynes, 2008; Wegner & Wheatley, 1999). Despite a persistent lack of consensus
among the philosophical and scientific communities on the issue, members of the general public
believe strongly in free will, across ages and cultures (Nahmias, Morris, Nadelhoffer, & Turner,
2005; Nichols, 2004; Sarkissian et al., 2010). Furthermore, folk beliefs in this concept are
relevant to a variety of real-world moral consequences. For example, an experimental
manipulation that reduces free will beliefs causes an increase in antisocial behavior (Baumeister,
Masicampo & DeWall, 2009; Protzko, Ouimette, & Schooler, 2016; Vohs & Schooler, 2008) and
a decrease in punitiveness (Shariff et al., 2014). Performing antisocial acts and declining to
punish others who misbehave both can weaken the moral consensus that helps society benefit its
members.
1.2.1 Motivated free will judgments. Many factors contribute to free will beliefs,
including the powerful subjective experience of conscious will (Wegner, 2002, 2003) and
observations about causality (Nichols, 2004). However, one driving factor is a strong motive to
justify punishing immoral behaviors (Clark, Baumeister, & Ditto, 2017). In five studies, Clark
and colleagues (2014) demonstrated that exposure to the morally bad actions of others, and
subsequent desires to punish immoral actions lead people to attribute more free will to such
actions and to believe more in the general human capacity for free action. These findings
ARE MORALLY GOOD ACTIONS EVER FREE?
!
7!
provided empirical support for Nietzsche’s (1889, 1954) contention that the concept of free will
was invented to satisfy instincts to judge and punish others.
When it comes to moral responsibility judgments, the general consensus is that people
ought to be punished only for actions that they freely chose (Nichols & Knobe, 2007).
Accordingly, when people desire to punish others for harmful actions, they increase their
perceptions that the misdeeds were performed freely. In other words, desires to hold others
morally responsible (particularly for immoral actions) influence judgments about the very factor
necessary to warrant moral responsibility: whether the person performed the action of their own
free will. Although people generally regard their judgments as based on logical reasoning,
preferences for certain conclusions often influence the judgments people make and the beliefs
they hold (e.g., Baumeister & Newman, 1994; Ditto & Lopez, 1992; Kunda, 1990). Moral
judgments in particular are less often characterized by rigorous contemplation of moral
principles than by affective and intuitive reactions to potentially morally relevant stimuli (Clark,
Chen, & Ditto, 2015; Ditto, Pizarro, & Tannenbaum, 2009). Instead of preceding moral
judgment, moral reasoning often follows as a post-hoc justification for one’s desired moral
conclusions (Haidt, 2001; Liu & Ditto, 2013). In the case of free will judgments, desires to blame
and punish immoral actions lead people to perceive such actions as freely performed, thus
providing a crucial justification for punishing such actions (e.g., Alicke, 2000; Clark et al.,
2017).
One analysis of these motivated free will judgments suggests that attributing free will to
actors signifies that the actor is accountable for the outcome and capable of change or learning to
produce better outcomes in the future (Feldman et al., 2016). Consistent with that view, people
perceive others as incapable of learning in a universe without free will (Feldman &
ARE MORALLY GOOD ACTIONS EVER FREE?
!
8!
Chandrashekar, 2017). Much like punishment is an effective tool for teaching others to cooperate
(Cushman, 2013), ascribing free will to an individual may lead to feelings of personal
accountability, which could encourage individuals to learn from past mistakes (Alquist,
Ainsworth, Baumeister, Daly, & Stillman, 2015). Thus in addition to justifying punitive
responses, ascribing free will to immoral actors may effectively promote more desirable behavior
in the future.
1.3 Retributive Punishment
In principle, punishment could be a purely dispassionate, pragmatic response, engaged in
only for the sake of deterring bad behavior—or it could be affectively motivated. Although
punishment does effectively deter antisocial behavior, and people express support for deterrent
punitiveness (e.g., Crockett et al., 2014), the actual motivations underlying punishment appear
predominantly retributive (e.g., Carlsmith, 2008; Carlsmith et al., 2002). For example, Carlsmith
(2006) found that when determining how much to punish someone, people prefer information
related to retribution (e.g., the severity of the harm) over information related to the likely utility
of the punishment (e.g., whether the punishment will become public knowledge). Other work has
demonstrated that people will punish unfair others even when there will be no deterrent benefits
(Crockett et al., 2014). Punishing any particular instance of harmful behavior may effectively
deter that person from reoffending and deter others from engaging in similar harmful behaviors,
however, this does not appear to be the proximate reason humans punish one another. Rather,
humans want to punish harmdoers, even when the harm is directed toward unknown others, and
even when punishing comes at a personal cost (Fehr & Gächter, 2002; Henrich et al., 2005;
Henrich et al., 2006). Exposure to another’s harmful action elicits a strong affective response, to
wit, anger, which leads people to punish (e.g., Fehr & Fischbacher, 2004; Nelissen &
ARE MORALLY GOOD ACTIONS EVER FREE?
!
9!
Zeelenberg, 2009). These strong affective responses lead people to punish regardless of
potentially relevant characteristics of the harmful act and the punishment, such as whether
punishing the harmful action will have any utility.
Consider the concepts “blinded by rage” and “crime of passion.” The implications here
are that anger and desires for revenge can crowd out rational considerations and limit attention to
other important features of a situation. These strong punitive tendencies play an integral role in
society (e.g., Cushman, 2013; Fehr & Gächter, 2005). Unchecked human behavior is often selfish
(e.g., Karau & Williams, 1993; Kerr & Bruun, 1983; Latané, Williams, & Harkins, 1979), and
instances of selfish behavior can lead others to defect to the detriment of entire social groups
(Kerr, 1983; Orbell & Dawes, 1981). Punishment serves the crucial social function of deterring
selfish behavior (e.g., Fehr, Gächter, Kirchsteiger, 1997), and thus punishing rule breakers is
likely an indispensable feature of human cultures (e.g., Clutton-Brock & Parker, 1995; Henrich
et al., 2010).!
1.4 Punishment is Stronger than Praise
With respect to influencing future behavior, blame and punishment appear to be more
impactful than praise and reward. For example, Rasmussen and Newland (2008) found that
punishment had a greater impact on an individual’s future behavior than reinforcement of
equivalent magnitude. This asymmetry has been found in a number of dimensions related to
moral judgments, including evaluations of character, control, and intentionality. For example,
people are hypersensitive to moral violations, even at the expense of failing to recognize cases of
virtuous action (Monroe, Ainsworth, Vohs, & Baumeister, 2017). In fact, some analyses suggest
that rewarding prosocial behavior will sometimes undermine intrinsic motivation to help others
(Deci, Koestner, & Ryan, 1999; Fabes, Fultz, Eisenberg, May-Plumlee, & Christopher, 1989;
ARE MORALLY GOOD ACTIONS EVER FREE?
!
10!
Warneken & Tomasello, 2008, though also see Cameron, Banko & Pierce, 2001). Hence it seems
far more useful to punish bad behavior than to reward good behavior.
Nonetheless, both punishment and reward do seem to effectively mold future behavior
(e.g., Kubanek, Snyder, & Abrams 2015). Just as crime and the resulting punishment are a
normal part of society and play a significant role in defining socially acceptable behavior,
prosocial behavior, ranging from basic cooperation to altruism, and the recognition and
rewarding of that behavior contribute to successful group functioning (Irwin, 2009). Much like
punishment, reward can promote prosocial behavior—at least in some circumstances (Angrist &
Lavy, 2009; Ariely, Bracha, & Meier, 2007; Croson & Gneezy, 2009; Gneezy, Meier, & Rey-
Biel, 2011; Lacetera & Macis, 2010, Szolnoki & Perc, 2010). One study found that people
reward positive behavior approximately as often as they punish negative behavior, and
furthermore, that reward and punishment both increase net payoffs, suggesting that both
punishment and reward are important for optimal group functioning (Fehr et al., 1997). Due to
the value of rewarding morally good behavior, people may be motivated to see good actions as
freely performed and good actors as morally responsible.
1.5 Affective Responses to Morally Positive Behavior
As with punishment, reward could be motivated by either affectively charged feelings or
by dispassionate calculation of what will benefit the self and society. That is, people may reward
good behavior because they are compelled by positive emotional responses (like anger responses
lead to retributive punitive desires) — or they might confer rewards based on the notion that
rewarding virtuous behavior increases such behavior, thereby making society better for its
members (i.e., utilitarian concerns). Only recently have researchers begun investigating affective
responses to morally virtuous acts. The positive emotional response triggered by exposure to
ARE MORALLY GOOD ACTIONS EVER FREE?
!
11!
morally good acts has been termed elevation, which is associated with feelings of being uplifted
and increased desires to behave prosocially toward others (Haidt, 2003).
To our knowledge, no work has attempted to tease apart affective vs. more utilitarian
motives to praise or reward morally good behavior (as has been done with retributive vs.
deterrent punitiveness). There is, however, some reason to expect that praise and reward
responses would be more sensitive to utilitarian concerns than punitive motives. The Broaden-
and-Build Theory of positive emotions argues that positive emotions broaden the scope of
attention and expand awareness, thereby allowing people to consider more contextual
information (e.g., Frederickson, 2001, 2013; Frederickson & Branigan, 2005) and to incorporate
more relevant information into their decision-making, as compared to negative emotional states
(Isen, Rosenzweig, & Young, 1991). To our knowledge, there are no analogous concepts to
“blind rage” to describe relatively thoughtless and insensitive desires to praise or reward. A
recent literature review concluded that elevation actually expands awareness and flexible
thinking (Pohling & Diessner, 2016). Therefore, when evaluating moral responsibility for
morally positive actions, people may consider more contextual information, such as whether
rewarding positive behavior will encourage future good behavior.
1.6 The Present Research
When people are exposed to morally bad actions, they may be overwhelmed by strong
affective desires to punish, which renders their judgments insensitive to various features of the
act itself as well as features of the punishment (such as whether there will be any deterrent
benefits; e.g,. Crockett et al., 2014). When people are exposed to morally good actions, positive
feelings may broaden attention to contextual features of the specific act, allowing for more
critical evaluation and consideration of the likely utility of reward. If so, people might ascribe
ARE MORALLY GOOD ACTIONS EVER FREE?
!
12!
free will to morally bad acts regardless of whether punishment would have deterrent effects, due
to an overwhelming urge to hold wrongdoers morally responsible. But when people evaluate
morally good acts, their free will ascriptions might be sensitive to a variety of relevant features,
including whether rewards would encourage morally good behavior in the future.
The present work tested the hypothesis that people perceive morally positive actions as
more freely performed than morally neutral actions. We acknowledge that there are stronger
motives to interpret bad actions as freely performed than there are for morally good actions—
but exposure to good actions could also produce a non-negligible increment in perceived
freedom. Hence morally positive actions (relative to more neutral actions) might also inspire
motives to perceive moral actors, actions, and mankind, as free. This was tested in Studies 1a,
1b, and a mini-meta-analysis of the two studies. Studies 2a and 2b moved on to explore potential
reasons for this asymmetry by testing the hypothesis that free will judgments for immoral actions
are driven predominantly by affective punitive motives, whereas free will judgments for morally
positive actions are more sensitive to a variety of contextual features of morally good acts, in
particular, whether rewards will have utility.
2. Study 1a
Study 1a provided a first test of the hypothesis that people would believe more in free
will after considering morally good actions than morally neutral ones. Participants read and
evaluated an anti-free will research passage. Embedded in that passage were remarks from a
fictional academic about the implications of the research, which indicated that people are not
responsible for their everyday, mundane behavior (neutral condition), should not be praised for
their good behavior, or should not be punished for their bad behavior. We predicted that people
would most critically evaluate the anti-free will research in the morally bad condition and least
ARE MORALLY GOOD ACTIONS EVER FREE?
!
13!
critically in the neutral condition, with the morally good condition falling between these two.
This would indicate that people desire to uphold the idea of free will the most when they wish to
punish bad behavior, but also to a (lesser) degree when they wish to reward morally virtuous
behavior.
2.1 Method
Participants (n = 234) were recruited through Mechanical Turk (MTurk). Procedures
were adapted from Scurich and Shniderman (2014). The target sample size was 240
(~80/condition, based on the experimenter’s experience conducting similar research);1 however
research funds ran out after 234 participants. After reporting demographics, participants
completed a standard measure of free will belief, the free will subscale of the Free Will and
Determinism scale (FAD-Plus; Paulhus & Carey, 2011), on which people report their agreement
with a variety of statements such as “People have complete control over the decisions they
make” on a 5-point scale from Strongly disagree to Strongly agree, α = .88. Participants were
then randomly assigned to read and evaluate one of three newspaper articles, which described
recent scientific research opposing the existence of free will. In the morally bad condition, the
lead researcher commented that these results indicate that people should not be held responsible
and punished for their bad behavior. In the morally good condition, the researcher stated that
people should not be praised for their altruistic or morally good behavior. In the neutral
condition, the researcher stated that individuals are not responsible for everyday choices like
what clothes they wear or what route they take to work. Exact phrasing of the materials for each
study is available in the supplementary materials.
