DataPDF Available

Are morally good actions ever free? Supplemental Materials

Are Morally Good Actions Ever Free? Supplementary Materials
Initiative for Open Science Statement
No analyses were performed before the corresponding data collection was complete. The only
participant exclusions were those described in Study 1a. The attention checks that resulted in
those participant exclusions were not included in Studies 1b, 2a, and 2b simply because a
different experimenter designed those studies. The results of Study 1a without participant
exclusions are presented in Footnote 2. With the exception of demographic variables, suspicion
probes, and other potential identifiers (e.g., ip address), no other data were collected that are not
described in the paper. There were no additional undisclosed manipulations or conditions. All
datasets (with certain demographics, suspicion probes, and other potential identifiers removed)
and syntax files are publically available. Study 1a data are only available with participant
exclusions because the original data without participant exclusions were lost after the
experimenter of Study 1a switched institutions.
Two other studies with methods very similar to Study 1b are not reported in the paper. The main
difference was that in both of these unreported studies, all participants read an anti-free will
argument before receiving the morality manipulation (this anti-free will argument was not
included in Study 1b). Upon receiving feedback from colleagues that the inclusion of the anti-
free will argument obscured the meaning of these results, we decided to retain only Study 1b,
which had a superior experimental design. Nonetheless, the methods and results of these studies
are reported at the end of this supplementary file under heading Unreported Studies X and Y. The
results of both of these studies are nearly identical to those of Study 1b, with one exception: in
Unreported Study Y, higher attributions of free will in the morally bad condition relative to the
morally neutral condition were not significantly mediated by higher desires to blame (though, the
morally bad condition did elicit higher desires to blame and higher desires to blame were
associated with higher attributions of free will). We discuss this in more detail in the methods
and results below.
Study 1a was also conducted as an improvement upon one study that is not reported in the paper.
Unfortunately, this study was conducted several years ago at the experimenter’s former
institution and the exact methods were not retained, nor were the data. We are reluctant to
speculate about the methodological differences and the pattern of results based off of our
memories alone.
Study 1a Materials
Free Will Subscale of the Free Will and Determinism scale (FAD-Plus; Paulhus & Carey,
2011)
1. People have complete control over the decisions they make.
2. People must take full responsibility for any bad choices they make.
3. People can overcome any obstacles if they truly want to.
4. Criminals are totally responsible for the bad things they do.
5. People have complete free will.
6. People are always at fault for their bad behavior.
7. Strength of mind can always overcome the body's desires.
5-point scales from Strongly disagree to Strongly agree
Morally Bad Condition:
Morally Neutral Condition:
Morally Good Condition:
Questions:
To what extent do you agree with the outcome of the study reported in the article? (DV1)
7-point scale from Not Very Strongly to Very Strongly
How persuasive was the neuroscience reported in the study? (DV2)
7-point scale from Not Very Persuasive to Very Persuasive
How much does the study change your attitude about free will? (DV3)
7-point scale from Not At All to Completely
How much do you want to read more about the study mentioned in the news article? (DV4)
7-point scale from Not At All to Extremely
Do you believe the results of the study are likely to be replicable in the future? (DV5)
7-point scale from Very Unlikely to Very Likely
Was the reporting in the article fair and impartial (objective)? (DV6)
7-point scale from Not Very Objective to Very Objective
How scientific was the study described in the article? (DV7)
7-point scale from Not Very Scientific to Very Scientific
How valid are the results of the study described in the article? (DV8)
7-point scale from Not At All Valid to Completely Valid
How likely is it that most scientists would agree with the study’s conclusions? (DV9)
7-point scale from Very Unlikely to Very Likely
This type of research should receive more funding. (DV10)
Neuroscience is relevant to social policy. (DV11)
7-point scales from Strongly disagree to Strongly agree
Attention Check:
To ensure the survey is working properly, please select "strongly agree.”
7-point scale from Strongly disagree to Strongly agree
Comprehension Checks:
To make sure that you remember the facts of this article, please respond to the following
questions. Mark the response that applies best to the article you have just read.
The researchers used what kind of science in the study reported in the article?
