1

The Moral Identity Picture Scale (MIPS): Measuring the Full Scope of Moral Identity

Amelia Goranson, Connor O’Fallon, & Kurt Gray

University of North Carolina at Chapel Hill

Corresponding Author:
Kurt Gray
Department of Psychology and Neuroscience
University of North Carolina at Chapel Hill
CB #3270, 27599
[email protected]

in press, Self and Identity

 2

Abstract

Morality is core to people’s identity. Existing moral identity scales measure good/moral vs.

bad/immoral, but the Theory of Dyadic Morality highlights two-dimensions of morality: valence

(good/moral vs. bad/immoral) and agency (high/agent vs. low/recipient). The Moral Identity

Picture Scale (MIPS) measures this full space through 16 vivid pictures. Participants receive

scores for each of four moral roles: hero, villain, victim, and beneficiary. The MIPS can also

provide summary scores for good, evil, agent, and patient, and possesses test-retest reliability

and convergent/divergent validity. Self-identified heroes are more empathic and higher in locus

of control, villains are less agreeable and higher in narcissism, victims are higher in depression

and lower in self-efficacy, and beneficiaries are lower in Machiavellianism. Although people

generally see themselves as heroes, comparisons across known-groups reveals relative

differences: Duke MBA students self-identify more as villains, UNC social work students self-

identify more as heroes, and workplace bullying victims self-identify more as victims. Data also

reveals that the beneficiary role is ill-defined, collapsing the two-dimensional space of moral

identity into a triangle anchored by hero, villain, and victim.

Keywords: morality, self-perception, dyadic morality, measurement, moral character

 3

The Moral Identity Picture Scale (MIPS): Measuring the Full Scope of Moral Identity

People often grapple with questions about their identity, wondering whether they are

better understood as professional or a parent (Maurer et al., 2001), or as a soldier or a civilian

(Vest, 2013). Given the centrality of morality to identity (Strohminger & Nichols, 2014), people

also likely wonder how they fit into the moral world. Do they see themselves as an agent of good

or of evil, or as a recipient of other people’s kindness or cruelty? Although there is much prior

work that assesses variability in moral judgments (e.g., Greene & Haidt, 2002; Haidt, 2001;

Hofmann et al., 2014), and many measures that assess constructs related to moral identity (e.g.,

Barriga et al., 2001; Reimer & Wade-Stein, 2004), there are fewer measures that focus moral

identity per se. Those that do examine moral identity often examine only side of morality,

assessing either general self-perceived goodness (e.g. Aquino & Reed, 2002) or evilness (e.g.,

Christie & Geis, 2013). Here we present a new theoretically grounded image set—the Moral

Identity Picture Scale (MIPS)—that can be used for many purposes, including a way to measure

a fuller scope of moral identity.

This Picture Scale draws from an emerging perspective the Theory of Dyadic Morality,

which argues that morality revolves around common template of two—an intentional agent and a

vulnerable patient (see Gray, Young, & Waytz, 2012, Schein & Gray, 2018). Dyadic morality

thus argues for two intersecting dimensions of morality: valence (good/moral vs. evil/immoral)

and agency (high/agent vs. low/patient). Morality contains not only the doers/agents of moral or

immoral acts—heroes and villains—but also its recipients/patients—victims and beneficiaries.

Rather than the two moral roles of heroes and villains, there are four: hero (doer/agent of

goodness), villain (doer/agent of evil), victim (recipient/patient of evil), beneficiary

(recipient/patient of good). The research in this paper leverages this theory to measure a more
 4

expansive view of moral self-perceptions through images, which makes these four moral

exemplars both intuitive and vivid. Four studies demonstrate the validity of this approach.

Moral Identity

Identity is important. Much of our lives are spent attempting to answer the question of

“who am I?” Indeed, this question is at the very core of the study of psychology. Our identity

helps to organize our thoughts and direct our actions (Hutcheson, 1726), and helps define and

shape who we see as an “in-group” member (Cunningham, 2005; Gaertner et al., 1996). People’s

identities can reveal what kind of personality traits people are likely to have (Lilgendahl, 2015;

Luyckx et al., 2014); for example, someone who holds dear the identity of “volunteer” might be

more likely to also have personality traits of helpfulness or altruism. Identities are so powerful

that when we think or do something that violates an identity we hold dear, we often experience

powerful cognitive dissonance (Alicke et al., 1995; Barkan, 2015; Festinger, 1962; Stets &

Carter, 2011).

Moral identity is one of the most central facets of a person’s identity (Aquino et al., 2009;

Aquino & Reed, 2002; Stets & Carter, 2011; Strohminger & Nichols, 2014). Some research finds

that morality can even be considered a basic psychological need, and is crucial for our “peak”

experiences in life (Prentice et al., 2019). Indeed, morality seems to direct our cognitive

processing such that people identify peak experiences in their life using morality need

satisfaction as a barometer (Prentice et al., 2019). People often ask themselves whether they are a

morally good person who helps others, or a morally bad person who acts only in self-interest (S.

J. Reynolds & Ceranic, 2007)? Following past work on “positive illusions” about our traits and

abilities (Taylor & Brown, 1988), recent work finds that these positive illusions extend self-

perceptions of morality as well, with most people tending to see themselves as morally good
 5

(Tappin & McKay, 2016). We are so strongly motivated to see ourselves as morally good that we

deliberately work to alter or forget autobiographical memories about past moral transgressions

(Stanley & De Brigard, 2019). When we are unable to forget our sins, we attempt to strategically

compare our recent immoral behaviors to past immoral behaviors to create a narrative of

personal moral improvement over time. While most people generally see themselves as morally

good, the moral world is both dynamic and diverse. Our sense of our own moral identity can

shift based on social comparisons (compared to Mother Teresa, am I really a good person?) and

based on what exactly a group our culture defines as “good” (Janoff-Bulman & Carnes, 2013b;

Rai & Fiske, 2011). A history of killing is acceptable for a wartime soldier’s identity, but likely

bad for a civilian’s identity (Watkins & Laham, 2020).

There is a rich body of work emphasizing the importance of understanding—and

measuring—morality. However, the vast majority of the work on measuring morality focuses on

perceptions of others’ actions, intentions, motivations, desires, beliefs, and mental states to

evaluate (im)moral actions rather than on self-perceptions (Critcher et al., 2012; Gray et al.,

2012; Pizarro et al., 2003; Reeder, 2009). For example, there are studies about how liberals and

conservatives differentially evaluate acts (Graham et al., 2009; Janoff-Bulman et al., 2008;

Janoff-Bulman & Carnes, 2018), the importance of intention and causation, and the role of mind

perception (Schein & Gray, 2015). Most relevant to judgments of one’s own moral character are

studies on how we evaluate other’s character (Alicke, 2000; Goodwin et al., 2014; Pizarro &

Tannenbaum, 2011; Tannenbaum et al., 2011; for a review see Hartman et al., in press). People

ascribed good moral character are those who take environmental inputs and translates them into

socially acceptable outputs (Helzer & Critcher, 2018). The way that we see others’ moral
 6

character has strong implications for the way that we treat and talk about them (Aquino & Reed,

2002; Kohlberg, 1964).

Much work examines moral character judgments of others, but there is relatively less

work on self-perceptions of moral character. There is one popular scale developed by Aquino

and Reed (2002) that measures the self-importance of moral goodness, but otherwise little other

explores this topic—despite the time, effort, and thought each of us put into creating and

projecting our moral identities (Strohminger & Nichols, 2014). Some work shows that

perceptions of our moral identities can direct our behavior in identity-consistent ways (Hertz &

Krettenauer, 2016; Reed et al., 2007) and that we feel distressed when we perform moral

identity-inconsistent actions (Stets & Carter, 2011). Other work suggests that self-esteem is

often directly tied to perceptions of ourselves as a good person (Crocker & Park, 2004;

Rosenberg, 1965; Sheldon et al., 2001).

Despite little work measuring moral self-perceptions per se, there are many measures that

get at aspects of moral self-perceptions. For example, self-report scales on narcissism (Gentile et

al., 2013) or Machiavellianism (Christie & Geis, 2013) may capture part of self-perceived

immorality. Conversely, measure of empathy, such as the Interpersonal Reactivity Index (Davis,

1983), or positive social value orientation (Van Lange et al., 1997) may approximate positive,

agentic moral self-regard. Beyond these measures, other scales help to capture self-perceptions

of varieties of goodness or badness (e.g., self-esteem, self-efficacy), which likely feed into self-

perceptions of morality. Despite the usefulness of these scales, they are relatively narrow,

assessing one aspect (or sub-aspect) of morality. We suggest that the study of moral identity

could benefit from a stimulus set that allows a broader examination of moral identity.
 7

A Broader Structure of the “Who” of Morality

While work on the centrality, importance, and influence of our moral identities is

abundant, much of this work uses a relatively narrow definition of morality: that one is either

good or evil (Ayala, 2010). This valence-based definition of morality has been broadened in the

last few decades to include the influence of different types of acts (Graham et al., 2013), identity

(Hester & Gray, 2019), and character (Blasi, 2005). Even here, the vast majority of this research

has focused on the “active” side of morality (e.g., Hardy & Carlo, 2005)—the heroes who do

morally good acts, and the villains who do morally bad acts. One reason for this strong focus on

the active side of morality is because we tend to think of ourselves in terms of agency—"doing”

(Abele & Wojciszke, 2013). More generally, we are fascinated by moral agents/doers but not

moral recipients/patients. There are many thousands of books on heroes and villains but many

fewer on victims. Because of this domination of agency in social cognition, people may clearly

be able to simulate the experience of being a moral agent, but that it may be harder to simulate

the experience of being a moral patient. Although the doers—or “agents”—of morality are

undoubtedly important, an emerging perspective highlights a fuller understand of the moral

world

The Theory of Dyadic Morality (TDM; Schein & Gray, 2018) suggests that people

understand the moral world through a dyadic template of an agent harming (or helping) a patient.