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
1 Adam Shniderman was the experimenter for this study. Scurich and Shniderman (2014)
recruited approximately 85 participants per condition and ended up with approximately 75
participants per condition in each of their two studies.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
14!
Following the newspaper article, participants evaluated the scientific quality of the
research on 11 questions, which were presented in randomized order (e.g., “Do you believe the
results of the study are likely to be replicable in the future?” “How likely is it that most scientists
would agree with the researchers’ conclusions?”), on relevant 7-point scales, α = .90. Because
people more critically evaluate information that opposes their beliefs than information that
supports their beliefs (e.g, Munro & Ditto, 1997), a person’s beliefs can be inferred from his or
her evaluation of information. More critical evaluations of the anti-free will argument would
indicate stronger belief in free will. This same method for assessing free will beliefs was also
used in Clark et al. (2014), as this more indirect measurement approach reduces the likelihood of
demand characteristics.
As comprehension checks, participants were asked to identify the technique used in the
fictional study (i.e., neuroscience) and to identify the topic of the news article (i.e., free will).
Participants also completed a simple attention check that stated, “To ensure this survey is
working properly, please click Strongly Agree.” Participants who failed the comprehension
checks (n = 20) or attention check (n = 7) were removed prior to analyses, resulting in a final
sample of 207 participants (98 female; Mage = 35.30). With this sample size, we would be able to
detect Cohen’s d effect sizes of roughly .42 (at p < .05 with 80% power; G*Power; Faul,
Erdfelder, Buchner, & Lang, 2009; Faul, Erdfelder, Lang, & Buchner, 2007).
2.2 Results
An Univariate Analysis of Variance (ANOVA) showed a significant effect of condition
on perceptions of scientific quality F(2, 204) = 4.39, p = .014, ηp2 = .041. Tukey’s post-hoc
analysis revealed that participants evaluated the anti-free will research as significantly less valid
in the morally bad condition (M = 3.18, SD = 1.20; n = 69) than the neutral condition (M = 3.77,
ARE MORALLY GOOD ACTIONS EVER FREE?
!
15!
SD = 1.17; n = 68), p =.009, Cohen’s d = .50, consistent with prior research (Clark et al., 2014).
As predicted, evaluations of the anti-free will research in the morally good condition (M = 3.48,
SD = 1.10; n = 70) fell between the morally bad (p = .301; Cohen’s d = .26) and morally neutral
(p = .294; Cohen’s d = .26) conditions, but these differences failed to reach statistical
significance.2
Because these results were statistically inconclusive, we also conducted a linear trend
analysis, which revealed a significant linear trend demonstrating a decrease in motivation to
believe in free will from the morally bad to morally good to morally neutral conditions, F(1, 204)
= 8.79, p = .003. Though note that the ANOVA in the preceding paragraph was planned, whereas
this analysis was not.
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2 Repeating these analyses without participant exclusions revealed a marginal effect, F(2, 228) =
2.85, p = .059. Tukey’s post-hoc analysis showed significant differences (p = .048; Cohen’s d =
.37) between the morally bad (M = 3.21, SD = 1.13) and the morally neutral conditions (M =
3.64, SD = 1.19). No other significant differences were found.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
16!
Figure 1. Scientific quality ratings by condition in Study 1a (scale range: 1-7).
2.3 Discussion
Consistent with expectations, participants were most critical of the anti-free will passage
in the morally bad condition, followed by the morally good condition, followed by the neutral
condition. However, the morally good condition did not significantly differ from either. A
possible explanation for the weak findings was that our manipulation was too subtle. For this
reason, we followed up with Study 1b, in which participants read about concrete instances of
morally good and morally bad behavior in order to elicit stronger desires to blame and praise.
We then conducted mini-meta-analyses of the results of these two initial studies to increase
confidence in this pattern of findings.
3. Study 1b
In Study 1b, participants read what was ostensibly a brief news report about a morally
bad, morally good, or morally neutral behavior, and then evaluated how freely performed the
behavior was and reported their desire to blame or praise the actor. Participants also completed a
general measure of free will belief. For both free will attributions and free will beliefs, we
predicted that free will judgments would be highest in the morally bad condition, followed by the
morally good condition, followed by the neutral condition. Furthermore, we expected desires to
blame to mediate the difference between the morally bad and neutral conditions on both free will
attributions and free will beliefs, and we expected desires to praise to mediate the difference
between the morally good and neutral conditions on both free will attributions and free will
beliefs.
3.1 Method
ARE MORALLY GOOD ACTIONS EVER FREE?
!
17!
Participants (175 undergraduates; 125 Female; Mage = 20.65) were randomly assigned to
read one of three ostensible news stories about 1) a thief stealing expensive lasers from a
pediatric cancer hospital (morally bad condition), 2) a hospital administrator purchasing
expensive lasers for a pediatric cancer hospital (neutral condition), or 3) a contributor donating
expensive lasers to a pediatric cancer hospital (morally good condition). The target sample size
was 180 (~60/condition, based on the experimenter’s experience conducting similar research); 5
participants who signed up for the study failed to complete the experimental procedures,
resulting in 175.3 With this sample size, we would be able to detect Cohen’s d effect sizes of
roughly .47 (at p < .05 with 80% power; G*Power; Faul et al., 2009; 2007).
Participants then rated the extent to which the actor behaved of his own free will on three
items: whether the action was freely chosen, whether the actor could have made other choices,
and whether the actor exercised his own free will in choosing to perform the action on 7-point
scales from Not at all to Very much so,
= .92. Participants in the morally good and neutral
conditions then reported the extent to which the actor was morally praiseworthy, and participants
in the morally bad and neutral conditions reported the extent to which the actor was morally
blameworthy on 7-point scales from Not at all to Very much so. Last, participants reported their
free will beliefs on the FAD-Plus (Paulhus & Carey, 2011;
= .77).4 Participants also reported
demographic information.
3.2 Results
3.2.1 Free will attributions. An ANOVA revealed a significant effect of condition on
free will attributions, F(2, 170) = 24.60, p < .001, ηp2 = .224 (see Figure 2). As hypothesized,
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
3 Cory Clark was the experimenter for this study. Clark et al. (2017) aimed for approximately 60
participants per condition in each of their three experimental studies (Studies 3-5).
4 Participants also completed the scientific and fatalistic determinism subscales of the FAD-Plus,
but these were not analyzed.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
18!
Tukey’s post-hoc tests revealed that participants attributed more free will to the morally good
actor (n = 49; M = 5.71, SD = 1.15) than the morally neutral actor (n = 54; M = 4.56, SD = 1.07),
p<.001, Cohen’s d = 1.04. Replicating prior work, participants also attributed more free will to
the morally bad actor (n = 70; M = 6.05, SD = 1.32) than the morally neutral one, p<.001,
Cohen’s d = 1.24 (Clark et al., 2014). As in Study 1a, participants attributed more free will to the
bad actor than to the good actor with a small to medium effect size, Cohen’s d = .27, but this
difference failed to reach statistical significance, p = .297.5
Figure 2. Attributions of free will by condition in Study 1b (scale range: 1-7).
3.2.2 Free will beliefs. There was a marginal effect of condition on general free will
beliefs, F(2, 165) = 2.81, p = .063, ηp2 = .033 (see Figure 3). Mirroring the pattern for free will
attributions, free will beliefs were higher in the morally good condition (n = 47; M = 3.68, SD =
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
5 Please see the supplementary file under the heading Unreported Studies X and Y, which
describes two similar studies that replicated this pattern of results.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
19!
.51) than the neutral condition (n = 51; M = 3.49, SD = .72), though this difference failed to
reach statistical significance, p = .272, Cohen’s d = .30. Free will beliefs were marginally higher
in the morally bad condition (n = 70; M = 3.75, SD = .56) than the neutral condition, p = .053,
Cohen’s d = .40, whereas the morally bad and morally good conditions did not differ, p = .805,
Cohen’s d = .13.
Figure 3. Free will beliefs by condition in Study 1b (scale range: 1-5).
3.2.3 Praise and blame. Independent samples t-tests revealed that participants felt the
morally good actor was more morally praiseworthy (n = 49; M = 6.00, SD = 1.21) than the
morally neutral actor (n = 52; M = 5.10, SD = 1.26), t(99) = 3.68, p <.001, Cohen’s d = .73, and
that the morally bad actor was more morally blameworthy (n = 72; M = 5.69, SD = 1.89) than the
morally neutral actor (n = 54; M = 3.24, SD = 1.59), t(124) = 7.71, p <.001, Cohen’s d = 1.40.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
20!
A bootstrap mediation analysis (10,000 resamples; PROCESS macro [Hayes, 2013])
revealed that perceptions that the morally good actor was more deserving of moral praise
partially accounted for higher attributions of free will to the morally good actor, 95% CI [.109,
.682]. This model is depicted in Figure 4. A similar but weaker pattern emerged for general free
will beliefs, such that controlling for perceived praiseworthiness reduced the effect size of the
neutral vs. good condition on free will belief, from semipartial r = .15, p = .156, to semipartial r
= .06, p = .523, though the mediation was not statistically significant, 95% CI [-.024, .362]. This
is likely due to the relatively small difference in free will beliefs between the neutral and morally
good conditions.
Replicating Clark and colleagues (2014), perceptions that the morally bad actor was more
deserving of moral blame partially accounted for higher attributions of free will to the morally
bad actor, 95% CI [.484, 1.429],6 and fully mediated the influence of the morally bad condition
on general free will beliefs, 95% CI [.020, .345].
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
6 Please see the supplementary file under the heading Unreported Studies X and Y for results of
two similar studies, in which one study replicated this mediation pattern and one did not.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
21!
3.3 Discussion
With closely matched behaviors, participants attributed more free will to the morally
good actor than the morally neutral actor, and this relationship was mediated by stronger desires
to praise the morally good actor. Though a similar pattern emerged for general free will beliefs,
the results were not statistically significant. As expected, these relationships were generally
stronger in the morally bad conditions such that participants attributed the most free will to the
morally bad actor and believed marginally more in free will in the morally bad condition.
Further, greater desires to blame mediated the influence of the morally bad condition on both
free will attributions and free will beliefs. It is clear that harmful actions that elicit desires to
blame and punish have a very strong influence on free will judgments, but the present results
indicate that morally good actions that elicit desires to praise have a similar albeit weaker effect.
Undoubtedly, people ascribe more blame and praise to an actor who had more freedom
and control over causing a particular harmful or helpful outcome (e.g., consider the distinctions
between murder, voluntary manslaughter, and involuntary manslaughter). The present results
suggest that the reverse causal direction may also occur: people may increase perceptions that a
behavior was performed freely when desiring to blame or praise. Of course, it is challenging to
make causal conclusions from mediation models; often, both causal patterns (IV to mediator to
DV and IV to DV to mediator) are statistically significant. This was the case in Study 1b as well:
all significant mediation models were also significant if the mediator and DV were reversed.
Nonetheless, our proposed causal order might make more sense a priori. As our IVs were
intended to manipulate the mediator directly, it seems less likely that participants would only
desire to blame or praise harmful and helpful actions because they were perceived as more freely
performed. Independent of increased perceptions of freedom, it is likely that people would
ARE MORALLY GOOD ACTIONS EVER FREE?
!
22!
desire to praise morally good actions simply because they are morally good and blame morally
bad actions simply because they are morally bad. Thus, it seems likely that there is bidirectional
causality such that high perceptions of freedom increase blame and praise, but desires to blame
and praise can also increase perceptions of freedom.
4. Mini Metas
Because some of the results of Studies 1a and 1b were statistically inconclusive, we
conducted three mini meta-analyses of the effect sizes of the differences in free will judgments
between neutral and good, neutral and bad, and good and bad. In recent years, psychological
science has encouraged greater emphasis on meta-analyses of overall patterns of findings across
studies rather than on the statistical significance of individual studies (e.g., Maner, 2014). A
meta-analysis of the results of studies testing similar questions provides better evidence of
whether a given effect is real than the statistical significance of individual tests (e.g., Braver,
Thoemmes, Rosenthal, 2014). We employed many of the suggested procedures outlined by Goh,
Hall, and Rosenthal (2016) to conduct our three mini metas. Given that a large body of work has
demonstrated that people attribute more responsibility to morally bad than morally good actions
and neutral actions, we were primarily interested in meta-analyzing whether people attributed
significantly more free will to morally good actors than morally neutral ones. Because we are
specifically interested in interpreting the statistical significance of Studies 1a and 1b (particularly
for the good vs. neutral comparison), we interpret fixed effects (Goh et al., 2016). Random
effects are likely too conservative to detect significant effects with such a limited number of
studies, nonetheless, we also report random effects as suggested by Goh and colleagues (2016;
see also Hedges & Vevea, 1998).