Physiological science
Econometrics Studies
Neuroimmunology
Neuroscience
Psychological Science
The subject of the article was:
Prison
Free Will
Violent Video Games
The Death Penalty
Physiology
Study 1b Materials
Morally Good Condition:
On November 3, 2013, a contributor donated five laser devices to the pediatric cancer center in
Charleston Memorial Hospital in South Carolina. These lasers cost $100,000 each and are used
for endoscopic neurosurgery to remove brain tumors. These lasers help improve the quality of
treatment and speed recovery time, saving the lives of many child patients.
Questions in Morally Good Condition:
To what extent were the contributor's actions freely chosen?
To what extent could the contributor have made other choices?
To what extent did the contributor exercise their own free will in choosing to donate the lasers?
To what extent is the contributor morally praiseworthy for donating the lasers?
7-point scales from Not at all to Very much so
Morally Neutral Condition:
On November 3, 2013, a hospital administrator approved a budget to purchase five laser devices
for the pediatric cancer center in Charleston Memorial Hospital in South Carolina. These lasers
cost $100,000 each and are used for endoscopic neurosurgery to remove brain tumors. These
lasers help improve the quality of treatment and speed recovery time, saving the lives of many
child patients.
Questions in Morally Neutral Condition:
To what extent were the administrator's actions freely chosen?
To what extent could the administrator have made other choices?
To what extent did the administrator exercise their own free will in choosing to approve the
budget?
To what extent is the administrator morally praiseworthy for approving the budget?
To what extent is the administrator morally blameworthy for approving the budget?
7-point scales from Not at all to Very much so
Morally Bad Condition:
On November 3, 2013, a thief stole five laser devices from the pediatric cancer center in
Charleston Memorial Hospital in South Carolina. These lasers cost $100,000 each and were used
for endoscopic neurosurgery to remove brain tumors. These lasers helped improve the quality of
treatment and speed recovery time, saving the lives of many child patients.
Questions in Morally Bad Condition:
To what extent were the thief's actions freely chosen?
To what extent could the thief have made other choices?
To what extent did the thief exercise their own free will in choosing to steal the lasers?
To what extent is the thief morally blameworthy for stealing the lasers?
7-point scales from Not at all to Very much so
Free Will, Fatalistic Determinism, and Scientific Determinism Subscales of the Free Will
and Determinism scale (FAD-Plus; Paulhus & Carey, 2011)
For each statement below, indicate how much you agree or disagree.
FAD01 1. I believe that the future has already been determined by fate.
FAD02 2. People’s biological makeup determines their talents and personality.
FAD04 3. People have complete control over the decisions they make.
FAD05 4. No matter how hard you try, you can’t change your destiny.
FAD06 5. Psychologists and psychiatrists will eventually figure out all human behavior.
FAD08 6. People must take full responsibility for any bad choices they make.
FAD09 7. Fate already has a plan for everyone.
FAD10 8. Your genes determine your future.
FAD12 9. People can overcome any obstacles if they truly want to.
FAD13 10. Whatever will be, will be – there’s not much you can do about it.
FAD14 11. Science has shown how your past environment created your current
intelligence and personality.
FAD16 12. Criminals are totally responsible for the bad things they do.
FAD17 13. Whether people like it or not, mysterious forces seem to move their lives.
FAD18 14. As with other animals, human behavior always follows the laws of nature.
FAD21 15. People have complete free will.
FAD22 16. Parents' character will determine the character of their children.
FAD23 17. People are always at fault for their bad behavior.
FAD24 18. Childhood environment will determine your success as an adult.
FAD26 19. Strength of mind can always overcome the body's desires.
5-point scales from Strongly disagree to Strongly agree
Science: (FAD02 + FAD06 + FAD10+ FAD14 + FAD18 + FAD22 + FAD24)/7
Fate: (FAD01 + FAD05 + FAD09 + FAD13 + FAD17)/5
Free will: (FAD04 + FAD08 + FAD12 + FAD16 + FAD21 + FAD23 + FAD26)/7
Studies 2a and 2b Materials
Morally Bad, High Deserving, High Utility Condition:
John is a district manager for a large department store corporation that employs nearly 200,000
people nationwide. His district includes six stores, and so he is responsible for the jobs of nearly
720 people. A couple years ago he began falsely reporting earnings and managed to steal over
$9.2 million dollars unnoticed. As a result, three of the stores he managed closed and others had
to cut back employees, causing hundreds of people to lose their jobs. Eventually, he was caught,
and soon he will be charged in court.