TDM was initially developed to predict people’s judgments of (im)moral deeds and explain why

perceptions of harm robustly predict moral judgments across diverse scenario (Gray et al., 2012,

2014; Gray & Keeney, 2015; Gray & Schein, 2012). However, this theory can also be

understood as a map of different kinds of moral roles, defined through the intersection of two

dimensions of moral perceptions: a continuum of good versus evil, and a continuum of doer
 8

versus recipient of moral acts. These two dimensions of agency (high: hero, villain; low:

beneficiary, victim) and valence (good: hero, beneficiary; evil: villain, victim) divide morality

into four cells: heroes who help others, villains who harm others, victims who receive harm, and

beneficiaries who receive help. See Figure 1.

Figure 1. The two-dimensional structure of moral perceptions suggested by the Theory of

Dyadic Morality

Most past work has examined the “agent” side of morality through perceptions of

goodness and evil, but it can be useful to capture the “patient” side of morality. For example,

moral emotions appear to map onto this two-dimensional space (Gray & Wegner, 2011a), with

heroes like the Dalai Lama evoking emotions such as admiration or awe (Algoe & Haidt, 2009),

villains like Hitler evoking disgust or anger (Rozin et al., 1999), victims of wrongdoing evoking
 9

sympathy (Batson et al., 1981), and beneficiaries of help evoking relief (Cialdini et al., 1987).

This two-dimensional framework may also well capture self-perceptions, as different people

seem to see themselves along the axes of agency and valence—as heroes, villains, victims and

beneficiaries.

Heroes. Heroes are good moral exemplars and plenty of research suggests that people see

themselves in a positive light; for example, individuals tend to rate themselves as better than

average on a wide variety of traits or abilities (Alicke & Govorun, 2005), a finding that may be

especially prevalent in moral situations (Tappin & McKay, 2016). Despite generally heroic self-

perceptions, there is variance in these perceptions: negative affective states can suppress self-

perceived heroism (Pacini et al., 1998) and narcissism can inflate these perceptions (John &

Robins, 1994). Given the other-focused orientation of moral exemplars (Han et al., 2017), self-

perceptions of heroes should also be high in self-rated empathy (Fagin-Jones & Midlarsky, 2007;

Harvey et al., 2009; Jayawickreme & Di Stefano, 2012; Midlarsky et al., 2005) and the tendency

to help others in need (Franco & Zimbardo, 2006; Kohen et al., 2019).

Heroic people may also be invested in developing and maintaining their self-perceived

heroic identity. People like Martin Luther King Jr. and Nelson Mandela were driven in their

quests to not only do good things, but also to be good people. For heroes, it is likely important

that they see themselves as holding positive characteristics like being fair, caring, and honest

(Aquino & Reed, 2002). Those who report moral identity to be important to them also reported

higher levels of volunteerism and higher amounts of intrinsic satisfaction with participation. As

being a moral agent is not only about inner goodness, but about doing good deeds, and so we

suggest that seeing oneself as a hero should also involve higher perceptions of self-efficacy or an

internal locus of control.

 10

Villains. At first blush, few would seem to identify as villains. Who would want to see

themselves as instrumental in the suffering of others? However, research suggests that there is

variance in self-views, and that some individuals do hold negative self-views (Bernichon et al.,

2003; Malle & Horowitz, 1995). Moreover, villains may not be as negative as one might initially

think. In movies, villains are often more interesting than heroes, not only possessing moral

complexity, but also doing the kind of anti-social deeds that many fantasize about (Gollwitzer &

Denzler, 2009), such as enacting revenge or seizing power.

Perhaps most importantly, the harm that villains do is often instrumental, harming others

for the expressed goal of helping some subset of people, or bettering the world (i.e., burning it

down to build it back up). For example, the comic book hero Magneto is bent on the destruction

of human beings but only because they pose an existential threat to his fellow mutants (Lee,

1963). Recent work has identified that instrumental harm plays a large role in those who hold a

utilitarian moral ideology (Everett & Kahane, 2020; Kahane et al., 2018). This can be contrasted

to those who follow more deontological or rule-based morality where doing harm may be

considered wrong no matter the benefit. Indeed, many acts that people see as villainous may in

fact be perceived as necessary by the perpetrator. Acts like honor killings, torture, following

orders from authority or God can be rationalized into the “correct” choice (Fiske & Rai, 2014).

Those who identify as villains may see themselves as enacting necessary evils for

eventual benefits, and may they recognize that others see them as morally flawed—and accept

such judgements. Those who see themselves as doing necessary immoral deeds may also identify

as a hero (because sometimes heroes must o make tough decisions). We can therefore expect

some overlap between the hero and villain dimension, especially because both are agents/doers

of deeds. Ultimately, viewing oneself as a villain can allow for more self-serving actions. The
 11

confidence in one’s own moral compass—however skewed—and the willingness to harm others

suggests that those who identify as villains should be high on narcissism and low on empathy.

Victims. Victims occupy the “evil/patient” quadrant of the moral space as people who

receive bad deeds. This is an important identity to investigate because many people experience

victimization in one form or another throughout life, either by sexual assault (nearly half a

million each year; RAINN, 2020), crime to one’s home (about 25% of households; Gallup, 2014),

or domestic violence (1 in 4 women, 1 in 9 men; Truman & Morgan, 2014). Sometimes people

try to take on the mantle of victimhood in intergroup interactions (i.e., competitive victimhood;

Sullivan et al., 2012), this is often employed to escape culpability for immoral actions like

discrimination. Those who signal to others that they are victims with good morality also tend to

be associated with the dark triad, and in particular Machiavellianism (Ok et al., 2020). Some

people may use victimhood to gain resources like money and social support from others.

Though these findings seem to indicate a benefit of victimhood, it seems that identifying

as a victim likely has downsides. Socially, victims can feel stigmatized as weak, vulnerable, and

in some cases be blamed for their victimization (Fohring, 2018; Hafer, 2000). Victims may also

feel like they will be socially cut off if they do not fill these perceptions. Negative reactions to

disclosing a traumatic event can lead to poor health outcomes (Hakimi et al., 2018) and, as the

recipient of harm, victims likely feel powerless and low in self-efficacy. Given that

powerlessness can induce depression (Swearer et al., 2001) and is tied to emotional instability

(Glasø et al., 2007), we suggest that those who identify as victims will be high on these

constructs. Many people experience events in which they feel as if they have been victimized.

People who have been repeatedly victimized may also identify as being a victim more generally
 12

and while they might gain certain benefits, they also likely incur costs and adverse physical,

social, and mental effects.

Beneficiaries. Beneficiaries are those who receive help. Out of all four roles,

beneficiaries are the least studied, perhaps because they (arguably) represent the least pressing of

social challenges. It is obviously important to stop people from doing evil and to foster good

deeds, which is why most people study good and evil agents. It is also important to understand

the psychology of victims because they clearly need help, but understanding those who have

been helped seems less urgent. However, one could debate how important it is to understand

identities of beneficiaries who are already receiving help. On one hand, those who identify as

beneficiaries may feel positive emotions—with all their benefits (Cohn et al., 2009; Fredrickson,

2001)—and increased feelings of self-efficacy after being helped (DeSteno et al., 2010; Isen et

al., 1976). On the other hand, they may also still feel “patient-like” and powerless (Aujoulat et

al., 2007).

Identifying as a beneficiary could be strongly tied to identifying as a victim (because both

are patients) or as a hero (because both are on the “good” side of valence). Victims are people

who need help, and beneficiaries are those who receive help, so it makes sense that people who

see themselves as beneficiaries likely have been victims at one point—they are now receiving the

help they needed. In fact, the Theory of Dyadic Morality (Schein & Gray, 2018) argues that

victims are the pre-eminent moral patient, and that—at least in third-party judgments—

beneficiaries are simply a variety of victims (victim who have been helped). While those who

identify as victims might not necessarily identify as a beneficiary, those who identify as

beneficiaries may remember negative situations or parts of their life that also lead them to

identifying as a victim. Thus, we may see a correlation between victims and beneficiaries.
 13

We also could see a correlation between beneficiaries and heroes. Given prior work that

shows that individuals are agency-focused rather than communion-focused in their self-related

thoughts (Wojciszke et al., 2011), it is possible that individuals will more easily identify with

agent-focused roles, especially given that we are asking participants to put themselves in the

shoes of the individuals depicted in each scenario. This could make the beneficiary role

particularly difficult to identify with, especially given that negative events tend to be more

salient, so the most noticeable or impactful patient role is victimhood. Receiving help can be tied

to feelings of empathy and gratitude (Bartlett & DeSteno, 2006; Dongjie et al., 2018).