ARE MORALLY GOOD ACTIONS EVER FREE?
!
23!
We included one effect size for each study (i.e., the effect sizes for free will attributions
and free will beliefs were averaged in Study 1b; Card, 2012), thus two effect sizes were included
in each mini meta (the minimum necessary for a meta-analysis; Goh et al., 2016; Valentine,
Pigott, & Rothstein, 2010). We computed r effect sizes from the Ms, SDs, and ns. For fixed
effects, all rs were Fisher’s z transformed to rz which were then weighted and averaged using the
following formula: Weighted r̄z = Σ ([N-3]rz) / Σ (N-3). The weighted r̄z s were then converted
back to Pearson correlations for presentation. To determine statistical significance, we utilized
the Stouffer’s Z test, in which the p values for each r effect size were converted to Zs. The Zs
were then combined using the following formula: Zcombined = Σ Z / sqrt(k), and then converted
back to ps for presentation. For random effects, we conducted one-sample t-tests of the effect
sizes.7 All of these procedures are more thoroughly described in Goh et al. (2016).
As can be seen in summary Table 1, consistent with our main hypothesis, free will
judgments were significantly higher in the good than neutral conditions with a small to medium
effect size, r = .21, p = .0007 (random effects: r = .23, p = .254). Consistent with Clark and
colleagues (2014), free will judgments were significantly higher in the bad than neutral
conditions, r = .30, p < .0001 (random effects: r = .31, p = .134). There was also a marginal
effect such that free will judgments were higher in the morally bad than morally good conditions,
r = .12, p = .064 (random effects: r = .12, p = .083).8,9
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
7 These data and the syntax are publically available.
8 If we add the effect sizes for the good vs. bad comparisons in the upcoming Studies 2a and 2b,
this difference is significant, r = .14, p < .0001 (random effects: r = .14, p = .006). This
relationship is also consistent with the large body of similar research showing higher attributions
of responsibility for morally bad than morally good actions. Thus we feel there is little reason to
doubt that this effect is real, even if it is somewhat small here.
9 If we also add the effect sizes for two unreported studies that were similar in design to Study 1b
(see Unreported Studies X and Y and Table 2S in Supplementary File), all conditions
significantly differ, ps < .0001 (random effects: ps < .020).
ARE MORALLY GOOD ACTIONS EVER FREE?
!
24!
Table 1. Meta-analyzed Effect Sizes for Good vs. Bad, Good vs. Neutral, and Bad vs. Neutral
Free Will Measure/s
Comparison
Good vs. Bad
Good vs. Neutral
Bad vs. Neutral
Study 1a
Evaluation of Anti-Free
r
0.13
0.13
0.24
Will Research (reversed)
n
139
138
137
p
0.127
0.129
0.0047
Study 1b
Attributions and Beliefs
r
0.10
0.32
0.37
(combined)
n
119
103
124
p
0.279
0.001
<0.0001
Unweighted (random)
r
0.12
0.23
0.31
p
0.083
0.254
0.134
Weighted (fixed)
r
0.12
0.21
0.30
p
0.064
0.0007
<.0001
5. Study 2a
The mini meta of Studies 1a and 1b supported our first hypothesis that free will
judgments would be higher for morally positive actions than morally neutral ones, and (to some
extent) that free will judgments would be lower for morally positive actions than for morally bad
ones (see Footnote 8 for mini meta-analysis results including upcoming Study 2a and 2b data,
which confirm this relationship). In Studies 2a and 2b, we moved on to test our second
hypothesis that free will judgments for immoral actions would be predominantly sensitive to
affective punitive motives, whereas free will judgments for morally positive actions would be
sensitive to other relevant information such as the utility of reward. Research has shown that the
motives underlying punishment appear to be predominantly retributive and that the actual utility
of punishment has little influence on punishment decisions (e.g., Carlsmith et al., 2002). When
people are exposed to others’ morally bad actions, they experience strong impulses to punish the
harmdoer, and they are relatively insensitive to considerations such as whether the punishment
ARE MORALLY GOOD ACTIONS EVER FREE?
!
25!
would deter bad behavior. Therefore, we expected that free will judgments for immoral actions
would be sensitive to retributive concerns only.
In contrast, positive emotional reactions to morally good actions may allow individuals to
more critically evaluate various aspects of the act itself, including the likely utility of the reward.
Relative to negative emotions, positive emotions broaden the scope of awareness and allow
people to consider more relevant information in their decision-making. Therefore, when deciding
whether a morally good act is deserving of praise, people might consider the likely utility of the
reward, and other features of the act in question. We expected that free will judgments for
morally positive actions would be more contextually sensitive than those for immoral actions,
and in particular, more sensitive to utilitarian concerns.
Methods for Studies 2a and 2b were adapted from studies by Carlsmith and colleagues
(2002), in which the deservingness of punishment (retributive motive) was manipulated by
varying the severity of the immoral act, and utility (deterrent motive) was manipulated by
varying the publicity of the punishment. We applied this paradigm to the present work by having
participants read about either a morally good or morally bad act and manipulating the
deservingness and utility of the punishment or reward. As in Carlsmith and colleagues’ (2002)
studies, deservingness was manipulated by varying the magnitude of the action, and utility was
manipulated by specifying whether the reward or punishment would become public knowledge
with the explicit purpose of either deterring others from doing the same (in the morally bad
condition) or encouraging others to do the same (in the morally good condition).
Participants then recommended a punishment or reward for the actor, rated the magnitude
of the action, rated the likely utility of the punishment or reward, rated how freely performed the
action was, and reported their general free will belief. We expected that in the morally bad
ARE MORALLY GOOD ACTIONS EVER FREE?
!
26!
condition, people would punish more severely, attribute more free will, and believe more in free
will in the high deservingness condition than the low deservingness condition, but that these
judgments would not differ between the high and low utility conditions. In the morally good
condition, we expected that reward, free will attributions, and free will beliefs would be higher in
the high deservingness condition than the low deservingness condition, but also in the high
utility condition than the low utility condition. These results would suggest that praise motives
and ascriptions of free will for morally positive behaviors are sensitive to both deservingness and
utility.
For exploratory purposes, participants also reported their emotional reaction, rated how
normative the behavior was, and rated how much willpower was exerted. If free will judgments
for immoral actions are primarily driven by affective punitive responses, these judgments should
be predicted most strongly by participants’ emotional reactions. If free will judgments for
morally positive actions are more carefully considered and context-dependent, they might be
better predicted by ratings of normativity and required willpower.
5.1 Method
In a 2 (Deservingness: High vs. Low) × 2 (Utility: High vs. Low) × 2 (Moral Valence:
Bad vs. Good) between-subjects design, participants (451 undergraduates; 285 female; Mage =
19.27) were randomly assigned to read one of eight short stories about an ostensible recent event,
involving John, a district manager at a large department store, who performed either a morally
admirable or morally deplorable action. The target sample size was 480 (again ~60/condition)10
or however many participants could be recruited by the end of the semester. The end of the
semester came first, resulting in 451 participants. Participants were randomly assigned to read
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
10 The target numbers of participants per condition were the same in Studies 1b and 2a because
the same experimenter (Cory Clark) performed these studies.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
27!
that John had either falsely reported earnings to steal from the company (morally bad condition)
or sacrificed his own bonus to implement an employee incentive program that increased profits
for the company (morally good condition). The amount of money stolen or earned was also
manipulated: $9.2 million in the high deservingness conditions vs. $24,000 in the low
deservingness conditions. Last, the utility of John’s punishment or reward was manipulated by
indicating whether the punishment or reward would become public knowledge in order to deter
others (in the morally bad condition) or encourage others (in the morally good condition).
Participants then rated John’s free will on the same three free will attribution items from
Study 1b, on 9-point scales. Participants in the morally good condition then reported how much
John should be rewarded and participants in the morally bad condition reported how much John
should be punished on two questions. The first question asked how severely John should be
punished or how highly John should be rewarded on a 7-point scale from Not at all to Extremely.
They were then given concrete punishment or reward options. Participants in the morally bad
condition were asked what an appropriate prison sentence for John would be with 14 options
increasing in severity from no prison to 50 years to life. Participants in the morally good
condition were asked what an appropriate reward bonus for John would be with 15 options
increasing in amount from no bonus to more than $1,000,000. The two punishment questions
and the two reward questions were converted to Z-scores and combined into individual
punishment and reward indices.
Participants then evaluated the magnitude of the action by reporting either how serious
John’s crime was or how generous John’s actions were, which also served as a deservingness
manipulation check. Participants also rated the utility of the response by indicating how likely it
is that John’s punishment would deter others from doing the same thing or that John’s reward
ARE MORALLY GOOD ACTIONS EVER FREE?
!
28!
would encourage others to do the same thing, which also served as a utility manipulation check.
To gauge emotional reactions, participants were asked how morally outraged or morally uplifted
they were by John’s behavior. Action magnitude, response utility, and emotional reactions all
were reported on 7-point scales from Not at all to Extremely.
Participants then rated how normative the behavior was by indicating how likely it is that
someone else in John’s position would have done the same thing rated on a 7-point scale from
Not at all likely to Extremely likely and how much John used his willpower rated on a 7-point
scale from Not at all to Completely. Last, participants completed the free will subscale of the
FAD-Plus (Paulhus & Carey, 2011), reported demographic information, and were debriefed.
5.2 Results
5.2.1 Action magnitude manipulation check. A 2 (deservingness) × 2 (utility) × 2
(moral valence) ANOVA revealed a main effect of the deservingness manipulation on action
magnitude ratings, F(1, 443) = 13.06, p < .001, η2
p = .029, such that participants in the high
deservingness conditions rated John’s actions as more serious/generous (M = 5.97, SD = 1.09; n
= 228) than participants in the low deservingness conditions (M = 5.64, SD = 1.20; n = 223).11
5.2.2 Response utility manipulation check. A 2×2×2 ANOVA revealed a main effect of
the utility manipulation, F(1, 443) = 9.86, p = .002, η2
p = .022, such that participants in the high
utility conditions reported that John’s punishment/reward would be more likely to
discourage/encourage others (M = 4.89, SD = 1.41; n = 224) than participants in the low utility
conditions (M = 4.42, SD = 1.82; n = 227).
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
11 For space purposes, only significant or marginal effects that are relevant to the present
hypothesis are reported in text.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
29!
5.2.3 Punishment. A 2 (deservingness) × 2 (utility) ANOVA on the punishment index
revealed just one main effect for deservingness, F(1, 219) = 16.07, p <.001, η2
p = .068, such that
participants in the high deservingness condition recommended harsher punishment (M = .22, SD
= .91; n = 116) than those in the low deservingness condition (M = -.24, SD = .77; n = 107).
5.2.4 Reward. A 2 (deservingness) × 2 (utility) ANOVA on the reward index revealed a
main effect for deservingness, F(1, 222) = 4.51, p = .035, η2
p = .020. Participants in the high
deservingness condition recommended higher reward (M = .12, SD = .95) than those in the low
deservingness condition (M = -.12, SD = .81). There was no main effect for the utility condition,
but there was an interaction between the utility and deservingness manipulations, F(1, 222) =
4.75, p = .030, η2
p = .021. In the low utility condition, deservingness of the reward had no
influence on how highly participants rewarded (high deservingness M = .017, SE = .12; n = 56;
low deservingness M = .023, SE = .12; n = 58), F(1, 222) = .002, p = .968, η2
p = .000. In the high
utility condition, participants in the high deservingness condition recommended higher rewards
(M = .24, SE = .12; n = 54) than participants in the low deservingness condition (M = -.27, SE =
.12; n = 58), F(1, 222) = 9.17, p = .003, η2
p = .040. See Figure 5 for graphs of the interactions
between the deservingness and utility manipulations on desires to punish and reward in both
Studies 2a and 2b.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
30!
Figure 5. Interactions between the deservingness and utility manipulations on desires to punish
(left column) and reward (right column) in Study 2a (top row) and Study 2b (bottom row).