The CEO of the company decided to report John's actions and any criminal charges John
receives to the rest of the company and the media so others can learn from his lesson and to
ensure this never happens again.
!
Morally Bad, High Deserving, Low Utility Condition:
John is a district manager for a large department store corporation that employs nearly 200,000
people nationwide. His district includes six stores, and so he is responsible for the jobs of nearly
720 people. A couple years ago he began falsely reporting earnings and managed to steal over
$9.2 million dollars unnoticed. As a result, three of the stores he managed closed and others had
to cut back employees, causing hundreds of people to lose their jobs. Eventually, he was caught,
and soon he will be charged in court.
The CEO of the company decided to keep John's actions and any criminal charges John receives
hidden from the rest of the company and the media.
!
Morally Bad, Low Deserving, High Utility Condition:
John is a district manager for a large department store corporation that employs nearly 200,000
people nationwide. His district includes six stores, and so he is responsible for the jobs of nearly
720 people. A couple years ago he began falsely reporting earnings and managed to steal over
$24,000 unnoticed. Eventually, he was caught, and soon he will be charged in court.
The CEO of the company decided to report John's actions and any criminal charges John
receives to the rest of the company and the media so others can learn from his lesson and to
ensure this never happens again.
Morally Bad, Low Deserving, Low Utility Condition:
John is a district manager for a large department store corporation that employs nearly 200,000
people nationwide. His district includes six stores, and so he is responsible for the jobs of nearly
720 people. A couple years ago he began falsely reporting earnings and managed to steal over
$24,000 unnoticed. Eventually, he was caught, and soon he will be charged in court.
The CEO of the company decided to keep John's actions and any criminal charges John receives
hidden from the rest of the company and the media.
Questions in Bad Conditions:
To what extent were John's actions freely chosen?
To what extent could John have made other choices?
To what extent did John exercise his own free will in stealing from his company?
9-point scales from Not at all to Completely
Relative to typical, every day sorts of actions, to what extent were John's actions more or less
free? (Study 2b only)
9-point scale from Much less free to Much more free
How severely should John be punished?
7-point scale, Not at all severely to Extremely severely
What would be an appropriate prison sentence for John?
No prison (1)
A few days (2)
A few weeks (3)
A few months (4)
3 to 6 months (5)
6 months to 1 year (6)
1 to 3 years (7)
3 to 5 years (8)
5 to 8 years (9)
8 to 12 years (10)
12 to 20 years (11)
20 to 30 years (12)
30 to 50 years (13)
50 years to life (14)
How serious was John's crime?
How likely is it that John's punishment will deter others from doing the same thing?
How morally outraged are you by John's behavior?
7-point scales from Not at all to Extremely
How likely is it that someone else in John's position would have done the same thing?
7-point scale from Not at all likely to Extremely likely
How much did John use his willpower in stealing from his company?
7-point scale from Not at all to Completely
Morally Good, High Deserving, High Utility Condition:
John is a district manager for a large department store corporation that employs nearly 200,000
people nationwide. His district includes six stores, and so he is responsible for the jobs of nearly
720 people. A couple years ago he began sacrificing his own annual bonus to implement an
employee incentive program that has drastically increased productivity and employee morale.
His district successfully increased earnings over $9.2 million dollars, and as a result, was able to
hire hundreds of new employees.
Eventually, the CEO of the company took notice and decided to reward him a bonus. The CEO
also decided to report John’s actions and his reward to the rest of the company and the media to
promote this sort of behavior and encourage others to do the same.
Morally Good, High Deserving, Low Utility Condition:
John is a district manager for a large department store corporation that employs nearly 200,000
people nationwide. His district includes six stores, and so he is responsible for the jobs of nearly
720 people. A couple years ago he began sacrificing his own annual bonus to implement an
employee incentive program that has drastically increased productivity and employee morale.
His district successfully increased earnings over $9.2 million dollars, and as a result, was able to
hire hundreds of new employees.