Plenty of work has also found people’s tendency to “pay-it forward” when someone else

does something good for them (Bartlett & DeSteno, 2006; Chang et al., 2012; Chiang &

Takahashi, 2011; Dongjie et al., 2018; Gray et al., 2014; Horita et al., 2016; Tang et al., 2021;

Tsvetkova & Macy, 2014) . Those who receive help, give help, which is why those who see

themselves as beneficiaries may also see themselves as heroes. How people view their life stories

can also impact their identity. Studies of life narratives find that those with a generative lifestyle

(one that helps others; the hero) are more likely to discusses times in their life in which they were

a beneficiary or received an early advantage (McAdams et al., 1997, 2001). This is another

reason why we would expect the hero and beneficiary parts of the scale to correlate.

Altogether, there is reason to suspect that the beneficiary role may collapse into other

roles in self-perceptions of moral identity. Consistent with this idea, recent research suggests a

more “triangular” structure for moral character judgments of other people (Giner-Sorolla et al.,

n.d.; Hartman et al., in press). This triangle is still mapped on the space of valence and agency,

but excludes beneficiary—the vertices are hero, villain and victim.

 14

The Moral Identity Picture Scale

In the current studies, we provide and assess a scale that helps to assess the fuller scope

of self-perceived moral identity: the Moral Identity Picture Scale (MIPS). The “pictures” of the

MIPS are stylized drawings of pairwise interactions, and they depict good and bad moral agents

(heroes and villains) as well as moral patients (victims and beneficiaries), reflecting the four-cell

design laid out by the Theory of Dyadic Morality. For the individual files of all 16 pictures,

please see the OSF link: https://osf.io/faz85/. For the full scale, see Appendix A. For the scoring

procedure, see Appendix B.

The MIPS consists of 16 images presented in a random order, each on its own screen.

Each picture captures the connection between adjacent cells depicted in Figure 1. Four images

depict a hero and villain, four depict a hero and beneficiary, four depict a villain and victim, and

four depict a victim and beneficiary. Although TDM emphasizes heroes and beneficiaries and—

especially—villains and victims—it is useful to measure within-agency pairs (i.e., both agents or

both patients) to better contrast the roles.

To provide increased generalizability and reliability, there are four versions of each

pairing, and each was created with a few criteria in mind. In creating the items, we first wanted

scenarios that would show a dyad interacting. Second, we chose scenarios that could be drawn

out with minimal context needed to understand what was happening between the two individuals

pictured. Lastly, we wanted to make these scenarios diverse and relatable. We wanted to include

a variety of actions from mundane situations—like comforting a friend who is feeling down or

saying something cruel to a colleague—to more extreme examples—like being rescued from a

burning building—that you might see in a comic book or television show. With these criteria in

mind, we brainstormed possible scenarios with our extended lab group. We chose the 16 with the
 15

highest level of consensus for clarity, simplicity, and vividness. Then, a professional artist—

Canadian cartoonist and illustrator Shawn Daley—drew out each situation. We did another round

of review and edited wording for clarity, which led to our final set of photos. The sampling of

this space allows the MIPS to better cover the range of possible moral scenarios that might

resonate with people. See Figures 2 and 3 for examples.

Not only does the MIPS differ from past measures by assessing a broader scope of moral

identity, it also differs from past measures by using pictures rather than words. For each picture,

participants indicate how much they identify with each person in the frame, which provides 32

individual ratings which are then combined into 4 moral identity subscale scores: one for each of

hero, villain, victim, and beneficiary.

Picture-based questionnaires in research have been shown to increase engagement with

the research process (Puleston, 2011) and be more fun, without sacrificing quality of responses

(Puleston, 2013). Further, we chose this type of measure because it provides a richer narrative for

each participant to engage in. Past work shows that being immersed in a narrative increases

motivation (Barraza et al., 2015), improves memory (Cahill & McGaugh, 1995; Heath & Heath,

2007), and can even improve theory of mind (Kidd & Castano, 2013). Some have even gone so

far as to claim that this penchant for narratives and storytelling is the very essence of our

humanity (Gottschall, 2013). Whether or not the human mind is truly built for story, pictures can

at least capture situations more succinctly than words. Again, see supplementary material for all

items (Appendix A) and scoring procedures (Appendix B).

 16

Figure 2. MIPS image 7: Villain/Victim

Figure 3. MIPS image 16: Hero/Beneficiary

Current Studies

In four studies, we use the MIPS to explore moral identity, testing its construct validity,

test-retest reliability, and convergent validity with real-world known groups. Study 1 examines
 17

the internal consistency of the hero, villain, victim, and beneficiary roles, and uses

multidimensional scaling to explore the structure of moral self-perceptions. We expect that the

multidimensional scaling will result in two dimensions with one representing valence

(good/moral vs. bad/immoral) while the other closely resembles agency (high/agent vs

low/recipient/patient). Study 2 examines the test-retest reliability of the MIPS over a period of

approximately 30 days. We believe that moral identities will be stable over this time frame and

that Time 1 identification should robustly predict of Time 2 identification.

Studies 3a and 3b examine the convergent and divergent validity of the MIPS with a

variety of validated scales that should relate to moral self-identity, such as measures of

depression, self-efficacy, and the importance of moral identity. We predict that hero

identification should correlate with self-efficacy and empathy, villains should show high levels

of Machiavellianism, and victims should be high in levels of depression and low in self-efficacy.

Because of the relative vagueness of beneficiaries, we do not hold any specific predictions about

how it will correlate with other key scales. We also predict that those who see themselves as

heroic will score highly on the self-importance of moral identity questionnaire, which asks

participants to imagine that they have positive moral characteristics like being kind, helpful, or

honest (Aquino & Reed, 2002).

Study 4 examines known groups validation by giving the MIPS to members of real-world

groups that are perceived to occupy different moral roles in our society. We expect that those

who are Masters of Social Work students at the University of North Carolina should be high in

our hero identification. The villain identity is expected to be endorsed more by Masters of

Business Administration students at Duke University, while those who identify as having been
 18

bullied in the past should be endorse the victim identity. There was no group specifically selected

to represent the beneficiary category due to theoretical indistinctness.

We present these studies the order that allows us to create a logical narrative flow, rather

than the order in which they were conducted in the laboratory or online. We include information

on a priori power analysis and pre-registration where appropriate. All materials including data,

code, pre-registrations, and supplemental material can be found on https://osf.io/faz85/. All

studies were approved by the UNC IRB #: 16-2315. Before proceeding we note that this measure

is not argued to be “the best” measure of morality per se but rather an additional tool in the

toolbox of identity and moral psychology researchers.

Study 1: Internal Consistency and Moral Map

Study 1 provided an initial investigation into the Moral Identity Picture Scale (MIPS),

measuring the internal consistency reliability of each of four potential self-identifications—hero,

villain, victim, and beneficiary—might relate to one another. We also examined the correlations

between each four roles and submitted the scores to a multidimensional scaling procedure to

provide a “map” of the structure of morality identity. We predicted that we would reproduce the

structure found in Figure 1, with dimensions of agency (agent/patient) and valence (good/evil)

and each four roles in the expected quadrant.

Method

Preregistration. We pre-registered this study using AsPredicted. We pre-registered a

correlational design with 200 participants, which we deemed would provide sufficient power to

reveal correlations in this within-subjects study. Pre-registered analyses were Pearson

correlations between MIPS target identities, and multi-dimensional scaling of the relationships

between the four MIPS target identities.

 19

Participants. Two hundred and five participants (95 male, 109 female, 1 non-binary;

Mage = 37.08, SD = 24.19) were collected via Amazon’s Mechanical Turk (mTurk). After

screening out those who failed all attention checks, we were left with 174 participants (77 male,

96 female, 1 non-binary; Mage = 35.99, SD = 12.56).

Procedure. Each participant saw the 16 pictures from the MIPS. Below each picture,

participants answered, “How much do you identify with each person above?” on 1 (not at all) to

4 (extremely) Likert scale, for both the picture on the left and the picture on the right.

Results

Internal Consistency. Are each of the moral identity roles reliable across each of the 8

pictures that assess each role? Cronbach’s alphas for hero items (α = 0.79), villain items (α =

0.86), victim items (α = 0.82), and beneficiary items (α = 0.73) suggest reasonable—but not

extremely high—internal consistency. Given differences between story content in each picture

and the intentional variation of factors including sex of characters and kind of harm/help, it is not

surprising that there is variation within each role. Although all moral roles have alpha greater

than .70—the recommended minimum for a measure (Nunnally, 1978)—we note that the

beneficiary role is the least internally consistent. We suggest that this reflects the fuzziness of

this very construct. Although people seem to have strong archetypes about heroes, villains, and

victims—there seem to be little consensus about what a beneficiary is like. In fact, beneficiaries

are often talked about in the same terms as victims, or grouped into one “moral patient” category

in past research (e.g., Baumeister, 1997; Bernstein, 1998; Crimston et al., 2016). Moreover, as

we have noted before, beneficiaries are usually first victims, and those who are helped usually

help others making them hero like, further blurring the lines around this construct.
 20

Mean Scores and Correlations. As both Figure 4 and Table 1 reveal, people generally

self-identified most with the hero, next with the beneficiary, and next with the victim. They

identified least with villain. In other words, they saw themselves most highly as good-doers, least

highly as evil-doers, and intermediate moral recipients/patients, whether good or evil.

Correlations (Table 1) reveal that all roles are significantly related, except for

identification as a hero and a villain. Please see Figure 4 for identifications across roles. While it

might be intuitive to expect hero and villain to be inversely related, we suggest that the

underlying dimensions of agency could increase the association between these constructs.

Indeed, a dimensional structure—which we assess next, could give rise to positive associations

across many roles.