5.2.5 Free will judgments. For free will attributions, there was only a main effect of the
moral valence condition, F(1, 443) = 30.37, p < .001, η2
p = .064, such that participants attributed
more free will to the morally bad actor (M = 8.52, SD = .93; n = 223) than to the morally good
ARE MORALLY GOOD ACTIONS EVER FREE?
!
31!
actor (M = 7.86, SD = 1.53; n = 228). The same result was found for general free will belief.
Participants in the morally bad condition believed more in free will (M = 4.00, SD = .64; n =
222) than participants in the morally good condition (M = 3.84, SD = .70; n = 228), F(1, 442) =
6.74, p = .010, η2
p = .015. Contrary to expectations, there were no interactions with deservingness
or utility on either free will attributions or beliefs.
5.2.6 Exploratory analyses. Study 2a replicated the results of the mini meta and prior
work demonstrating that people ascribe more free will to morally bad actions than morally good
ones, and the deservingness and utility manipulations worked as expected. Also as expected,
punishment was only sensitive to the deservingness manipulation whereas reward was sensitive
to both deservingness and utility. However, against our hypothesis, no interactions were found
between the deservingness or utility manipulations with the moral valence condition on free will
attributions or beliefs.
We hypothesized that free will judgments for morally bad actions would be driven
primarily by affective retributive concerns (and not utility concerns), but that free will judgments
for morally good actions would be sensitive to utility concerns. Despite the failure of our
manipulations to demonstrate such findings, we included a variety of other evaluation variables
that could help test this hypothesis. For this reason, we conducted additional exploratory
analyses to test whether free will judgments for morally bad actions were primarily predicted by
participants’ ratings of the magnitude of the immoral act, emotional responses, and desires to
punish (outcomes relevant to affect and retribution), whereas free will judgments for morally
good actions might be more sensitive to less emotional features such as whether the reward
would be likely to encourage good behavior, how much willpower the action required, and the
degree of counternormativity of the action.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
32!
We analyzed the interactions between the moral valence condition and each of the six
following evaluation variables: action magnitude ratings, response utility ratings, emotional
reaction, normativity ratings, willpower ratings, and punish/reward response on both free will
attributions and free will beliefs. Six separate regressions were performed for both free will
attributions and free will beliefs. Each included the moral valence condition, one of the six
evaluation variables, and the interaction between the moral valence condition and that variable.
As can be seen in Tables 2 (free will attributions) and 3 (free will beliefs), there were significant
main effects for all six evaluation variables on both free will attributions and beliefs (with the
exception of normativity ratings on free will beliefs), such that higher free will judgments were
predicted by higher ratings of action magnitude, higher ratings that the punishment or reward
would have utility, stronger emotional reactions, lower ratings of normativity, higher ratings of
willpower, and stronger desires to punish or reward. Many of these were qualified by interactions
with the moral valence condition, which will be described below.
5.2.6.1 Action magnitude. Higher ratings of action magnitude predicted higher
attributions of free will in both the morally good and morally bad conditions. Simple slopes
within the morally good and morally bad conditions demonstrated that higher action magnitude
ratings were stronger predictors of higher free will attributions in the morally good condition (b
=.75), t = 10.54, p < .001, than in the morally bad condition (b =.21), t = 3.00, p = .003. This
interaction did not emerge for free will beliefs. Rather, action magnitude ratings predicted higher
free will beliefs in both the morally good and bad conditions to equal degrees.
5.2.6.2 Response utility. There was no interaction between the moral valence condition
and response utility on free will attributions or free will beliefs. Higher utility of the punishment
and reward both predicted higher free will attributions and beliefs.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
33!
5.2.6.3 Emotional reaction. In the morally bad condition, greater moral outrage
significantly predicted higher free will attributions (b =.54), t = 9.84, p < .001. However, this
effect was larger in the morally good condition such that feeling morally uplifted more strongly
predicted higher free will attributions (b =.75), t = 13.66, p < .001. There was no interaction for
free will beliefs. A stronger emotional response predicted higher free will beliefs in both the
morally good and bad conditions to equal degrees.
5.2.6.4 Normative. Participant ratings of the likelihood that other people in the same
position would have done the same thing did not predict free will attributions to the morally bad
actor (b = -.01), t = -0.18, p = .855. In contrast, participants attributed more free will to the
morally good actor as their perceived likelihood that others in the same position would have
done the same thing decreased (b =-.21), t = -3.26, p = .001. Normativity ratings did not predict
free will beliefs in either condition.
5.2.6.5 Willpower. In the morally bad condition, there was only a marginal relationship
between willpower ratings and free will attributions (b =.06), t = 1.77, p = .077, and no
relationship between willpower ratings and free will beliefs (b =.02), t = 1.36, p = .175. In the
morally good condition, participants who felt that the actor exerted more willpower attributed
him more free will (b =.54), t = 9.88, p < .001, and believed more in free will generally (b =.09),
t = 2.81, p = .005.
5.2.6.6 Punish/reward response. There was no interaction between the condition on the
extent to which desires to reward or punish increased free will attributions and beliefs. Stronger
desires to reward predicted higher free will attributions and beliefs in the morally good condition
and stronger desires to punish predicted higher free will attributions and beliefs in the morally
bad condition.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
34!
Table 2
Free Will Attributions Regressed on Moral Valence Condition, Evaluation Variables, and Interactions
in Study 2a (Exploratory Analyses)
F
R2
b
SE
t
p
95% CIs
Free Will Attributions (Model)
65.97
.51
<.001
Moral valence condition
1.04
.11
9.16
<.001
.82, 1.26
Action magnitude
.75
.07
10.30
<.001
.60, .89
Interaction
-.53
.10
-5.42
<.001
-.73, -.34
Free Will Attributions (Model)
11.92
.07
<.001
Moral valence condition
.69
.12
5.77
<.001
.38, .66
Response utility
.10
.05
2.06
.040
.04, .17
Interaction
-.06
.07
-0.78
.435
-.20, .09
Free Will Attributions (Model)
21.94
.13
<.001
Moral valence condition
.75
.12
6.41
<.001
.52, .98
Emotional reaction
.31
.06
5.41
<.001
.20, .43
Interaction
-.21
.08
-2.65
.008
-.36, -.05
Free Will Attributions (Model)
14.04
.09
<.001
Moral valence condition
.75
.12
6.06
<.001
.50, .99
Normative
-.21
.06
-3.30
.001
-.33, -.08
Interaction
.20
.09
2.27
.024
.03, .37
Free Will Attributions (Model)
42.55
.22
<.001
Moral valence condition
.99
.12
8.63
<.001
.77, 1.22
Willpower
.54
.06
9.39
<.001
.43, .66
Interaction
-.49
.07
-7.29
<.001
-.62, -.35
Free Will Attributions (Model)
19.41
.12
<.001
Moral valence condition
.66
.12
5.68
<.001
.43, .89
Punish/reward response
.37
.09
4.01
<.001
.19, .56
Interaction
-.08
.13
-.57
.571
-.34, .19
Note. For ease of comparison, bold indicates interactions that are significant in Study 2b, which had a
larger sample size.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
35!
Table 3
Free Will Beliefs Regressed on Moral Valence Condition, Evaluation Variables, and Interactions in
Study 2a (Exploratory Analyses)
F
R2
b
SE
t
p
95% CIs
Free Will Belief (Model)
9.73
.06
<.001
Moral valence condition
.27
.07
4.11
<.001
.14, .40
Action magnitude
.13
.04
3.05
.002
.05, .21
Interaction
.01
.06
0.18
.856
-.10, .12
Free Will Belief (Model)
8.80
.06
<.001
Moral valence condition
.20
.06
3.19
.002
.08, .32
Response utility
.08
.03
3.20
.001
.03, .13
Interaction
.00
.04
0.07
.948
-.07, .08
Free Will Belief (Model)
10.53
.07
<.001
Moral valence condition
.21
.06
3.34
.001
.09, .33
Emotional reaction
.12
.03
3.89
<.001
.06, .18
Interaction
-.03
.04
-0.77
.444
-.12, .05
Free Will Belief (Model)
3.18
.02
.024
Moral valence condition
.13
.07
2.03
.043
.00, .26
Normative
.04
.03
1.25
.212
-.02, .11
Interaction
-.01
.05
-0.15
.881
-.10, .08
Free Will Belief (Model)
5.11
.03
.002
Moral valence condition
.23
.07
3.42
.001
.10, .36
Willpower
.09
.03
2.67
.008
.02, .15
Interaction
-.07
.04
-1.73
.084
-.14, .01
Free Will Belief (Model)
7.81
.04
<.001
Moral valence condition
.17
.06
2.64
.008
.04, .29
Punish/reward response
.13
.05
2.69
.007
.04, .23
Interaction
.02
.07
0.27
.785
-.12, .16
Note. For ease of comparison, bold indicates interactions that are significant in Study 2b, which had a
larger sample size.
5.3 Discussion
Study 2a explored how deservingness and utility differentially predicted punitive and
reward responses as well as free will judgments for morally good and bad behaviors. Replicating
the results of the mini meta and prior work, participants attributed more free will to the morally
bad actor than the morally good actor. Furthermore, as hypothesized, participants were primarily
ARE MORALLY GOOD ACTIONS EVER FREE?
!
36!
concerned with how deserving the immoral actor was when determining punishment.
Participants advocated harsher punishments for more severe misdeeds, regardless of whether
anyone else would find out, and thus regardless of whether the punishment would be likely to
deter others from committing similar crimes. When determining rewards, there was some
evidence indicating that participants were concerned with both deservingness and the likelihood
that the reward would improve others’ moral behavior. Participants only advocated greater
rewards for more virtuous actions when the rewards would be publicly known. However, these
differences did not manifest in free will judgments as we had hypothesized.
Nonetheless, the exploratory analyses revealed a number of useful findings. Specifically,
free will judgments for morally bad actions were predominantly predicted by responses related to
retributive feelings (action magnitude ratings, emotional response, and desires to punish), but not
the less emotional responses (willpower and normativity ratings). The main exception was
perceived utility, which predicted higher free will attributions and beliefs for both morally good
and bad actions. Free will judgments for morally good actions were predicted by all of these,
including willpower and normativity.
Though these results seem generally consistent with our hypothesis that free will
judgments for morally good actions are more contextually sensitive than free will judgments for
morally bad acts, whereas free will judgments for morally bad acts are predominantly driven by
affective punitive motives, they were not predicted a priori. To increase confidence that these
effects were real, we conducted a near replication of Study 2a in order to confirm these results.
Furthermore, many of these effects emerged only for free will attributions and not free will
beliefs. We may have failed to find similar effects for free will beliefs because they were
measured at the very end of the study, or because free will beliefs are generally more stable and
ARE MORALLY GOOD ACTIONS EVER FREE?
!
37!
harder to manipulate. Nonetheless, we expected similar patterns of results for general free will
beliefs as for free will attributions, even if the effect sizes would be smaller. For this reason, a
larger sample size was collected for Study 2b.
6. Study 2b
Study 2b was designed to confirm and replicate Study 2a. Because some of the effects
that emerged for free will attributions did not emerge for free will beliefs, we recruited a larger
sample size to increase our ability to detect any similar effects on general free will beliefs and
any smaller influences on free will attributions.
6.1 Method
Study 2b (n = 784 MTurk participants; 475 female; Mage = 34.66) was a near identical
replication of Study 2a except the punish/response, action magnitude rating, utility rating, and
emotional response came before the free will attributions, which included one additional item:
“Relative to typical, every day sorts of actions, to what extent were John’s actions more or less
free?” rated on a 9-point scale from Much less free to Much more free,
α
= .74. Sample size was
determined by the maximum number of participants that $650 could pay for at $0.65/participant
including MTurk and TurkPrime fees.
6.2 Results
6.2.1 Action magnitude manipulation check. A 2×2×2 ANOVA replicated the main
effect of the deservingness manipulation, F(1, 776) = 48.66, p < .001, η2
p = .059, such that
participants in the high deservingness conditions rated John’s actions as more serious/generous
(M = 6.28, SD = .95; n = 389) than participants in the low deservingness conditions (M = 5.79,
SD = 1.16; n = 395).
ARE MORALLY GOOD ACTIONS EVER FREE?
!
38!
6.2.2 Response utility manipulation check. A 2×2×2 ANOVA replicated the main
effect of the utility manipulation, F(1, 776) = 31.82, p < .001, η2
p = .039, such that participants in
the high utility conditions reported that John’s punishment/reward would be more likely to
discourage/encourage others (M = 4.99, SD = 1.42; n = 393) than participants in the low utility
conditions (M = 4.35, SD = 1.80; n = 391).