Eventually, the CEO of the company took notice and decided to reward him a bonus, but to keep
John's actions and his reward a secret from the rest of the company and the media.
Morally Good, Low Deserving, High Utility Condition:
John is a district manager for a large department store corporation that employs nearly 200,000
people nationwide. His district includes six stores, and so he is responsible for the jobs of nearly
720 people. A couple years ago he began sacrificing his own annual bonus to implement an
employee incentive program that has increased productivity and employee morale. His district
successfully increased earnings over $24,000 dollars.
Eventually, the CEO of the company took notice and decided to reward him a bonus. The CEO
also decided to report John’s actions and his reward to the rest of the company and the media to
promote this sort of behavior and encourage others to do the same.
Morally Good, Low Deserving, Low Utility Condition:
John is a district manager for a large department store corporation that employs nearly 200,000
people nationwide. His district includes six stores, and so he is responsible for the jobs of nearly
720 people. A couple years ago he began sacrificing his own annual bonus to implement an
employee incentive program t hat has increased productivity and employee morale. His district
successfully increased earnings over $24,000 dollars.
Eventually, the CEO of the company took notice and decided to reward him a bonus, but to keep
John's actions and his reward a secret from the rest of the company and the media.
Questions in Good Conditions:
To what extent were John's actions freely chosen?
To what extent could John have made other choices?
To what extent did John exercise his own free will in sacrificing his own bonus?
9-point scales from Not at all to Completely
Relative to typical, every day sorts of actions, to what extent were John's actions more or less
free? (Study 2b only)
9-point scale from Much less free to Much more free
How highly should John be rewarded?
7-point scale, Not at all to Extremely
What would be an appropriate reward bonus for John?
No bonus (1)
Up to $100 (2)
$100-$250 (3)
$250-$500 (4)
$500-$1,000 (5)
$1,000-$2,500 (6)
$2,500-$5,000 (7)
$5,000-$10,000 (8)
$10,000-$25,000 (9)
$25,000-$50,000 (10)
$50,000-$100,000 (11)
$100,000-$250,000 (12)
$250,000-$500,000 (13)
$500,000 to $1,000,000 (14)
More than $1,000,000 (15)
How generous were John’s actions?
How likely is it that John's reward will encourage others to do the same thing?
How morally uplifted are you by John's behavior?
7-point scales from Not at all to Extremely
How likely is it that someone else in John's position would have done the same thing?
7-point scale from Not at all likely to Extremely likely
How much did John use his willpower in sacrificing his own bonus?
7-point scale from Not at all to Completely
Unreported Studies X and Y
Please see the second paragraph of the Initiative for Open Science Statement (at top of
this supplementary file) for discussion of why these two studies were not included. The methods
of Unreported Studies X and Y were very similar to those of Study 1b. The main difference was
that these two studies included an anti-free will argument prior to the morality manipulation. The
original intention of this methodological strategy was to attempt to lower base-level free will
beliefs in order to provide more room to detect differences between conditions (free will beliefs
tend to be high). Colleagues of the ours criticized this methodological strategy on the grounds
that the inclusion of the anti-free will argument prior to the manipulation obscured the meaning
of our results. Thus, we report the results of these studies only in this supplemental file. The
results demonstrated very similar patterns as Study 1b. Nonetheless, they should be interpreted
with caution as we do not know if or how the anti-free will argument might have altered the
influence of the morality manipulation. The data for these two studies are publicly available
should others find them useful.
Unreported Study X Method
Mturk participants (n = 201) read an argument opposing free will. This argument briefly
reported a series of neuroscientific and social psychological findings arguing against the
existence of free will. Participants were then randomly assigned to read one of the same three
news stories as in Study 1b. Participants then rated the extent to which the actor behaved of their
own free will on the same three items as in Study 1b: whether the action was freely chosen,
whether the actor could have made other choices, and whether the actor exercised their own free
will in choosing to perform the action on 8-point scales from Not at all to Completely.1 To
determine whether the desire to blame and/or praise mediated the influence of the condition on
attributions of free will, participants reported the extent to which the actor was blameworthy or
praiseworthy on a 9-point scale from Very Blameworthy to Very Praiseworthy.2
Unreported Study X Results
An ANOVA revealed a significant effect of condition on attributions of free will, F(2,
198) = 4.72, p = .010. As in Study 1b, Tukey’s post-hoc tests revealed that participants who read
about the morally good behavior (n = 60; M = 6.38, SD = 1.87) and the morally bad behavior (n
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
1 In Study 1b, these questions were answered on 7-point scales from Not at all to Very much so.
2!In Study 1b, participants in the morally good and neutral conditions 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. !