Table 1: MIP Subscale Scores and Correlations

Variable M SD α 1 2 3

1. Hero 2.75 0.62 0.79

2. Villain 1.61 0.64 0.86 .05

[-.10, .19]

3. Victim 2.07 0.68 0.82 .24** .38**

[.09, .37] [.25, .50]

4. Beneficiary 2.33 0.58 0.73 .56** .34** .44**

[.45, .66] [.21, .47] [.31, .55]

Note. M and SD are used to represent mean and standard deviation, respectively. Values in

square brackets indicate the 95% confidence interval for each correlation. * indicates p < .05. **

indicates p < .01.

 21

Self-Identification 2

0
Hero Villain Victim Beneficiary

Figure 4. MIPS Subscale Means in Study 1, across MTurk participants.

Multidimensional Scaling. Scores for each role were submitted to PROXSCAL in

SPSS, which yielded the structure in Figure 5. The model showed a stress value of 0.027,

indicating an excellent fit (Kruskal, 1964). Largely consistent with the structure outlined by the

Theory of Dyadic Morality (Figure 1), this MDS analysis appears to reveal two dimensions—a

valence dimension (good/evil) running left-right, and an agency dimension running up-down.

Although the villain, hero and victim are where one might predict, the beneficiary role appears

poorly distinguished. Again, we suggest that this because this role is cognitively ill-defined and

likely has stronger ties to other roles.

 22

Figure 5. Multi-Dimensional Scaling results for the MIPS (Study 1).

Discussion

Study 1 examined initial properties of the MIPS. Ratings revealed reasonable internal

consistency ratings for each role, indicating that the pictures assessed some breadth of situations

and interpretations of each role. Inter-correlations were positive for all roles except for heroes

and villains. There may be some common variance in just being willing to assign yourself a role

in a story, but part of this overlap is also likely explained by the findings of the MDS, which

revealed an underlying structure of moral self-perception consisting of two dimensions—one

related to valence (good/evil) and one related to agency (agent/patient) consistent with the

predictions of dyadic morality. Here we see the mutability of the beneficiary role as it seems to

fall directly between the victim and hero roles. As discussed previously this is likely due to those

who identify as victims and heroes also acknowledging that they’ve been helped along the way.
 23

This provides some basic confidence in the psychometric properties of the MIPS. Next, we

examined the temporal stability of the MIPS with a test-retest design.

Study 2: Temporally Stable Moral Identity

Although some aspect of people’s identities can vary from day to day and across

situations, identity is often considered to be something relatively stable, with self-perceptions of

“who I am” possessing some temporal integrity. Past work on moral typecasting (e.g., Gray &

Wegner, 2009) suggests that perceptions of the moral identity of others can remain stable over

time. In this study we examined the temporal stability of self-perceived moral identity as

assessed by the MIPS. We predicted a significant correlation for test-retest reliability across a

span of approximately one month.

Method

Participants. One hundred thirty-eight undergraduate students (35 male, 103 female;

Mage = 19.34, SD = 3.03) participated in this study for course credit, providing us with sufficient

statistical power to detect a small-to-medium effect with 80% power based on an a prior power

analysis.

Procedure. Participants signed up for two lab sessions, approximately one month apart

from each other. They completed the same procedure at each session, including evaluating the

MIPS and providing demographic information. The participants were fully debriefed upon

completion of the study. We hypothesize that moral self-identification should be relatively

stable across time. Thus, we would expect Time 1 responses to correlate with Time 2 responses

across the participants in Study 2.

 24

Results

We examined the relationship of participant answers at Time 1 to those at Time 2. We

first examined all item-level correlations across participants. Participants’ responses at Time 1

was significantly correlated to their responding at Time 2 across all items (ps < .05). For ease of

reporting, we collapsed item identification into four categories: hero (r = 0.68, p < .001), villain

(r = 0.79, p < .001), victim (r = 0.77, p < .001), and beneficiary (r = 0.70, p < .001), all of which

showed significant Time 1 - Time 2 agreement. Please see Table 2 for all Time 1 – Time 2

correlations.

Table 2: Scores and Test-Retest Reliability for MIPS Sub-Scales (Study 2)

Variable M SD 1 2 3 4 5 6 7

1. Hero (T1) 3.00 0.46

2. Hero (T2) 2.92 0.53 .68**

[.57, .76]

3. Villain (T1) 1.81 0.50 -.05 -.01

[-.21, .12] [-.18, .16]

4. Villain (T2) 1.79 0.54 -.03 .09 .79**

[-.20, .14] [-.08, .26] [.72, .84]

5. Victim (T1) 2.30 0.55 .36** .32** .20* .16

[.21, .50] [.16, .46] [.03, .35] [-.00, .32]

6. Victim (T2) 2.25 0.61 .30** .43** .13 .18* .77**

[.14, .45] [.28, .56] [-.04, .29] [.02, .34] [.69, .83]

7. Beneficiary
2.44 0.48 .50** .53** .09 .07 .41** .30**
(T1)
[.36, .61] [.39, .64] [-.08, .25] [-.09, .24] [.26, .54] [.14, .44]

8. Beneficiary
2.42 0.48 .45** .61** .08 .15 .35** .41** .70**
(T2)
[.31, .57] [.49, .70] [-.09, .24] [-.01, .31] [.19, .49] [.26, .54] [.60, .78]

Note. M and SD are used to represent mean and standard deviation, respectively. Values in
 25

square brackets indicate the 95% confidence interval. * indicates p < .05. ** indicates p < .01.

We next examined whether identification at Time 1 would predict identification at Time

2 in regression models. Hero identification at Time 2 was significantly predicted by both hero (b

= .63, t(133) = 7.54, p < .001) and beneficiary (b = .28, t(133) = 3.39, p = .001) identification at

Time 1; villain (b = -.01, t(133) = -.11, p = .91) and victim (b = .02, t(133) = 0.18, p = .79)

identification were not significant predictors, R2 = 0.51, F(4, 133) = 34.13, p <.001. Villain

identification at Time 2 was significantly predicted only by villain identification at Time 1 (b =

.86, t(133) = 14.35, p < .001); hero (b = .003, t(133) = 0.04, p = .97), victim (b = .01, t(133) =

0.12, p = 0.91), and beneficiary (b = .003, t(133) = 0.04, p = .97) identification at Time 1 were all

non-significant predictors, R2 = 0.62, F(4, 133) = 54.99, p <.001. Victim identification at Time 2

was significantly predicted only by victim identification at Time 1 (b = .87, t(133) = 12.07, p <

.001); hero (b = .06, t(133) = 0.71, p = .48), villain (b = -.02, t(133) = -.24, p = 0.81), and

beneficiary (b = -.05, t(133) = -.63, p = .53) identification at Time 1 were all non-significant

predictors, R2 = 0.59, F(4, 133) = 47.46, p <.001. Finally, beneficiary identification at Time 2

was significantly predicted only by beneficiary identification at Time 1 (b = .62, t(133) = 8.41, p

< .001); hero (b = .14, t(133) = 1.81, p = .07), villain (b = .02, t(133) = .40, p = 0.69), and victim

(b = .04, t(133) = 0.57, p = .57) identification at Time 1 were all non-significant predictors of

beneficiary identification, R2 = 0.51, F(4, 133) = 34.21, p <.001.

Discussion

The results of Study 2 supported our hypothesis that ratings on the MIPS are relatively

stable over time, at least within a month-long period. Not only do we find that Time 1

identifications positively correlate with corresponding identification at Time 2, but we also find
 26

in regression analyses that identification for each role at Time 1 is the key predictor of that role

at Time 2. This lends support to the idea that moral self-identification as measured by the MIPS

is a relatively stable over time, indicating that people may group themselves based on their moral

self-perceptions. This second study also finds ambiguity with the beneficiary role. Here

identification of being a beneficiary at Time 1 significantly predicted identifying as a hero at

Time 2. One explanation for this finding is that participants who identified with being helped

ended up helping others and therefore saw themselves as more of a hero in the follow-up. In

Studies 1 and 2, we find support for the 4 identities in our scale and support for the test-retest

reliability of the Moral Identity Picture Scale. In the next two studies, we turn to tests of validity.

Study 3a: Assessing Convergency with Other Measures

This study examined convergence between the elements of the MIPS and other measures

that have been previously — or plausibly — related to moral identity. Participants took a battery

of existing scales as well as the MIPS and we explored both convergent and divergent validity.

Among the predictions are that those who self-identify as 1) heroes would show high levels of

perceived self-efficacy or empathy, 2) villains would show high levels of Machiavellianism, 3)

victims would show high levels of depression and low self-efficacy. The predictions regarding

beneficiaries were less clear. One important thing to note is that this study omitted the self-

importance of moral identity measure (Aquino & Reed, 2002). Given the popularity of this

measure and its clear relevance to moral character, we wanted to have a separate dedicated study

to examine the links between this measure and the MIPS (Study 3b).

Method

Participants. Power analysis based on a small-to-medium effect size and 80% power to

detect effects demonstrated that our sample should include 70 participants. Seventy-one mTurk
 27

workers participated in this study (25 male, 46 female, Mage = 32.11, SD = 15.94), which took

approximately 20-30 minutes to complete. This timing was pre-tested by undergraduate research

assistants before collecting our final sample.

Procedure. Each participant rated the Moral Identity Picture Scale followed by 14

validation scales in randomized order. To yield scores for each of the four moral identity types,

we averaged across all images that assessed each type, reverse coding where appropriate.