6.2.3 Punishment. As in Study 2a, a 2 (deservingness) × 2 (utility) ANOVA on the
punishment index revealed just one main effect for deservingness, F(1, 389) = 82.93, p <.001, η2
p
= .176, such that participants in the high deservingness condition recommended harsher
punishment (M = .36, SD = .80; n = 196) than those in the low deservingness condition (M = -
.36, SD = .77; n = 197). There was again no main effect for the utility condition, p = .749.
6.2.4 Reward. As in Study 2a, a 2 (deservingness) × 2 (utility) ANOVA on the reward
index revealed a main effect for deservingness, F(1, 385) = 78.57, p < .001, η2
p = .169. Rather
than an interaction between deservingness and utility (as in Study 2a), there was a main effect of
utility, F(1, 385) = 5.29, p = .022, η2
p = .014. Participants in the high deservingness condition
recommended higher reward (M = .34, SD = .81; n = 192) than those in the low deservingness
condition (M = -.34, SD = .71; n = 197). Participants in the high utility condition also
recommended higher reward (M = .09, SD = .80; n = 195) than those in the low utility condition
(M = -.09, SD = .86; n = 194).
6.2.5 Free will. As in Study 2a, for free will attributions, there was only a main effect of
the moral valence condition, F(1, 776) = 13.15, p < .001, η2
p = .017, such that participants
attributed more free will to the morally bad actor (M = 8.14, SD = 1.05; n = 393) than to the
morally good actor (M = 7.85, SD = 1.21; n = 391). The same result was found for general free
will belief. Participants in the morally bad condition believed more in free will (M = 4.06, SD =
ARE MORALLY GOOD ACTIONS EVER FREE?
!
39!
.61; n = 393) than participants in the morally good condition (M = 3.96, SD = .65; n = 391), F(1,
776) = 4.80, p = .029, η2
p = .006.
6.2.6 Confirmatory analyses. We ran the same twelve regressions as in Study 2a. As can
be seen in Tables 4 (free will attributions) and 5 (free will beliefs), all main effects for the six
evaluation variables on free will attributions and free will beliefs replicated (again, with the
exception of the normative evaluation on free will beliefs). Higher free will judgments were
predicted by higher action magnitude, higher response utility, stronger emotional reaction, lower
normativity, higher willpower, and stronger punish/reward response. Most of these were again
qualified by interactions with the moral valence condition.
6.2.6.1 Action magnitude. As in Study 2a, higher action magnitude ratings predicted
higher attributions of free will in both the morally bad and good conditions, but once again were
stronger predictors in the morally good condition (b = .65), t = 10.29, p < .001, than in the
morally bad condition (b = .16), t = 3.56, p < .001. Unlike Study 2a, this effect also emerged for
free will beliefs. People increased their belief in free will as their perceived magnitude of the
morally bad action increased (b = .11), t = 3.42, p = .001, but they did so to a stronger degree as
their perceived magnitude of the morally good action increased (b =.25), t = 8.00, p < .001.
6.2.6.2 Response utility. As in Study 2a, response utility ratings predicted higher
attributions of free will and higher free will beliefs. Unlike Study 2a, the interaction between the
moral valence condition and response utility reached statistical significance for free will
attributions. Perceptions that punishing the immoral actor would deter bad behavior did not
predict free will attributions (b = -.001), t = -0.02, p = .985, but perceptions that rewarding the
morally good actor would encourage good behavior did predict higher free will attributions to the
ARE MORALLY GOOD ACTIONS EVER FREE?
!
40!
morally good actor (b = .12), t = 2.68, p = .007. Once again, the interaction did not emerge for
free will beliefs.
6.2.6.3 Emotional reaction. As in Study 2a, greater moral outrage predicted higher free
will attributions to the morally bad actor (b = .32), t = 10.12, p < .001, but feeling morally
uplifted predicted higher free will attributions to the morally good actor to a significantly
stronger degree (b = .56), t = 12.52, p < .001. Unlike Study 2a, this interaction also emerged for
general free will beliefs (morally bad condition [b = .11], t = 2.45, p = .014; morally good
condition [b = .17], t = 5.25, p < .001).
6.2.6.4 Normative. Unlike Study 2a, higher ratings of the likelihood that other people in
John’s position would have done the same thing predicted lower free will attributions to the
morally bad actor (b = -.10), t = -3.19, p = .001. However, as in Study 2a, this relationship was
stronger in the morally good condition, such that participants attributed less free will to John as
their perceived likelihood that others in John’s position would have done the same thing
increased (b = -.21), t = -4.61, p < .001. As in Study 2a, normativity judgments did not predict
free will beliefs in either condition.
6.2.6.5 Willpower. There were again no relationships between ratings of John’s
willpower and free will attributions (b = .00) or beliefs (b = .01) in the morally bad condition. In
the morally good condition, participants who felt that John exerted more willpower attributed
him more free will (b = .37), t = 8.27, p < .001, and believed more in free will generally (b =
.15), t = 4.71, p < .001.
6.2.6.6 Punish/reward response. As in Study 2a, both higher punishment
recommendations (b = .27), t = 4.27, p < .001, and higher reward recommendations (b = .47), t =
7.43, p < .001, predicted higher attributions of free will. Unlike Study 2a, the relationship was
ARE MORALLY GOOD ACTIONS EVER FREE?
!
41!
significantly stronger in the morally good condition. Once again, this interaction did not emerge
for free will beliefs.
Table 4
Free Will Attributions Regressed on Moral Valence Condition, Evaluation Variables, and Interactions
in Study 2b (Confirmatory Analyses)
F
R2
b
SE
t
p
95% CIs
Free Will Attributions (Model)
47.44
.15
<.001
Moral valence condition
.60
.08
7.42
<.001
.44, .76
Action magnitude
.65
.06
10.73
<.001
.53, .77
Interaction
-.49
.08
-6.44
<.001
-.64, -.34
Free Will Attributions (Model)
7.88
.03
<.001
Moral valence condition
.33
.09
4.05
<.001
.17, .50
Response utility
.12
.04
3.19
.001
.05, .19
Interaction
-.12
.05
-2.42
.016
-.22, -.02
Free Will Attributions (Model)
51.41
.17
<.001
Moral valence condition
.42
.08
5.56
<.001
.27, .57
Emotional Reaction
.46
.04
10.33
<.001
.37, .54
Interaction
-.24
.06
-4.10
<.001
-.37, -.13
Free Will Attributions (Model)
14.46
.05
<.001
Moral valence condition
.37
.08
4.54
<.001
.21, .53
Normative
-.21
.04
-4.77
<.001
-.29, -.12
Interaction
.11
.06
1.83
.068
-.01, .22
Free Will Attributions (Model)
32.16
.09
<.001
Moral valence condition
.59
.09
6.67
<.001
.41, .76
Willpower
.37
.05
8.15
<.001
.28, .45
Interaction
-.37
.05
-7.40
<.001
-.47, -.27
Free Will Attributions (Model)
27.44
.10
<.001
Moral valence condition
.30
.08
3.80
<.001
.14, .45
Punish/reward response
.47
.07
7.08
<.001
.34, .60
Interaction
-.20
.09
-2.19
.029
-.38, -.02
ARE MORALLY GOOD ACTIONS EVER FREE?
!
42!
Table 5
Free Will Beliefs Regressed on Moral Valence Condition, Evaluation Variables, and Interactions in
Study 2b (Confirmatory Analyses)
F
R2
b
SE
t
p
95% CIs
Free Will Belief (Model)
24.90
.09
<.001
Moral valence condition
.23
.05
5.06
<.001
.14, .32
Action magnitude
.25
.04
7.27
<.001
.18, .32
Interaction
-.15
.04
-3.31
.001
-.23, -.06
Free Will Belief (Model)
9.90
.04
<.001
Moral valence condition
.14
.05
3.16
.002
.05, .23
Response utility
.06
.02
3.15
.002
.02, .10
Interaction
.01
.03
0.27
.789
-.05, .06
Free Will Belief (Model)
23.63
.08
<.001
Moral valence condition
.15
.04
3.43
.001
.06, .24
Emotional reaction
.17
.03
6.49
<.001
.12, .22
Interaction
-.06
.03
-1.72
.086
-.13, .01
Free Will Belief (Model)
2.22
.01
.008
Moral valence condition
.11
.05
2.33
.020
.02, .20
Normative
.00
.02
-0.08
.936
-.05, .05
Interaction
-.03
.03
-0.83
.410
-.09, .04
Free Will Belief (Model)
13.73
.05
<.001
Moral valence condition
.23
.05
4.66
<.001
.13, .33
Willpower
.15
.03
5.90
<.001
.10, .20
Interaction
-.14
.03
-4.74
<.001
-.19, -.08
Free Will Belief (Model)
11.34
.04
<.001
Moral valence condition
.10
.04
2.24
.025
.01, .19
Punish/reward response
.10
.04
2.63
.009
.03, .17
Interaction
.07
.05
1.35
.178
-.03, .17
!
ARE MORALLY GOOD ACTIONS EVER FREE?
!
43!
!
Table 6
Summary of the relationships between each Evaluation Variable and Free Will Attributions and
Beliefs within each Moral Condition and indicators of which (if either) condition had a significantly
stronger relationship in Studies 2A (exploratory) and 2B (confirmatory)
Study
Evaluation
Outcome
Condition
Direction
Significant
Stronger
Variable
Condition
Study 2a
Action
Free Will
Morally Good
+
*
X
Magnitude
Attributions
Morally Bad
+
*
Free Will
Morally Good
+
*
Beliefs
Morally Bad
+
*
Study 2b
Action
Free Will
Morally Good
+
*
X
Magnitude
Attributions
Morally Bad
+
*
Free Will
Morally Good
+
*
X
Beliefs
Morally Bad
+
*
Study 2a
Response
Free Will
Morally Good
+
*
Utility
Attributions
Morally Bad
+
*
Free Will
Morally Good
+
*
Beliefs
Morally Bad
+
*
Study 2b
Response
Free Will
Morally Good
+
*
X
Utility
Attributions
Morally Bad
0
Free Will
Morally Good
+
*
Beliefs
Morally Bad
+
*
Study 2a
Emotional
Free Will
Morally Good
+
*
X
Reaction
Attributions
Morally Bad
+
*
Free Will
Morally Good
+
*
Beliefs
Morally Bad
+
*
Study 2b
Emotional
Free Will
Morally Good
+
*
X
Reaction
Attributions
Morally Bad
+
*
Free Will
Morally Good
+
*
X
Beliefs
Morally Bad
+
*
Study 2a
Normativity
Free Will
Morally Good
-
*
X
Attributions
Morally Bad
0
Free Will
Morally Good
+
Beliefs
Morally Bad
+
Study 2b
Normativity
Free Will
Morally Good
-
*
X(marg)
Attributions
Morally Bad
-
*
Free Will
Morally Good
0
Beliefs
Morally Bad
0
ARE MORALLY GOOD ACTIONS EVER FREE?
!
44!
Study 2a
Willpower
Free Will
Morally Good
+
*
X
Attributions
Morally Bad
+
*(marg)
Free Will
Morally Good
+
*
X(marg)
Beliefs
Morally Bad
+
Study 2b
Willpower
Free Will
Morally Good
+
*
X
Attributions
Morally Bad
0
Free Will
Morally Good
+
*
X
Beliefs
Morally Bad
0
Study 2a
Punish/Reward
Free Will
Morally Good
+
*
Response
Attributions
Morally Bad
+
*
Free Will
Morally Good
+
*
Beliefs
Morally Bad
+
*
Study 2b
Punish/Reward
Free Will
Morally Good
+
*
X
Response
Attributions
Morally Bad
+
*
Free Will
Morally Good
+
*
Beliefs
Morally Bad
+
*
Note. Direction indicates direction of relationship (b .01 considered 0). Significant indicates
whether the relationship is statistically significant at p < .05 (marg = p < .10). An X in the
Stronger Condition column indicates that there was a significantly stronger relationship between
the evaluation variable and the free will judgment in that condition than the other.
!
!
6.3 Discussion
Study 2b replicated the effects of Study 2a with an increased sample size and picked up
similar patterns of results across more of the evaluation variables and free will beliefs (see Table
6 for a summary of results of both Studies 2a and 2b). Specifically, free will judgments for
morally bad actions were predominantly predicted by affective retributive responses. The
strongest predictors of free will judgments for morally bad actions were desires to punish, moral
outrage, and perceived severity of the morally bad act. There was some indication that higher
response utility ratings and lower normativity ratings predicted free will judgments for morally
bad actions, but these relationships were generally weak and inconsistent across the two studies.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
45!