= 63; M = 6.43, SD = 1.83) attributed significantly more free will to the actor than participants
who read about the morally neutral behavior (n = 78; M = 5.56, SD = 2.01), ps = .034 and .021
respectively. As in Study 1b, attributions of free will were slightly higher in the morally bad than
morally good condition, but they did not significantly differ, p = .991. See Table 1S below for a
summary of effect sizes for differences in free will attributions between conditions in Studies 1b,
X, and Y.
An ANOVA revealed a significant effect of condition on desires to blame and praise as
well, F(2, 198) = 31.08, p < .001. Participants rated the morally good actor as the most
praiseworthy (M = 7.62, SD = 1.98) and the morally bad actor as the most blameworthy (M =
4.71, SD = 2.61), and ratings of the morally neutral actor fell between the two (M = 6.67, SD =
1.67). Tukey’s post-hoc tests revealed that all conditions significantly differed, ps < .025.
Two bootstrap mediation analyses (10,000 resamples; PROCESS macro [Hayes, 2013])
revealed that perceptions that the morally good actor was more deserving of praise than the
morally neutral actor accounted for higher attributions of free will to the morally good actor,
95% CI [.190, 1.130] and that perceptions that the morally bad actor was more deserving of
blame than the morally neutral actor accounted for higher attributions of free will to the morally
bad actor 95% CI [-.860, -.249].
Unreported Study Y Method
The methods of Study Y were identical to Study X, with the exception that there were
two additional participants (n = 203).
Unreported Study Y Results
An ANOVA revealed a significant effect of condition on attributions of free will, F(2,
200) = 16.41, p<.001. As in Study 1b and Study X, Tukey’s post-hoc tests revealed that
participants who read about the morally good behavior (n = 63; M = 7.08, SD = 1.17) and the
morally bad behavior (n = 63; M = 7.12, SD = 1.17) attributed more free will to the actor than
participants who read about the morally neutral behavior (n = 77; M = 6.03, SD = 1.45), ps<.001.
And again, attributions of free will were slightly higher in the morally bad than morally good
condition, but they did not significantly differ, p = .986.
An ANOVA revealed a significant effect of condition on desires to blame and praise as
well, F(2, 200) = 230.53, p < .001. Participants rated the morally good actor as the most
praiseworthy (M = 8.03, SD = 1.19) and the morally bad actor as the most blameworthy (M =
2.14, SD = 1.97), and ratings of the morally neutral actor fell between the two (M = 6.61, SD =
1.59). Tukey’s post-hoc tests revealed that all conditions significantly differed, ps < .001.
Two bootstrap mediation analyses (10,000 resamples; PROCESS macro [Hayes, 2013])
revealed that perceptions that the morally good actor was more deserving of praise than the
morally neutral actor accounted for higher attributions of free will to the morally good actor,
95% CI [.256, .802]. However, in this study (unlike Study 1b and Study X), perceptions that the
morally bad actor was more deserving of blame than the morally neutral actor did not account for
higher attributions of free will to the morally bad actor 95% CI [-.701, .574]. Even after
controlling for the significant influence of desires to blame on higher attributions of free will, the
condition remained a highly significant predictor, p = .002.
Table 1S. Cohen's ds for differences in free will
attributions between conditions in Studies 1b, X, and Y
Comparison
Study 1b
Study X
Good vs. Neutral
1.04
0.42
Bad vs. Neutral
1.24
0.45
Bad vs. Good
0.27
0.03
Mini Metas Including Studies X and Y
As can be seen in Table 2S below, after adding Studies X and Y to the mini meta-
analyses, all conditions significantly differed with both fixed and random effects, ps < .02.