The various scales were Deceptive Behavior Scale (Phillips et al., 2011), the Empathic

Concern and Perspective Taking Subscales of the Interpersonal Reactivity Index (IRI, Davis,

1983), Risk Taking Tendency Measure (Brache & Stockwell, 2011), Strength Self-Efficacy

Scale (Tsai et al., 2014), Machiavellianism Scale (Christie & Geis, 2013), Social Value

Orientation (Van Lange et al., 1997), Ten-Item Personality Inventory (TIPI, Gosling, Rentfrow,

& Swann, 2003), Narcissistic Personality Inventory (NPI, Gentile et al., 2013), Willingness to

Engage in Help Seeking (Hammer & Vogel, 2013), Rotter Locus of Control (Rotter, 1966),

Adult Victim Scale (Rigby & Slee, 1993), and Beck Depression Inventory (BDI, suicide question

omitted; Beck, Ward, Mendelson, Mock, & Erbaugh, 1961). In addition, participants took 2

scales that should be unrelated to our moral self-perception measures as controls: Paranormal

Belief Scale (Tobacyk, 2004) and Materialism Scale (Sirgy et al., 2012) and completed two

attention check items (“I can fly”—only accepted strongly disagree; “No one has ever disliked

me in my entire life”—only accepted “strongly disagree”), which were presented alongside the

measures. Finally, participants provided demographics and were debriefed.

Results

We analyzed the data to examine correlations between MIPS ratings and other existing

scales that should be related to aspects of moral self-identification. We primarily examined these
 28

correlations both within each of our four moral types – heroes, villains, victims, and beneficiaries

– as well as across agents and patients, and positive and negative moral valence. See Table 3 for

these correlations (full correlational table provided in the supplementary material).

Table 3. Correlations between MIPS items and other scales (Study 3a).
 29

Hero. Our analyses reveal that identifying with the hero character in our measure is

positively related to self-efficacy, r (69) = 0.51, p < .001, empathy, r(69) = .41, p = .001,

extraversion, r(70) = 0.28, p = .018, emotional stability, r(70) = 0.51, p < .001, and an internal

locus of control, r(67) = 0.27, p = .027, and negatively related to Machiavellianism, r(69) = -

0.41, p < .001 and depression, r(69) = - 0.24, p = .048.

Villain. Identifying with the villain in our measure is positively related to deceptive

behavior, r(64) = 0.49, p < .001, and materialism, r(68) = 0.37, p = .002, and narcissism, r(67) =

0.58, p < .001. Surprisingly, we did not observe a significant relationship with Machiavellianism,

r(68) = 0.01, p = .91. However, it was negatively related to agreeableness, r(67) = - 0.45, p <

.001.

Victim. Identifying positively with our victim character was positively related to adult

victimhood, r(70) = 0.32, p = .008, depression, r(69) = 0.53, p < .001, and deceptive behavior,

r(65) = 0.26, p = .034; victim identification was negatively related to self-efficacy, r(69) = -

0.32, p = .007, extraversion, r(70) = - 0.26, p = .032, and emotional stability r(70) = - 0.38, p <

.001.

Beneficiary. Lastly, identifying with the beneficiary character was positively correlated

with self-efficacy, r(69) = 0.32, p = .008 and empathy, r(68) = 0.26, p = .035. It was also

negatively related to Machiavellianism, r(69) = - 0.30, p = .013.

Agent. Averaging across heroes and villains to create an “agent” score, we find that

being rated high on agent was correlated significantly with self-efficacy, r(67) = 0.47, p < .001,

extraversion, r(68) = .31, p = .011, and narcissism r(66) = 0.46, p < .001. Agent identification

negatively correlated with Machiavellianism, r(67) = - 0.32, p = .011 and depression, r(67) = -

0.32, p = .008.
 30

Patient. Averaging across beneficiaries and victims to create a “patient” score, we find

that patients are likely to rate higher on depression, r(69) = 0.32, p = .008.

Positive moral valence. Averaging across hero and beneficiary, we created a “positive

moral valence” identification score. Self-efficacy, r(70) = 0.48, p < .001, empathy, r(69) = 0.38,

p < .001, and emotional stability, r(71) = 0.26, p = .03, were positively associated with positive

moral valence. Machiavellianism, r(70) = -0.41, p < .001, and an internal locus of control, r(68)

= -0.26, p = .03, were negatively associated with positive moral valence identification.

Negative moral valence. Averaging across villain and victim identification, we created a

“negative moral valence” identification score. Narcissism, r(68) = 0.24, p = .05, deceptive

behavior, r(65) = 0.47, p < .001, adult victim, r(70) = 0.33, p = .01, and depression, r(69) = 0.28,

p = .02, were all positively correlated with negative moral valence identification. Agreeableness,

r(68) = -0.35, p < .001, emotional stability, r(68) = -0.38, p < .001, and paranormal belief, r(65)

= -0.30, p = .02, were all negatively correlated with negative moral valence.

Discussion

These results revealed some convergent and divergent validity for self-perceived moral

identity as assessed by the MIPS. For each of hero, villain, victim, and beneficiary, moral

identity seem to cohere with the most relevant subscales. This provides us with further

confidence that identification on the MIPS will map onto (im)moral identification as an agent or

patient in the world at large. While we strived to use a wide variety of previously validated

measures in Study 3a, there are certainly other psychological constructs that could relate to moral

self-identification, such as subjective socio-economic status, importance of interpersonal

relationships, or emotion regulation. Future research should continue to examine how other
 31

potentially relevant social psychological constructs relate to moral self-identity. Next, we

examine the validity of the MIPS in the specific context of moral identity.

Study 3b: Convergence with a Moral Identity Measure

In Study 3b, we continue our investigation of the validity of the MIPS, this time using a

pre-registered study to compare the MIPS to the most popular moral self-identification measure:

the self-importance of moral identity scale (SIM-Q) proposed by Aquino & Reed (2002). Since

the SIM-Q asks participants to imagine themselves as having positive moral characteristics like

honesty or generosity, we hypothesize that those who find these important would identify as high

on the MIPS hero identity. Conversely, we hypothesize that participants who rate those same

types of traits as unimportant may rate highly on the MIPS villain identity.

Method

Participants. Two hundred and four (104 male, 100 female; Mage = 35.64, SD = 10.52)

mTurk workers completed our survey for compensation. The study was pre-registered using

AsPredicted. After screening out individuals who failed attention checks, we were left with 123

participants (57 male, 66 female; Mage = 36.96, SD = 11.50). This study was conducted at a time

when mTurk was overrun with bots/server farms, hence the high number of exclusions.

Procedure. Participants took the MIPS as well as the self-importance of moral identity

measure (SIM-Q) by Aquino & Reed (2002). Finally, they provided demographic information

and were debriefed.

Results

First, we examined internal consistency of our measures. We found sufficient internal

consistency for hero items (α = 0.85), villain items (α = 0.81), victim items (α = 0.83), and

beneficiary items (α = 0.78); we made indices for each subscale.

 32

We also examined internal consistency for the self-importance of moral identity scale,

which was sufficient (α = 0.78) and similar to the internal consistencies for the MIPS. Next, we

checked for sufficient internal consistence of the two Aquino & Reed (2002) subscales:

symbolization (α = 0.89) and internalization (α = -0.08)1.

Consistent with our predictions, we also find that self-identifying with the hero role on

the MIPS is positively related to seeing one’s moral identity as important on the self-importance

of moral identity scale as a whole (r = 0.40, p < .001), and on both subscales (symbolization: r =

0.34, p < .001; internalization: r = 0.38, p < .001). MIPS villain identification was not

significantly related to moral self-identification broadly (r = 0.11, p = .23), or on either subscale

(symbolization: r = 0.11, p = .25; internalization: r = 0.08, p = .40). Inter-correlations between

MIPS identifications are similar to earlier studies, showing consistency across samples. See Table

4 for all correlations.

Table 4. Correlations between MIPS subscales and SIM-Q.

Variable M SD 1 2 3 4 5 6

1. Hero 2.79 0.72

2. Villain 1.48 0.54 .03

[-.15, .20]

3. Victim 2.04 0.69 .39** .16

[.23, .53] [-.02, .33]

4. Beneficiary 2.30 0.64 .72** .24** .45**

[.62, .79] [.07, .40] [.30, .58]

5. SIM-Q (all) 3.18 0.62 .40** .11 .16 .40**

[.24, .54] [-.07, .28] [-.02, .33] [.24, .54]

6. SIM-Q
3.08 1.03 .34** .10 .14 .34** .96**
(symbolization)
[.18, .49] [-.07, .28] [-.03, .31] [.18, .49] [.95, .97]

1
This value is negative due to a negative covariance among items, a violation of reliability model assumptions.
However, item codings were verified accurate. Please interpret with caution.
 33

7. SIM-Q
3.27 0.38 .38** .08 .14 .39** .68** .45**
(internalization)
[.21, .52] [-.10, .25] [-.04, .31] [.23, .53] [.57, .76] [.30, .58]

Note. M and SD are used to represent mean and standard deviation, respectively. Values in square
brackets indicate the 95% confidence interval for each correlation. * indicates p < .05. ** indicates p <
.01.