Across both studies, all of the six evaluation variables tended to better predict free will
judgments within the morally good condition than the morally bad condition. Ratings of the
generosity of the morally good act, stronger reward motives, and feelings of being morally
uplifted were all stronger predictors in the morally good condition than their counterparts in the
morally bad condition. Furthermore, perceived response utility, perceived normativity, and
perceived willpower consistently predicted free will judgments in the morally good condition
only. Thus free will attributions to morally positive actions were more sensitive to all relevant
evaluations of the act including the affectively-charged evaluations (deservingness, desires to
reward, emotional reaction), and the less affectively-charged evaluations of the act (utility,
normativity, and willpower).
When exposed to morally bad actions, participants seemed to experience simple and
strong punitive responses, which led them to perceive such actions as more freely performed. For
morally good actions, a variety of features seemed to influence free will judgments, some
affectively charged and some not. Of course, this specific set of findings was not predicted at the
outset of this project. However, they are consistent with our general contention that
responsibility judgments for morally bad actions are driven primarily by affective retributive
responses, whereas free will judgments for morally good actions are more sensitive to a variety
of relevant features. Furthermore, that Study 2b very closely replicated the results of Study 2a
increases confidence that the present results are real and replicable.
7. General Discussion
A rather large body of work has demonstrated that people attribute more responsibility to
bad actions than good and neutral ones. This created the general impression that people are only
motivated to attribute responsibility to harmful actions, and not helpful ones. The present
ARE MORALLY GOOD ACTIONS EVER FREE?
!
46!
findings have upheld that people ascribe the most responsibility to bad actions — but they also
show that people ascribe responsibility to good actions. In particular, people see morally good
actions as more freely performed than morally neutral ones.
Studies 2a and 2b uncovered a number of factors that help distinguish the processes by
which people come to ascribe responsibility for morally negative and morally positive actions.
Specifically, free will judgments for morally bad actions were predominantly influenced by
affective punitive motives (i.e., desires to punish, moral outrage, and perceived severity of the
act). These results are consistent with prior work demonstrating that people punish retributively
with relatively little concern for the deterrent benefits of punishment. The negative emotional
experience of anger may crowd out other more rational or pragmatic considerations. In contrast,
free will ascriptions for morally positive actions were sensitive to all measured evaluations,
including those relevant to affective reward responses (e.g., desires to reward, feelings of being
morally uplifted, and perceived generosity of the act), but also less affectively-charged
considerations (the likely utility of reward, normativity ratings, and perceived willpower).
The positive emotional response of feeling morally uplifted may expand awareness and allow for
more careful consideration of these less affectively-charged components.
Though affective punitive motives were the strongest predictors of free will ascriptions to
the morally bad actor, the relationships between the emotional response items and free will
judgments were actually weaker in the morally bad condition than the morally good condition.
These results might suggest that even relatively low levels of affective punitive feelings can
increase free will judgments for morally bad actions. These results may help explain why bad
actions are generally seen as freer than good ones. People may have a relatively low threshold
for ascribing free will to immoral actors. When exposed to morally bad actions, strong punitive
ARE MORALLY GOOD ACTIONS EVER FREE?
!
47!
motives may overwhelm one’s ability or willingness to consider the nuances of a particular
situation. Because people are highly sensitive to harmful acts, contextual features such as
whether punishment serves any deterrent purpose may not be considered. In the morally good
condition, all measured evaluations (including affective responses) predicted higher free will
judgments to an even stronger degree, suggesting that higher levels of these reactions may be
required to perceive morally good actions as free. In other words, people might have stricter
criteria for ascribing free will to morally positive actions, leading to generally weaker tendencies
to see such actions as freely performed (as compared with morally bad acts).
7.1 Affect vs. Utility
Ultimately, humans are inclined to punish because punishment deters harmful behavior
(e.g., Fehr & Gächter, 2002; Henrich et al., 2006). Proximately, however, people are inclined to
punish to satisfy desires for retribution, and people do not demonstrate much concern for
whether a particular punishment would deter bad behavior (e.g., Carlsmith et al., 2002; Crockett
et al., 2014). The present results suggest that ascriptions of responsibility for morally bad actions
are similarly driven by affective desires for punishment. To our knowledge, this distinction
between affective and utilitarian motives for punitive behavior has not been made for reward
behavior. The results of Studies 2a and 2b suggest the possibility that people may ascribe
responsibility to morally good actions due to both affective desires to reward, but also more
pragmatic considerations such as whether rewards will encourage future good behavior, and
other contextual features of the act such as how counternormative the morally good behavior was
and how much willpower was required to perform the morally good action.
How much are people’s responses to moral actions influenced by the utility of reward or
punishment? Studies 2a and 2b manipulated the utility explicitly. The utility manipulation was
ARE MORALLY GOOD ACTIONS EVER FREE?
!
48!
successful, and, consistent with previous findings, did not influence punishment decisions. In
contrast, utility did produce some significant effects on reward decisions, albeit inconsistently. In
Study 2a, higher utility of reward increased desires to reward only in the high deservingness
condition. In Study 2b, higher utility of reward increased desires to reward in both the high and
low deservingness conditions. Furthermore, the utility manipulation had no main or interactive
effect on free will judgments. The regression analyses, which included participants’ perceptions
of the utility of punishment and reward, provided further suggestive evidence that free will
judgments for morally positive actions are sensitive to utilitarian concerns (i.e., perceived utility
of reward predicted higher free will attributions and beliefs in both Studies 2a and 2b), but
perceived utility of punishment also predicted higher free will judgments for morally negative
actions to some extent (though this effect was generally weaker and did not emerge for free will
attributions in Study 2b). Future work might employ different manipulations of utility to try to
replicate the present results experimentally.
7.2 Why Rewards May be More Sensitive to Context
Selfish behavior within a society can lead to collective disaster for the entire social group
(e.g., Kerr, 1983), thus reducing harmful actions is the top priority for benefiting the self and
society. It is not surprising then, that exposure to morally bad actions elicits strong desires to
punish and commands the strongest reactions. Rewarding good actions is less urgent, but also
has value (Fehr et al., 1997). Thus, as the present results demonstrate, people are sensitive to
morally good actions as well. The greater context-sensitivity of responses to morally good than
morally bad actions may reflect the greater need to discriminate. Punishing a harmful action may
be useful regardless of whether the action was counternormative: People are inclined to behave
in normative ways (e.g., Goldstein, Cialdini, & Griskevicius, 2008; Nolan, Schultz, Cialdini,
ARE MORALLY GOOD ACTIONS EVER FREE?
!
49!
Goldstein, & Griskevicius, 2008), and so punishment may be needed to deter harmful normative
behavior. In contrast, reward may not be needed to encourage helpful normative behavior. For
example, a person might regularly speed on the highway because most others do so, but not if he
or she sees a police car and fears receiving a ticket. In contrast, a person might hold the door
open for a stranger when entering a building because most others do so, and would do so
regardless of whether he or she received a reward for doing so. Rewards would not increase
normative good actions, because people are inclined to perform them even without rewards (e.g.,
Goldstein et al., 2008; Nolan et al., 2008). Therefore, normativity may matter more when
considering rewards, whereas punishment may be useful regardless of whether an action is
normative.
Similarly, punishing a harmful action may be useful regardless of whether the action was
performed willfully. Punishing even unintended and accidental harmful actions, such as
negligent behavior, may increase vigilance and reduce the risk of careless behavior in the future.
If a person knows that they might be punished for accidentally bringing about a harmful
outcome, they might seek to avoid behaviors that could potentially cause harmful outcomes. In
contrast, it may not be particularly useful to reward unwilled positive actions, such as happy
accidents and positive side-effects. People tend to perform their desired actions regardless of
whether they anticipate unintended positive side-effects, and so rewards might do little to
encourage such behaviors.
Punishment may be helpful for deterring bad behavior and bad outcomes regardless of an
agent’s inclinations, desires, or intentions. In contrast, reward may be helpful for encouraging
good behavior and good outcomes only when the actor would not have pursued the good
behavior otherwise. Thus, careful evaluation of context and motives for morally significant
ARE MORALLY GOOD ACTIONS EVER FREE?
!
50!
actions may be more useful when allocating rewards. Furthermore, because people are generally
cooperative on a day-to-day basis, it may be impractical to reward all good actions. It might
therefore behoove individuals to evaluate good actions and the implications of reward more
carefully. Future work might seek to manipulate normativity and willpower to test the
significance of these variables experimentally.
7.3 The Function of Free Will Beliefs
The present work also speaks to the function of free will beliefs. A large body of work
has demonstrated that belief in free will bolsters feelings of moral responsibility (e.g.,
Baumeister et al., 2009; Protzko et al., 2016; Shariff et al., 2014; Vohs & Schooler, 2008).
Furthermore, people bolster their own belief in free will so as to justify holding others morally
responsible (Clark et al., 2014; 2017). Another analysis suggests that there may be individual
differences such that people who are particularly inclined to see a variety of negative actions and
outcomes (e.g., homelessness, drug addiction) as having moral significance are also more
inclined to believe that people have freedom and control over those actions and outcomes
(Everett et al., 2017). Relatedly, people tend to downplay their own free will when considering
the morally negative consequences of their own behavior (Vonasch, Clark, Lau, Vohs, &
Baumeister, 2017). Such results seem to suggest that free will is largely about moral
responsibility.
However, some recent work suggests that people ascribe free will to individuals to
increase accountability more generally so as to signal that the person is capable of altering their
behavior in future situations (Alquist et al., 2015; Feldman et al., 2016; Feldman &
Chandrashekar, 2017). In other words, people might ascribe free will not only to increase moral
ARE MORALLY GOOD ACTIONS EVER FREE?
!
51!
responsibility, but also to increase accountability across both moral and amoral domains to
maximize positive outcomes in the future.
Given the strength and pervasiveness of free will beliefs, both general free will beliefs
and individual ascriptions of free will likely have many antecedents and many functions. One of
the strongest driving factors is likely the powerful subjective experience of conscious will (e.g.,
Wegner, 2002, 2003). These feelings of personal free will, and the belief that other humans have
free will, both are important if not indispensable parts of human existence and social group
functioning. Further work is needed to fully understand how these beliefs influence individual
behavior, and how these beliefs arise through both motivated cognition and experiential
observation.
7.4 Is Good Stronger than Neutral?
Whereas previous work has found that “bad is freer than good,” the present results
demonstrate that good is freer than neutral. However, it remains unknown whether this pattern of
findings would extend beyond responsibility attributions to the broader phenomenon of “bad is
stronger than good.” Is bad stronger than good, but good stronger than neutral? Perhaps attesting
to the phenomenon that bad is stronger than good, large bodies of work have compared bad
actions and outcomes to 1) good actions and outcomes, 2) neutral actions and outcomes, and 3)
less bad actions and outcomes, but little work compares good and neutral. Future work testing
positive-negative asymmetries should consider including neutral or control conditions to more
fully understand the significance of valence in how people understand and experience the world.
8. Conclusion
Societies dedicate far more resources toward punishing rule breakers than toward
rewarding rule followers. It is also evident that individuals care a great deal more about holding
ARE MORALLY GOOD ACTIONS EVER FREE?
!
52!
others accountable for their harmful behaviors than their helpful ones. This rather obvious moral
asymmetry has led researchers to focus almost exclusively on how individuals judge morally bad
actions at the expense of understanding responsibility judgments for morally good actions.
However, the quality of social life can be improved both by reinforcing positive actions and
discouraging harmful ones. On that basis, one might reasonably expect that people would care
about accountability for positive actions too. The present research confirms that people do care
about responsibility for good deeds: People ascribe more free will to morally positive actions
than neutral ones. Indeed, our participants not only attributed free will to virtuous actions — they
did so in a cognitively sophisticated manner that was sensitive to various specific features of the
action (the normativity and required willpower) and of the reward process (whether rewards
would encourage good behavior). Though people perceive harmful actions as more freely
performed than helpful ones, they still perceive helpful actions as more freely performed than
neutral ones. Our findings augment previous work and correct the false impression that people
only care about assigning responsibility and free choice to bad actions. The condemnation of
morally bad actions is simple and strong, whereas the recognition of morally good actions is
weaker and more sensitive to context.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
53!
References
Alicke, M. D. (1992). Culpable causation. Journal of Personality and Social Psychology, 63,
368-378.
Alicke, M. D. (2000). Culpable control and the psychology of blame. Psychological Bulletin,
126, 556-574.
Alquist, J. L., Ainsworth, S. E., Baumeister, R. F., Daly, M., & Stillman, T. F. (2015). The
making of might-have-beens: Effects of free will belief on counterfactual
thinking. Personality and Social Psychology Bulletin, 41, 268-283.
Angrist, J., & Lavy, V. (2009). The effects of high stakes high school achievement awards:
Evidence from a randomized trial. The American Economic Review, 99, 1384-1414.