Table 2S. 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
0.24
Will Research (reversed)
n
139
137
p
0.127
0.0047
Study 1b
Attributions and Beliefs
r
0.10
0.37
(combined)
n
119
124
p
0.279
<0.0001
Study 2a
Attributions and Beliefs
r
0.19
(combined)
n
451
p
<0.0001
Study 2b
Attributions and Beliefs
r
0.12
(combined)
n
784
p
0.0008
Study X
Attributions
r
0.014
0.22
n
123
141
p
0.878
0.009
Study Y
Attributions
r
0.017
0.38
n
126
140
p
0.850
<.0001
Unweighted (random)
r
0.10
0.3
p
0.019
0.006
Weighted (fixed)
r
0.12
0.30
p
<.0001
<.0001
Discussion
Despite the inclusion of the anti-free will argument prior to the manipulation, these two
studies closely replicated the results of Study 1b (though note, the effect sizes for the differences
in free will attributions between conditions were smaller in these two studies than in Study 1b,
perhaps due do the inclusion of the anti-free will argument). The exception was that desires to
blame did not mediate the influence of the morally bad condition (relative to the neutral
condition) on higher attributions of free will. Participants did report significantly higher
attributions of free will and higher desires to blame in the morally bad condition than the neutral
condition, and higher desires to blame were significantly associated with higher attributions of
free will. Nonetheless, the mediation was not significant. Thus, this component of this study
failed to replicate Study 1b, Unreported Study X, and similar work demonstrating that desires to
punish and perceptions of moral wrongness mediate the influence of exposure to harmful actions
on higher attributions of free will and free will beliefs (e.g., Clark et al., 2014; Everett et al.,
2017), and the large body of work demonstrating that desires to blame influence other
attributions of responsibility (e.g., Alicke, 2000). Studies 2a and 2b also found that desires to
punish predicted higher attributions of free will and free will beliefs in the morally bad condition.
Given that we are also concerned about the soundness of the experimental methods of this study,
we are hesitant to put too much weight on this non-significant mediation. However, we hope that
other researchers will take note of this non-significant result and consider including it in any
future meta-analyses on the relationship between exposure to harmful actions, desires to blame
and punish, and attributions of responsibility.
One other noteworthy discussion point is that these two unreported studies and Study 1b
all found small to nonexistent differences in free will attributions between the morally good and
morally bad conditions (though, the morally bad condition was always [very slightly] higher than
the morally good condition). These findings are somewhat inconsistent with the results of
Studies 2a and 2b, our mini meta-analysis of all six studies, and the larger body of work
demonstrating higher attributions of responsibility for morally bad than good actions (e.g.,
Alicke, 1992; Knobe, 2003; Reeder & Spores, 1983). Two possible explanations come to mind.
Possibly, the lack of difference in free will attributions between these two conditions may be due
to the specific morally good and morally bad actions used in these three studies (a person
stealing expensive lasers from a hospital vs. a person donating expensive lasers to a hospital). If
so, we would expect to find little to no differences between morally good and morally bad
actions when these exact vignettes are used (we wrote these vignettes ourselves, so these three
studies are the only three times these vignettes have been used), but that most other well-matched
morally good vs. morally bad actions would expose larger differences.
Alternatively, the magnitude of the difference in responsibility attributions between
morally good and bad actions might vary as a function of a variety of features of the actions. For
example, as Studies 2a and 2b demonstrated, the perceived magnitude of the action (i.e., how
morally good or morally bad the action was perceived to be) increased free will attributions for
both morally good and morally bad actions. Thus, if an extremely morally bad action is paired
with a moderately morally good action, this could inflate the observed difference in free will
attributions between morally good and morally bad actions. We trust most researchers who
compare responses to morally good and bad actions at least attempt to make them equivalent in
extremity, but in some cases, this may be impossible. For example, what is the morally good
equivalent of a murder? The difficulty in matching morally good and bad actions may have led
researchers (ourselves included) to overestimate the extent to which people are motivated to
ascribe moral responsibility to morally bad actions relative to morally good ones. We hope future
research will continue to explore this possibility.

File (1)

Content uploaded by Cory Clark
Author content
ResearchGate has not been able to resolve any citations for this publication.
ResearchGate has not been able to resolve any references for this publication.