Discussion

Study 3b provides additional evidence that the MIPS captures moral identity. When

compared to a previously validated measure of moral identity, we find similar results. Namely,

those who find a positive moral self-identity important by Aquino & Reed's (2002) measure also

seem to rate highly on the hero identity of the MIPS. Further, beneficiary identity is also related

to centrality of a positive moral self-identity on the SIM-Q, emphasizing that beneficiaries

structurally lie on the positive side of valence within the moral identity space. Identifying as a

villain, on the other hand, is not significantly related to Aquino & Reed's (2002) SIM-Q measure.

Since the SIM-Q measures moral self-identification with positive moral traits such as honesty

and generosity, the lack of relationship between villain identification and SIM-Q score suggests

that those identifying as villains may not value these characteristics. Interestingly, our results

suggest that self-identified villainy and the SIM-Q are not negatively related—this lack of

relationship between these variables may indicate that MIPS villains are simply not concerned

with possessing positive moral traits identified by the SIM-Q or that they are rationalizing what

others may deem to be immoral as necessary or beneficial for oneself. Rather than a mustache-

twirling, sadistic supervillain from traditional conceptions of immorality, these may be

individuals who do not place high value on being seen as or self-identifying with a positive

moral identity. They may see moral concerns as simply irrelevant or outdated, and may prioritize

self-interested actions regardless of whether they may be interpreted negatively by others. While
 34

they may not strive to be actively dishonest, they likely do not prioritize honesty in their personal

actions.

In Study 4, we continue our validation of the MIPS by testing all 4 identities/subscales on

known groups of individuals.

Study 4: Known Groups Validation

In Study 4, we examine this idea of moral self-perception in more detail. For example, do

those who dedicate their lives to helping others identify with the hero character in our measure?

We examine four groups of individuals who should classify themselves into the four characters

of our measure – Master of Social Work (MSW) students from the University of North Carolina,

Master of Business Administration (MBA) students from Duke University, individuals who self-

identify as being bullied in the workplace, and mTurk workers.

We predicted that UNC MSW students, who devote their careers to helping others when

they are in need, would identify more with our hero profile than other known groups (e.g., Duke

MBA students and workplace bullying victims). Indeed, past work shows that social workers

hold central the ideals that they can enact positive change on the world and aid those in need

(Fine & Teram, 2013; Olin, 2013).

We predicted that Duke MBA students might self-identify more with the villain profile

compared with the other known groups. Although there are many heroes among business leaders,

who lead the way with sustainable practices, charitable giving, and community engagement, past

work argues that those with MBAs are more likely to act in self-interested ways compared to

those without MBAs (Miller & Xu, 2019). In terms of perceptions, business people are typically

stereotyped as more narcissistic (Mark Young & Pinsky, 2006) and cold-hearted (Fiske, Cuddy,
 35

Glick, & Xu, 2002) than the general population, and MBAs may internalize those stereotypes as

self-perceptions. Acknowledging that no two business schools or business school students are

alike, we suggest that Duke MBA students might, on average, identity themselves as higher in

villain compared with the other groups. Note that we—as researchers—are not arguing anyone in

any sample is objectively higher in “villainy,” but instead suggest that our sample of Duke MBA

students may, relative to the other groups examined, self-identify more with pictured characters

who harm others for instrumental means, perhaps because they see instrumental harm as

important to achieving desired outcomes.

We predicted that those who stated they had experienced high levels of workplace

bullying would identify with the victim profile. Finally, we suggest that mTurk workers should

be relatively representative of the average person and would serve as a comparison condition for

the other three groups, especially in comparison to typical psychological study samples.

Research shows that mTurk workers are more demographically diverse than typical internet or

college student samples and provide reliable, high quality data (Buhrmester et al., 2011) from a

sample much larger than the typical university participant pool (Stewart et al., 2015).

Method

Participants. We had four groups of participants for this study: 94 UNC MSW students

(80 female, 12 male, 1 self-described, Mage = 26. 02, SD = 3.22), 37 Duke MBA students (12

female, 25 male, Mage = 28.08, SD = 3.90), 50 individuals who self-identified as having been

bullied in their place of work (28 female, 22 male, Mage = 33.90, SD = 9.76), and 101 mTurk

workers (53 female, 48 male, Mage = 36.53, SD = 10.31). While we strove for approximately

equal sample sizes across groups, we note that we experienced some difficulty in collecting our

in-person groups. We collected data from all MSW students to which we were able to gain
 36

access and note that our MBA sample was more difficult to collect than the other groups. Both

MSW and MBA students were collected on university campuses; anecdotally, the Duke MBA

students were generally unwilling to give a few moments to take the survey, some even telling

their peers—in front of our research assistants—not to participate as the task was a waste of

time. In stark contrast, we often did not have enough research assistants to accommodate all the

willing UNC MSW participants. Thus, it took longer to collect fewer MBA students in

comparison to MSW students. While we initially wanted to collect all samples in person, it was

very difficult to obtain both access and IRB approval to administer surveys to victim groups in

person, so we collected this group online using those who self-identified as having been bullied

at work.

Procedure. Each of the above groups were recruited either from on campus at large,

public universities, or were selected this task from mTurk. Once they consented to participate,

each participant took the MIPS. Unlike the other groups, those who identified as workplace-

bullying-victims first went through a screening process in which they filled out the Negative

Acts Questionnaire-Revised (Nam et al., 2010), which measures mistreatment at work with items

such as frequency of “Being humiliated or ridiculed in connection with your work” or “ Being

ignored or facing a hostile reaction when you approach.” Those who scored into the top category

of the scale (scores over 45; Notelaers & Einarsen, 2013) were categorized as victims of

workplace bullying and were included in the study in which they took the MIPS, which occurred

2-3 weeks after they took the screener.

 37

Results

Although a look at the results (Figure 6) reveals that all groups saw themselves self-

identified most as a hero—consistent with Study 1—we were interested in comparing subscale

scores across groups. To compare identification across these four groups, we ran a one-way

ANOVA. We found significant differences between groups for hero (F(3, 279) = 8.18, p < .001),

villain (F(3, 279) = 9.45, p < .001), victim (F(3, 279) = 3.27, p = .022), and beneficiary (F(3,

279) = 6.08, p = .001) groups. See Figure 6. Post-hoc tests reveal that, as predicted, MSW

students (M = 2.99, SE = 0.05) rate themselves as higher on our heroism profile when compared

to mTurk workers (M = 2.71, SE = 0.07, p = .004) and those who have been bullied (M = 2.54,

SE = 0.09, p < .001). MBA students (M = 2.92, SE = 0.07) also identify as more significantly

heroic than those who have been bullied (p = .01).

Consistent with hypotheses, MBA students (M =1.95, SE = .09) self-identify as

significantly higher on the villain profile when compared to MSW students (M =1.43, SE = .04, p

< .001), mTurk workers (M =1.58, SE = .06, p = .001), and those who have been bullied (M

=1.53, SE = .06, p < .001). Those who have been bullied (M =2.24, SE = .11) identify

significantly higher on the victim profile than MBA students (M =2.11, SE = .08, p = .02),

marginally higher than MSW students (M =2.01, SE = .06, p = 0.18), and non-significantly

different than mTurk workers (M = 2.11, SE = .06, p = .66).

Finally, those who have been bullied (M =2.14, SE = .07) identify significantly less on

the beneficiary profile than MSW (M =2.49, SE = .04, p < .001) or MBA (M = 2.41, SE = .08, p

= 0.05) students; mTurk workers were not significantly different in beneficiary identification (M

=2.30, SE = .06, p = 0.23). See supplementary material for full post hoc results. Finally, we also

ran a repeated-measures ANOVA to examine MIPS identification within each group. Please see
 38

supplemental materials for results of this test (Appendix D) and for visualizations of the

multidimensional scaling of each group (Appendix E).

3 MSW MBA
mTurk Bullied
Self-Identification

0
hero villain victim beneficiary

Figure 6. MIPS self-identification by known-groups. MSW: University of North Carolina

Masters of Social Work students at UNC; MBA: Duke University Masters of Business
Administration students; mTurk: sample of mTurk workers; Bullied: sample of people who self-
report being bullied. Error bars represent standard errors.
 39

Discussion

Study 4 provides further support that the MIPS measures moral self-identification.

Known groups responded to the MIPS in the anticipated fashion. Relative to the other groups,

UNC MSW students identified more as heroes, Duke MBA students identified more as villains,

and those who have been bullied identified more as victims. We find that multiple groups

identify with the beneficiary role, as both MSW students and MBA students identify

significantly higher as beneficiaries in comparison to those who have been bullied. This fits with

prior work, such as that on redemptive life narrative, that suggests that those who have achieved

success in life often craft narratives that include themes of gratefulness at having received help

from others in attaining that success (McAdams, 2013; McAdams et al., 2001).

It is again important to note that we, as researchers, are not claims that these groups are

necessarily captured by these moral roles, but rather than members of groups themselves self-

identify with these moral roles. It also bears noting again that these are relative differences across

groups: within each group everyone sees themselves as generally more like a hero than any other

role, consistent with the results from our previous studies.

General discussion

These four studies provide support for the MIPS as a measure of moral self-perception

for hero, villain, and moral patient identities. Study 1 provides initial evidence that the MIPS taps

into self-identification along two axes: as a moral agent (hero, villain) or patient (victim,

beneficiary) and positive (hero, beneficiary) or negative (villain, victim) valence. As potentially

expected, the beneficiary identity was less defined than the identities of hero, villain, and victim.

Rather than a full 2 x 2 space created by agency and valence, the scope of self-perceived moral

identity is more of a triangle, anchored by heroes, villains, and victims (Figure 5).
 40

Study 2 suggests that moral identity self-perceptions seem to be stable over time. In our

sample, participants self-rated moral identity was very similar approximately thirty days apart.