Ariely, D., Bracha, A., & Meier, S. (2007). Doing good or doing well? Image motivation and
monetary incentives in behaving prosocially. The American Economic Review, 120, 544-
555.
Bargh, J. A. (2008). Free will is un-natural. In J. Baer, J. C. Kaufman, & R. F. Baumeister (Eds.),
Are we free? Psychology and free will (pp. 128–154). Oxford, England: Oxford
University Press.
Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001). Bad is stronger than
good. Review of General Psychology, 5, 323-370.
Baumeister, R. F., Clark, C. J., & Luguri, J. (2015). Free will: Belief and reality. In A. R. Mele,
(Ed.). Surrounding free will: philosophy, psychology, neuroscience (pp. 49-71). Oxford,
England: Oxford University Press.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
54!
Baumeister, R. F., Masicampo, E. J., & DeWall, C. N. (2009). Prosocial benefits of feeling free:
Disbelief in free will increases aggression and reduces helpfulness. Personality and
Social Psychology Bulletin, 35, 260-268.
Baumeister, R. F., & Newman, L. S. (1994). Self-regulation of cognitive inference and decision
processes. Personality and Social Psychology Bulletin, 20, 3-19.
Braver, S. L., Thoemmes, F. J., & Rosenthal, R. (2014). Continuously cumulating meta-analysis
and replicability. Perspectives on Psychological Science, 9, 333-342.
Cameron, J., Banko, K. M., & Pierce, W. D. (2001). Pervasive negative effects of rewards on
intrinsic motivation: The myth continues. The Behavior Analyst, 24, 1-44.
Card, N. A. (2012). Applied meta-analysis for social science research. New York, NY: Guilford
Carlsmith, K. M. (2006). The roles of retribution and utility in determining punishment. Journal
of Experimental Social Psychology, 42, 437-451.
Carlsmith, K. M. (2008). On justifying punishment: The discrepancy between words and
actions. Social Justice Research, 21, 119-137.
Carlsmith, K. M., Darley, J. M., & Robinson, P. H. (2002). Why do we punish? Deterrence and
just deserts as motives for punishment. Journal of Personality and Social Psychology, 83,
284-299.
Clark, C. J., Bauman, C. W., Kamble, S. V., & Knowles, E. D. (2016). Intentional sin and
accidental virtue? Cultural differences in moral systems influence perceived
intentionality. Social Psychological and Personality Science, 8, 74-82.
Clark, C. J., Baumeister, R. F. & Ditto, P. H. (2017). Making punishment palatable: Belief in
free will alleviates punitive distress. Consciousness and Cognition, 51, 193-211.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
55!
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, 501-513.
Clark, C. J., Chen, E. E. & Ditto, P. H. (2015). Moral coherence processes: Constructing
culpability and consequences. Current Opinion in Psychology, 6, 123-128.
Clutton-Brock, T. H. & Parker, G. A. (1995). Punishment in animal societies. Nature, 373, 209-
216.
Crockett, M. J., Özdemir, Y., & Fehr, E. (2014). The value of vengeance and the demand for
deterrence. Journal of Experimental Psychology: General, 143, 2279-2286.
Croson, R., & Gneezy, U. (2009). Gender differences in preferences. Journal of Economic
Literature, 47, 448-474.
Cushman, F. (2013). The role of learning in punishment, prosociality, and human uniqueness. In
K. Sterelny, R. Joyce, B. Calcott, & B. Fraser (Eds.), Cooperation and its evolution (Life
and mind: Philosophical issues in biology and psychology) (pp. 333–372). Cambridge,
MA: MIT Press.
Cushman, F., Knobe, J., & Sinnott-Armstrong, W. (2008). Moral appraisals affect
doing/allowing judgments. Cognition, 108, 281-289.
Deci, E. L., Koestner, R., & Ryan, R. M. (1999). A meta-analytic review of experiments
examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin,
125, 627-668.
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, 568-584.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
56!
Ditto, P. H., Pizarro, D. A., Tannenbaum, D. (2009). Motivated moral reasoning. In B. H. Ross
(Series Ed.) & D. M. Bartels, C. W. Bauman, L. J. Skitka, & D. L. Medin (Eds.),
Psychology of learning and motivation, Vol. 50: Moral judgment and decision making
(pp. 307-338). San Diego, CA: Academic Press.
Everett, J. A. C., Clark, C. J., Luguri, J. B., Earp, B. D., Ditto, P. H., & Shariff, A. F. (2017).
Political differences in free will belief are driven by differences in moralization.
PsyArXiv. Available at h5ps://osf.io/preprints/psyarxiv/nx9rj.
Fabes, R. A., Fultz, J., Eisenberg, N., May-Plumlee, T., & Christopher, F. S. (1989). Effects of
rewards on children's prosocial motivation: A socialization study. Developmental
Psychology, 25, 509-515.
Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible statistical
power analysis program for the social, behavioral, and biomedical sciences. Behavior
Research Methods, 39, 175-191.
Faul, F., Erdfelder, E., Buchner, A., & Lang, A.-G. (2009). Statistical power analyses using
G*Power 3.1: Tests for correlation and regression analyses. Behavior Research
Methods, 41, 1149-1160.
Fehr, E., & Fischbacher, U. (2004). Third-party punishment and social norms. Evolution and
Human Behavior, 25, 63-87.
Fehr, E., & Gächter, S. (2002). Altruistic punishment in humans. Nature, 415, 137-140.
Fehr, E., & Gächter, S. (2005). Human behaviour: Egalitarian motive and altruistic punishment
(reply). Nature, 433, E1-E2.
Fehr, E., Gächter, S. & Kirchsteiger, G. (1997). Reciprocity as a contract enforcement device:
Experimental evidence. Econometrica, 65, 833-860.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
57!
Feldman, G., & Chandrashekar, S. P. (in press). Laypersons’ beliefs and intuitions about free
will and determinism: New insights linking the social psychology and experimental
philosophy paradigms. Social Psychological and Personality Science.
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, 26-40.
Fredrickson, B. L. (2001). The role of positive emotions in positive psychology. The broaden-
and-build theory of positive emotions. American Psychologist, 56, 218–226.
Fredrickson, B. L. (2013). Positive emotions broaden and build. Advances in Experimental
Social Psychology, 47, 53.
Fredrickson, B. L., & Branigan, C. (2005). Positive emotions broaden the scope of attention and
thought-action repertoires. Cognition & Emotion, 19, 313-332.
Goldstein, N. J., Cialdini, R. B., & Griskevicius, V. (2008). A room with a viewpoint: Using
social norms to motivate environmental conservation in hotels. Journal of Consumer
Research, 35, 472-482.
Gneezy, U., Meier, S., & Rey-Biel, P. (2011). When and why incentives (don't) work to modify
behavior. The Journal of Economic Perspectives, 25, 191-209.
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, 535-549.
Haidt J. (2001). The emotional dog and its rational tail: A social intuitionist approach to moral
judgment. Psychological Review, 108, 814–834.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
58!
Haidt, J. (2003). Elevation and the positive psychology of morality. In C. L. M. Keyes & J. Haidt
(Eds.), Flourishing: Positive psychology and the life well-lived (pp. 275–289).
Washington, DC: American Psychological Association.
Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A
regression-based approach. New York, NY: The Guilford Press.
Hedges, L. V., & Vevea, J. L. (1998). Fixed-and random-effects models in meta-analysis.
Psychological Methods, 3, 486–504.
Henrich, J., Boyd, R., Bowles, S., Camerer, C., Fehr, E., Gintis, H., ... & Tracer, D. (2005).
“Economic man” in cross-cultural perspective: Behavioral experiments in 15 small-scale
societies. Behavioral and Brain Sciences, 28, 795-815.
Henrich, J., Ensminger, J., McElreath, R., Barr, A., Barrett, C., Bolyanatz, A., ……., Ziker, J.
(2010). Markets, religion, community size, and the evolution of fairness and punishment.
Science, 327, 1480-1484.
Henrich, J., McElreath, R., Barr, A., Ensminger, J., Barrett, C., Bolyanatz, A., ... & Ziker, J.
(2006). Costly punishment across human societies. Science, 312, 1767-1770.
Irwin, K. (2009). Prosocial behavior across cultures: The effects of institutional versus
generalized trust. Advances in Group Processes, 26, 165-198.
Isen, A. M., Rosenzweig, A. S., & Young, M. J. (1991). The Influence of Positive Affect on
Clinical Problem-Solving. Medical Decision Making, 11, 221–227.
Kanouse, D. E., & Hanson. L. R. (1971). Negativity in evaluations. In E. E. Jones, D. E.
Kanouse, H. H. Kelley, R. E. Nisbett, S. Valins, & B. Weiner (Eds.) Attribution:
Perceiving the causes of behavior (pp. 47-62). Morristown, NJ: General Learning Press.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
59!
Karau, S. J., & Williams, K. D. (1993). Social loafing: A meta-analytic review and theoretical
integration. Journal of Personality and Social Psychology, 65, 681-706.
Kerr N. L. (1983). Motivation losses in task-performing groups: A social dilemma analysis.
Journal of Personality and Social Psychology, 45, 819-828.
Kerr, N.L., & Bruun, S. (1983). The dispensability of member effort and group motivation
losses: Free-rider effects. Journal of Personality and Social Psychology, 44, 78–94.
Knobe, J. (2003). Intentional action in folk psychology: An experimental investigation.
Philosophical Psychology, 16, 309-324.
Knobe, J., & Fraser, B. (2008). Causal judgment and moral judgment: Two experiments. Moral
Psychology, 2, 441-448.
Kubanek, J., Snyder, L. H., & Abrams, R. A. (2015). Reward and punishment act as distinct
factors in guiding behavior. Cognition, 139, 154-167.
Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108, 480-498.
Lacetera, N., & Macis, M. (2010). Social image concerns and prosocial behavior: Field evidence
from a nonlinear incentive scheme. Journal of Economic Behavior & Organization, 76,
225-237.
Latané B., Williams K., Harkins S. (1979). Many hands make light the work: The causes and
consequences of social loafing. Journal of Personality and Social Psychology, 37, 822-
832.
Lewicka, M., Czapinski, J., & Peeters, G. (1992). Positivenegative asymmetry or ‘When the
heart needs a reason’. European Journal of Social Psychology, 22, 425-434.
Libet, B. (1985). Unconscious cerebral initiative and the role of conscious will in voluntary
action. Behavioral and Brain Sciences, 8, 529-566.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
60!
Liu, B. S., & Ditto, P. H. (2013). What dilemma? Moral evaluation shapes factual belief. Social
Psychological and Personality Science, 4, 316-323.
Maner, J. K. (2014). Let’s put our money where our mouth is: If authors are to change their
ways, reviewers (and editors) must change with them. Perspectives on Psychological
Science, 9, 343-351.
Monroe, A. E., Ainsworth, S. E., Vohs, K. D., & Baumeister, R. F. (2017). Fearing the future?
Future-oriented thought produces aversion to risky investments, trust, and
immorality. Social Cognition, 35, 66-78.
Munro, G. D., & Ditto, P. H. (1997). Biased assimilation, attitude polarization, and affect in
reactions to stereotype-relevant scientific information. Personality and Social Psychology
Bulletin, 23, 636-653.
Nahmias, E., Morris, S., Nadelhoffer, T., & Turner 1, J. (2005). Surveying freedom: Folk
intuitions about free will and moral responsibility. Philosophical Psychology, 18, 561-
584.
Nelissen, R. M., & Zeelenberg, M. (2009). Moral emotions as determinants of third-party
punishment: Anger, guilt and the functions of altruistic sanctions. Judgment and Decision
Making, 4, 543-553.
Nichols, S. (2004). The folk psychology of free will: Fits and starts. Mind and Language, 19,
473-502.
Nichols, S., & Knobe, J. (2007). Moral responsibility and determinism: The cognitive science of
folk intuitions. Nous, 41, 663-685.
Nietzsche, F. (1954). Twilight of the idols (W. Kaufmann, Trans.). New York, NY: Penguin
Books. (Original work published 1889).
ARE MORALLY GOOD ACTIONS EVER FREE?
!
61!
Nolan, J. M., Schultz, P. W., Cialdini, R. B., Goldstein, N. J., & Griskevicius, V. (2008).
Normative social influence is underdetected. Personality and Social Psychology
Bulletin, 34, 913-923.
Orbell, J., & Dawes, R. (1981). Social dilemmas. In G. Stephenson, & J. H. Davis (Eds.),
Progress in applied social psychology (pp.37-65.). London, England: Wiley.
Paulhus, D. L., & Carey, J. M. (2011). The FAD–Plus: Measuring lay beliefs regarding free will
and related constructs. Journal of Personality Assessment, 93, 96-104.