Studies 3a and 3b provide evidence of the MIPS’s convergent validity—those who identify as

heroes also self-rate as possessing more empathy and self-efficacy; those who identify as villains

rate themselves as higher on traits like narcissism; those who identify as victims also rate

themselves as higher on previously validated victimhood measures; beneficiaries are higher on

empathy. Further, hero identification is positively related to finding your moral identity

important on Aquino & Reed's (2002) measure. Study 4 examines target groups and finds that

UNC MSW students score comparatively higher on the hero profile, Duke MBA students scored

comparatively higher on the villain profile, and those who have been bullied at work scored

comparatively higher on the victim profile. While we believe these studies establish preliminary

support for the MIPS as a useful tool in moral psychology research, future work is needed to

further establish credibility and usefulness of the MIPS.

We hope that, in providing this new measure of moral identity, future work can examine

a broader sense of the moral world—beyond simple identifications of good vs. evil—using our

expanded measure that captures not only valence but also role as a moral agent or patient. This

measure expands upon previous measures related to moral identity (e.g., Aquino & Reed, 2002;

Barriga et al., 2001; Reimer & Wade-Stein, 2004), replicating prior work that we divide the

moral world up into good and evil, but demonstrating that the moral identification space includes

another component as well: moral agency and moral patiency. Most past work has examined this

“agent” side of moral identity—heroes and villains—but we can gain a fuller and more nuanced

view of the moral world if we also examine their counterparts—moral patients/recipients. The
 41

MIPS provides us with the ability to examine moral identity across these 2 dimensions of

valence (positive vs. negative) and agency (agent vs. patient).

Limitations & Future Directions

Taken together, these findings suggest that the MIPS is a measure of moral self-

perception that could be adapted to many areas of research. This scale not only measures moral

self-perception in a stable and valid way, but is also short and engaging, making it well suited for

multiple experimental designs. While this paper provides initial support for the use of MIPS in

moral self-perception research, future research should examine responding on this measure in a

wider variety of samples. For example, it might be useful to examine how an individual

identifies immediately after doing a good (“heroic) or bad (“villainous”) act, and to compare this

to how an individual responds without this moral or immoral prime. Further validating this

measure with other target groups that are typically seen as heroes, villains, victims, or

beneficiaries would also be helpful in identifying the applicability of this measure to broader

populations; such an examination would also be helpful in identifying any boundary conditions

or special cases to be aware of with this type of measure.

Future work could also expand upon this research by examining the MIPS directly in

relation to moral behavior. Study 4 reveals that those who work to actively improve the lives of

those in need—Master of Social Work students—identify more strongly as heroes than those

who have been bullied or the general mTurk worker, which may be influenced by their training

of helping those in need. Duke MBA students, in contrast, identify more strongly as villains

compared MSW students, reflecting the self-perception that they may have to harm others to

achieve their business goals. It bears noting again, that not all MBA students—or programs—are

alike and many to go pains to teach ethics, and emphasize pro-social goals, such as sustainability
 42

and positive social impact. Interestingly, both these MSW and MBA students—those who exert

moral agency regularly in day-to-day life—rate lower on victim self-identification than those

who have experienced bullying at work. While Study 4 provides a promising glimpse of how the

MIPS might relate to real-world behavior, future studies should further examine this relationship

between behavioral tasks in laboratories and real life and moral self-identification on this

measure.

Further research can also investigate the way social categorization interacts with

individual’s moral identification. For instance, recent work has shown that women compared to

men are more likely to be seen as a victim (T. Reynolds et al., 2020). Because of societal linked

gender roles, it’s possible that women are less likely to categorize themselves into either of the

agent roles and instead place themselves more into the patient roles. There is also potential that

other categorized populations like race and sexual orientation could present meaningful

differences in identification. When these stimuli were developed, we focused on including

gender diversity in our stimuli, but have not explicitly examined gender differences in response

patterns to the MIPS. Further, these stimuli could be further diversified to be more inclusive of

ages, race and ethnic backgrounds, and gender identities. Future research should work to increase

inclusivity of both stimuli and research participants.

Additionally, it will be useful to examine moral self-perception over a longer time scale.

While Study 2 suggests that these perceptions are likely to remain stable over time, this was only

tested within one time frame: approximately 30 days. In the future, examining moral self-

perception over a longer timeframe—months, or years—will be useful in determining how stable

these perceptions are throughout the lifespan. For example, perhaps these perceptions are quite

malleable through adolescence, but solidify in adulthood. Or perhaps they can be strongly
 43

molded by transformative or traumatic life events such as the birth of a child or the onset of a

medical condition. While there is certainly much to be explored in this area, this paper provides

initial evidence that MIPS will be a useful, stable tool by which moral self-perception can be

measured.

Finally, future work should seek to further understand how individuals see the moral

patient role—both in themselves and others. People clearly make distinctions between moral

agents with positive (hero) and negative (villain) valence. The distinctions between moral

patients, however, are less clear. While our work suggests that individuals think of themselves as

victims—our participants who experienced workplace bullying, for example—it is less murky

whether these victimhood perceptions are meaningfully different than those who received needed

help—beneficiaries. We see consistently that identifying as a beneficiary correlates well with the

hero role. Study 3 suggests that beneficiary identification is related valuing a positive moral self-

identity. In Study 4, we find that UNC MSW and Duke MBA students identify more strongly as

beneficiaries than do those who experienced workplace bullying. Many questions about

beneficiaries remain—is this role meaningfully separate from victimhood? Is a catch-all “moral

patient” a more appropriate model of the moral self-identification space? Do people lack a

distinct identification for beneficiaries and instead think that everyone gets help at some point in

time? The lack of literature on moral beneficiary leaves this a large, open question for future

work.

Conclusion

We hope this paper serves as a call to think about the broader nature of morality and

provides a measure that will be of use to future research. While much research examines the way

that individuals think about good and evil when making moral judgments of others, we argue that
 44

considering one’s self-perception as a hero, villain, beneficiary, and victim allows a fuller

understanding of moral identity.

 45

Acknowledgements
This work was supported by a grant from the Charles Koch Foundation to the Center for the
Science of Moral Understanding. And an NSF-GRF to A. Goranson. We thank Shawn Daley for
drawing the images in the MIPS.

Disclosure
The authors report no conflict of interests.
 46

References

Abele, A. E., & Wojciszke, B. (2013). The Big Two in social judgment and behavior.

Algoe, S. B., & Haidt, J. (2009). Witnessing excellence in action: The ‘other-praising’ emotions of

elevation, gratitude, and admiration. The Journal of Positive Psychology: Dedicated to Furthering

Research and Promoting Good Practice, 4(2), 105–127.

https://doi.org/10.1080/17439760802650519

Alicke, M. D. (2000). Culpable control and the psychology of blame. Psychological Bulletin, 126(4), 556–

574.

Alicke, M. D., & Govorun, O. (2005). The better-than-average effect. In The Self in Social Judgment (p.

304). Psychology Press.

Alicke, M. D., Klotz, M. L., Breitenbecher, D. L., Yurak, T. J., & et al. (1995). Personal contact,

individuation, and the better-than-average effect. Journal of Personality and Social Psychology,

68(5), 804–825. https://doi.org/10.1037/0022-3514.68.5.804

Aquino, K., Freeman, D., Reed, A., Lim, V. K. G., & Felps, W. (2009). Testing a social-cognitive model of

moral behavior: The interactive influence of situations and moral identity centrality. Journal of

Personality and Social Psychology, 97(1), 123–141. https://doi.org/10.1037/a0015406

Aquino, K., & Reed, A., II. (2002). The self-importance of moral identity. Journal of Personality and Social

Psychology, 83(6), 1423–1440. https://doi.org/10.1037//0022-3514.83.6.1423

Aujoulat, I., Luminet, O., & Deccache, A. (2007). The Perspective of Patients on Their Experience of

Powerlessness. Qualitative Health Research, 17(6), 772–785.

https://doi.org/10.1177/1049732307302665

Ayala, F. J. (2010). The difference of being human: Morality. Proceedings of the National Academy of

Sciences, 107(Supplement_2), 9015–9022. https://doi.org/10.1073/pnas.0914616107

 47

Barkan, R. (2015). Ethical dissonance, justifications, and moral behavior. Current Opinion in Psychology,

6.

Barraza, J. A., Alexander, V., Beavin, L. E., Terris, E. T., & Zak, P. J. (2015). The heart of the story:

Peripheral physiology during narrative exposure predicts charitable giving. Biological

Psychology, 105, 138–143. https://doi.org/10.1016/j.biopsycho.2015.01.008

Barriga, A. Q., Morrison, E. M., Liau, A. K., & Gibbs, J. C. (2001). Moral cognition: Explaining the gender

difference in antisocial behavior. Merrill-Palmer Quarterly (1982-), 532–562.

Bartlett, M. Y., & DeSteno, D. (2006). Gratitude and prosocial behavior: Helping when it costs you.

Psychological Science: A Journal of the American Psychological Society / APS, 17(4), 319–325.

https://doi.org/10.1111/j.1467-9280.2006.01705.x

Batson, C. D., Duncan, B. D., Ackerman, P., Buckley, T., & Birch, K. (1981). Is empathic emotion a source

of altruistic motivation? Journal of Personality and Social Psychology, 40(2), 290–302.

https://doi.org/10.1037/0022-3514.40.2.290

Baumeister, R. F. (1997). Evil. Inside human violence and cruelty. New York: W. H. Freeman and

company.