Pohling, R., & Diessner, R. (2016). Moral elevation and moral beauty: A review of the empirical
literature. Review of General Psychology, 20, 412-425.
Protzko, J., Ouimette, B., & Schooler, J. (2016). Believing there is no free will corrupts intuitive
cooperation. Cognition, 151, 6-9.
Rasmussen, E. B., & Newland, M. C. (2008). Asymmetry of reinforcement and punishment in
human choice. Journal of the Experimental Analysis of Behavior, 89, 157-167.
Reeder, R. D., & Spores, J. M. (1983). The attribution of morality. Journal of Personality and
Social Psychology, 44, 736-745.
Rozin, P., & Royzman, E. B. (2001). Negativity bias, negativity dominance, and contagion.
Personality and Social Psychology Review, 5, 296-320.
Sarkissian, H., Chatterjee, A., De Brigard, F., Knobe, J., Nichols, S., & Sirker, S. (2010). Is
belief in free will a cultural universal?. Mind & Language, 25, 346-358.
Scurich, N., & Shniderman, A. (2014). The Selective Allure of Neuroscientific Explanations.
PLoS ONE, 9, 1-6.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
62!
Shariff, A. F., Greene, J. D., Karremans, J. C., Luguri, J. B., Clark, C. J., Schooler, J. W., ... &
Vohs, K. D. (2014). Free Will and Punishment A Mechanistic View of Human Nature
Reduces Retribution. Psychological Science, 25, 1563-1570.
Skowronski, J. J., & Carlston, D. E. (1992). Caught in the act: When impressions based on
highly diagnostic behaviours are resistant to contradiction. European Journal of Social
Psychology, 22, 435-452.
Soon, C. S., Brass, M., Heinze, H. J., & Haynes, J. D. (2008). Unconscious determinants of free
decisions in the human brain. Nature Neuroscience, 11, 543-545.
Szolnoki, A., & Perc, M. (2010). Reward and cooperation in the spatial public goods game. EPL,
92, 38003.
Valentine, J. C., Pigott, T. D., & Rothstein, H. R. (2010). Tutorial: How many studies do you
need? A primer on statistical power for meta-analysis. Journal of Educational and
Behavioral Statistics, 35, 215-247.
Vohs, K. D., & Schooler, J. W. (2008). The value of believing in free will encouraging a belief in
determinism increases cheating. Psychological Science, 19, 49-54.
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, 56-66.
Walster, E. (1966). Assignment of responsibility for an accident. Journal of Personality and
Social Psychology, 3, 73-79.
Warneken, F., & Tomasello, M. (2008). Extrinsic rewards undermine altruistic tendencies in 20-
month-olds. Developmental Psychology, 44, 1785-1788.
Wegner, D.M. (2002). The illusion of conscious will. Cambridge, MA: MIT Press.
ARE MORALLY GOOD ACTIONS EVER FREE?
!
63!
Wegner, D. M. (2003). The mind’s best trick: How we experience conscious will. Trends in
Cognitive Sciences, 7, 65-69.
Wegner, D. M., & Wheatley, T. (1999). Apparent mental causation: Sources of the experience of
will. American Psychologist, 54, 480-492.
... First, like the term soul mate, the term free will is likely not a logically and rigorously defined philosophical term. Rather, the term free will likely describes a feeling (Wegner, 2002) that serves a variety of social functions (e.g., Baumeister & Brewer, 2012;Baumeister, Masicampo & DeWall, 2009;Clark, Baumeister & Ditto, 2017). There are probably many folk theories that vary both between (Feltz & Cokely, 2008;Schulz, Cokely, & Feltz, 2011) and within individuals (Deery et al., 2014), and it is probable that many of them are inconsistent with each other. ...
... Study 2 directly manipulated motives to hold others morally responsible in order to determine whether such motives influence compatibilist responding. Past research has shown that motives to punish (Clark et al., 2014) and praise (Clark, Shniderman, Luguri, Baumeister & Ditto, 2017) underlie free will beliefs. Hence people should be more likely to agree that compatibilist free will is sufficient for moral responsibility when considering concrete morally virtuous and wicked actions than when contemplating abstractly (consistent with Nichols & Knobe, 2007). ...
... Participants were asked to assume the passage was correct while responding to two questions, "If the argument is correct, can people have free will?" and "If the argument is correct, can people be held morally responsible?" on 7-point scales from "Definitely no" to "Definitely yes." Participants were then directed to a second passage adapted from Clark et al. (2017), which argued that free will is unnecessary for moral responsibility, "…recently, philosophers and scientists are realizing that free will is not a requirement for holding people morally responsible… If humans know that bad behavior will be punished, they will be less likely to perform that bad behavior. Holding people morally responsible is not an issue of free will; it is merely a means of discouraging behavior destructive to society." ...
Article
See updated full text here: https://www.researchgate.net/publication/319944911_Forget_the_Folk_Moral_Responsibility_Preservation_Motives_and_Other_Conditions_for_Compatibilism
... A complementary line of research has shown that the belief in free will is also predictive of positive social outcomes. Free will beliefs have been associated with higher honesty (Vohs & Schooler, 2008), higher morality (Carey & Paulhus, 2013;Caspar, Vuillaume, Magalhães De Saldanha Da Gama, & Cleeremans, 2017;Clark, Shniderman, Baumeister, Luguri, & Ditto, 2017;Feldman, Wong, & Baumeister, 2016), less prejudice (Zhao, Liu, Zhang, Shi, & Huang, 2014), more prosocial behavior and less aggression (Baumeister, Masicampo, & DeWall, 2009), better learning from and more guilt over own misdeeds , more cooperation (Protzko, Ouimette, & Schooler, 2016), less objectification of others (Baldissarri, Andrighetto, Gabbiadini, & Volpato, 2016), and higher gratitude (MacKenzie, Vohs, & Baumeister, 2014). It would seem that there is overwhelming evidence for free will beliefs as predicting positive outcomes both for the self and for social behavior. ...
Article
Full-text available
In three studies we examined the relationship between free will beliefs and job satisfaction over time and across cultures. Study 1 examined 252 Taiwanese real-estate agents over a three-months period. Study 2 examined job satisfaction for 137 American workers on an online labor market over a six-months period. Study 3 extended to a large sample of 14,062 employees from 16 countries and examined country-level moderators. We found a consistent positive relationship between the belief in free will and job satisfaction. The relationship was above and beyond other agency constructs (Study 2), mediated by perceived autonomy (Studies 2-3), and stronger in countries with a higher national endorsement of the belief in free will (Study 3). We conclude that free-will beliefs predict outcomes over time and across cultures beyond other agency constructs. We call for more cross-cultural and longitudinal studies examining free-will beliefs as predictors of real-life outcomes.
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
Full-text available
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.
Article
Full-text available
We linked between the social-psychology and experimental-philosophy paradigms for the study of folk intuitions and beliefs regarding the concept of free will to answer three questions: (1) what intuitions do people have about free-will and determinism? (2) do free will beliefs predict differences in free-will and determinism intuitions? and (3) is there more to free-will and determinism than experiencing certainty or uncertainty about the nature of the universe? Overall, laypersons viewed the universe as allowing for human indeterminism, and they did so with certainty. Examining intuitions of prosociality, future-orientation, learning, meaningfulness, human uniqueness, and well-being, ratings were highest in the indeterministic universe condition and lowest in the deterministic universe condition, both significantly different from the uncertain universe condition. Participants' free will beliefs had only weak impact on realism, happiness, and learning intuitions, but did not reverse the general intuition favoring indeterminism, and showed no impact on other intuitions.
Article
Full-text available
Introduction: It is widely believed that addiction entails a loss of free will, even though this point is controversial among scholars. There is arguably a downside to this belief, in that addicts who believe they lack the free will to quit an addiction might therefore fail to quit an addiction. Methods: A correlational study tested the relationship between belief in free will and addiction. Follow-up studies tested steps of a potential mechanism: 1) people think drugs undermine free will 2) people believe addiction undermines free will more when doing so serves the self 3) disbelief in free will leads people to perceive various temptations as more addictive. Results: People with lower belief in free will were more likely to have a history of addiction to alcohol and other drugs, and also less likely to have successfully quit alcohol. People believe that drugs undermine free will, and they use this belief to self-servingly attribute less free will to their bad actions than to good ones. Low belief in free will also increases perceptions that things are addictive. Conclusions: Addiction is widely seen as loss of free will. The belief can be used in self-serving ways that may undermine people's efforts to quit.
Article
Full-text available
Moral elevation is defined as the emotional response to witnessing acts of moral beauty. Studies have found that elevation entails pleasant feelings of warmth in the chest, feeling uplifted, moved, and optimistic about humanity. Elevation motivates affiliation with others as well as moral action tendencies. The main goal of this review was to gather and organize the empirical findings from the last 16 years of elevation research with regard to psychological and physiological characteristics, motivational tendencies, behavioral outcomes, neuronal mechanisms, moderators, and correlates of elevation. A secondary goal was to examine whether elevation is congruent with Fredrickson’s (2001) broaden-and-build theory of positive emotions. It was concluded that there is strong evidence that elevation broadens the thought-action repertoire and relatively weak evidence that it builds lasting resources. Potential evolutionary functions, the forms of measurement of elevation, the process of how elevation is triggered, practical applications and directions for future research were also addressed.
Article
Full-text available
Indians and U.S. Americans view harmful actions as morally wrong, but Indians are more likely than U.S. Americans to perceive helping behaviors as moral imperatives. We utilize this cultural variability in moral belief systems to test whether and how moral considerations influence perceptions of intentionality (as suggested by theories of folk psychology). Four experiments found that Indians attribute more intentionality than U.S. Americans for helpful but not harmful (Studies 1–4) or neutral side effects (Studies 2 and 3). Also, cross-cultural differences in intentionality judgments for positive actions reflect stronger praise motives (Study 3), and stronger devotion to religious beliefs and practices among Hindus (Study 4). These results provide the first direct support for the claim that features of moral belief systems influence folk psychology, and further suggest that the influence is not inherently asymmetrical; motivation to either blame or praise can influence judgments of intentionality.
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
Punishing wrongdoers is beneficial for group functioning, but can harm individual well-being. Building on research demonstrating that punitive motives underlie free will beliefs, we propose that free will beliefs help justify punitive impulses, thus alleviating the associated distress. In Study 1, trait-level punitiveness predicted heightened levels of anxiety only for free will skeptics. Study 2 found that higher state-level incarceration rates predicted higher mental health issue rates, only in states with citizens relatively skeptical about free will. In Study 3, participants who punished an unfair partner experienced greater distress than non-punishers, only when their partner did not have free choice. Studies 4 and 5 confirmed experimentally that punitive desires led to greater anxiety only when free will beliefs were undermined by an anti-free will argument. These results suggest that believing in free will permits holding immoral actors morally responsible, thus justifying punishment with diminished negative psychological consequences for punishers.
Article
Three experiments tested the behavioral effects of prospection on risk tolerance, trust, and moral judgment. While previous research shows that people generally hold positive beliefs about their futures, our analysis demonstrates that actively thinking about one's future engenders conservative behaviors to avoid potential losses or harm. In Experiment 1, writing about one's future self, compared to one's present self, caused people to favor low-risk, low-payoff investments rather than high-risk, high-payoff ones. In Experiment 2, restating sentences about the future (as opposed to the present) reduced behavioral trust. Participants shared less money with various partners in a potentially lucrative investment game. That same manipulation of prospection in Experiment 3 increased blame for misdeeds - while having no effect on praise for virtuous action. These finding suggest that, although people often hold optimistic beliefs about the future, prospection may lead people to behave cautiously as if wary about what could go wrong.
Article
We outline the need to, and provide a guide on how to, conduct a meta-analysis on one's own studies within a manuscript. Although conducting a “mini meta” within one's manuscript has been argued for in the past, this practice is still relatively rare and adoption is slow. We believe two deterrents are responsible. First, researchers may not think that it is legitimate to do a meta-analysis on a small number of studies. Second, researchers may think a meta-analysis is too complicated to do without expert knowledge or guidance. We dispel these two misconceptions by (1) offering arguments on why researchers should be encouraged to do mini metas, (2) citing previous articles that have conducted such analyses to good effect, and (3) providing a user-friendly guide on calculating some meta-analytic procedures that are appropriate when there are only a few studies. We provide formulas for calculating effect sizes and converting effect sizes from one metric to another (e.g., from Cohen's d to r), as well as annotated Excel spreadsheets and a step-by-step guide on how to conduct a simple meta-analysis. A series of related studies can be strengthened and better understood if accompanied by a mini meta-analysis.