Beck, A. T., Ward, C. H., Mendelson, M., Mock, J. E., & Erbaugh, J. K. (1961). An inventory for measuring

depression. Archives of General Psychiatry, 4, 561–571.

Bernichon, T., Cook, K. E., & Brown, J. D. (2003). Seeking self-evaulative feedback: The interactive role of

global self-esteem and specific self-views. Journal of Personality and Social Psychology, 84(1),

194–204. https://doi.org/10.1037/0022-3514.84.1.194

Bernstein, M. H. (1998). On moral considerability: An essay on who morally matters. Oxford University

Press.

Blasi, A. (2005). Moral Character: A Psychological Approach.

 48

Brache, K., & Stockwell, T. (2011). Drinking patterns and risk behaviors associated with combined alcohol

and energy drink consumption in college drinkers. Addictive Behaviors, 36(12), 1133–1140.

https://doi.org/10.1016/j.addbeh.2011.07.003

Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon’s Mechanical Turk: A new source of

inexpensive, yet high-quality, data? Perspectives on Psychological Science, 6(1), 3–5.

https://doi.org/10.1177/1745691610393980

Cahill, L., & McGaugh, J. L. (1995). A Novel Demonstration of Enhanced Memory Associated with

Emotional Arousal. Consciousness and Cognition, 4(4), 410–421.

https://doi.org/10.1006/ccog.1995.1048

Chang, Y. P., Lin, Y. C., & Chen, L. H. (2012). Pay It Forward: Gratitude in Social Networks. Journal of

Happiness Studies, 13(5), 761–781. https://doi.org/10.1007/s10902-011-9289-z

Chiang, Y. S., & Takahashi, N. (2011). Network homophily and the evolution of the pay-it-forward

reciprocity. PLoS ONE, 6(12). https://doi.org/10.1371/journal.pone.0029188

Christie, R., & Geis, F. L. (2013). Studies in Machiavellianism. Academic Press.

Cialdini, R. B., Schaller, M., Houlihan, D., Arps, K., Fultz, J., & Beaman, A. L. (1987). Empathy-based

helping: Is it selflessly or selfishly motivated? Journal of Personality and Social Psychology, 52(4),

749–758.

Cohn, M. A., Fredrickson, B. L., Brown, S. L., Mikels, J. A., & Conway, A. M. (2009). Happiness Unpacked:

Positive Emotions Increase Life Satisfaction by Building Resilience. Emotion (Washington, D.C.),

9(3), 361–368. https://doi.org/10.1037/a0015952

Crimston, D., Bain, P. G., Hornsey, M. J., & Bastian, B. (2016). Moral expansiveness: Examining variability

in the extension of the moral world. Journal of Personality and Social Psychology, 111(4), 636–

653. https://doi.org/10.1037/pspp0000086
 49

Critcher, C. R., Inbar, Y., & Pizarro, D. (2012). How quick decisions illuminate moral character. Social

Psychological and Personality Science, 4(3), 308–315.

https://doi.org/10.1177/1948550612457688

Crocker, J., & Park, L. E. (2004). The Costly Pursuit of Self-Esteem. Psychological Bulletin, 130(3), 392–

414. https://doi.org/10.1037/0033-2909.130.3.392

Cunningham, G. B. (2005). The importance of a common in-group identity in ethnically diverse groups.

Group Dynamics: Theory, Research, and Practice, 9(4), 251–260. https://doi.org/10.1037/1089-

2699.9.4.251

Davis, M. H. (1983). The effects of dispositional empathy on emotional reactions and helping: A

multidimensional approach. Journal of Personality, 51(2), 167–184.

DeSteno, D., Bartlett, M. Y., Baumann, J., Williams, L. A., & Dickens, L. (2010). Gratitude as moral

sentiment: Emotion-guided cooperation in economic exchange. Emotion, 10(2), 289–293.

https://doi.org/10.1037/a0017883

Dongjie, X. I. E., Hao, L. U., & Yanjie, S. U. (2018). Pay-forward effect of resource allocation in

preschoolers: Role of theory of mind and empathy. Acta Psychologica Sinica, 50(9), 1018–1028.

Everett, J. A. C., & Kahane, G. (2020). Switching Tracks? Towards a Multidimensional Model of Utilitarian

Psychology. Trends in Cognitive Sciences, 24(2), 124–134.

https://doi.org/10.1016/j.tics.2019.11.012

Fagin-Jones, S., & Midlarsky, E. (2007). Courageous altruism: Personal and situational correlates of

rescue during the Holocaust. The Journal of Positive Psychology, 2(2), 136–147.

https://doi.org/10.1080/17439760701228979

Festinger, L. (1962). A Theory Of Cognitive Dissonance. Stanford University Press.

 50

Fine, M., & Teram, E. (2013). Overt and covert ways of responding to moral injustices in social work

practice: Heroes and mild-mannered social work bipeds. British Journal of Social Work, 43(7),

1312–1329.

Fiske, A. P., & Rai, T. S. (2014). Virtuous violence: Hurting and killing to create, sustain, end, and honor

social relationships. Cambridge University Press.

Fiske, S. T., Cuddy, A. J. C., Glick, P., & Xu, J. (2002). A model of (often mixed) stereotype content:

Competence and warmth respectively follow from perceived status and competition. Journal of

Personality and Social Psychology, 82(6), 878–902. https://doi.org/10.1037/0022-3514.82.6.878

Fohring, S. (2018). What’s in a word? Victims on ‘victim.’ International Review of Victimology, 24(2), 151–

164. https://doi.org/10.1177/0269758018755154

Franco, Z., & Zimbardo, P. (2006). The banality of heroism. Greater Good, 3(2), 30–35.

Fredrickson, B. L. (2001). The role of positive emotions in positive psychology: The broaden-and-build

theory of positive emotions. American Psychologist, 56(3), 218–226. https://doi.org/10.1037/0003-

066X.56.3.218

Gaertner, S. L., Dovidio, J. F., & Bachman, B. A. (1996). Revisiting the contact hypothesis: The induction of a

common ingroup identity. International Journal of Intercultural Relations, 20(3), 271–290.

https://doi.org/10.1016/0147-1767(96)00019-3

Gallup. (2014). About One in Four U.S. Households Victimized by Crime. Gallup.Com.

https://news.gallup.com/poll/179174/one-four-households-victimized-crime.aspx

Gentile, B., Miller, J. D., Hoffman, B. J., Reidy, D. E., Zeichner, A., & Campbell, W. K. (2013). A test of two

brief measures of grandiose narcissism: The Narcissistic Personality Inventory–13 and the

Narcissistic Personality Inventory-16. Psychological Assessment, 25(4), 1120–1136.

https://doi.org/10.1037/a0033192
 51

Giner-Sorolla, R., Hilton, D., Erb, H.-P., Durante, F., Flaßbeck, C., Fülöp, E., Mari, S., Petrović, N., Sekerdej,

M., Studzinska, A., Skitka, L. J., Washburn, A. N., & Zadora, A. (n.d.). Assigning moral roles within

the Second World War in Europe: National similarities, differences, and implications for group-level

moral representations. Asian Journal of Social Psychology, n/a(n/a).

https://doi.org/10.1111/ajsp.12450

Glasø, L., Matthiesen, S. B., Nielsen, M. B., & Einarsen, S. (2007). Do targets of workplace bullying portray a

general victim personality profile? Scandinavian Journal of Psychology, 48(4), 313–319.

https://doi.org/10.1111/j.1467-9450.2007.00554.x

Gollwitzer, M., & Denzler, M. (2009). What makes revenge sweet: Seeing the offender suffer or delivering a

message? Journal of Experimental Social Psychology, 45(4), 840–844.

https://doi.org/10.1016/j.jesp.2009.03.001

Goodwin, G. P., Piazza, J., & Rozin, P. (2014). Moral character predominates in person perception and

evaluation. Journal of Personality and Social Psychology, 106(1), 148–168.

https://doi.org/10.1037/a0034726

Gosling, S. D., Rentfrow, P. J., & Swann, W. B. (2003). A very brief measure of the Big-Five personality

domains. Journal of Research in Personality, 37(6), 504–528. https://doi.org/10.1016/S0092-

6566(03)00046-1

Gottschall, J. (2013). The Storytelling Animal: How Stories Make Us Human (1 edition). Mariner Books.

Graham, J., Haidt, J., Koleva, S., Motyl, M., Iyer, R., Wojcik, S. P., & Ditto, P. H. (2013). Moral Foundations

Theory. In Advances in Experimental Social Psychology (Vol. 47, pp. 55–130). Elsevier.

https://doi.org/10.1016/B978-0-12-407236-7.00002-4

Graham, J., Haidt, J., & Nosek, B. A. (2009). Liberals and conservatives rely on different sets of moral

foundations. Journal of Personality and Social Psychology, 96(5), 1029–1046.

https://doi.org/10.1037/a0015141
 52

Gray, K., & Keeney, J. E. (2015). Impure, or just weird? Scenario sampling bias raises questions about the

foundation of moral cognition. Social Psychological and Personality Science, 6(8), 859–868.

Gray, K., & Schein, C. (2012). Two minds vs. two philosophies: Mind perception defines morality and

dissolves the debate between deontology and utilitarianism. Review of Philosophy and Psychology,

3(3), 1–19. https://doi.org/10.1007/s13164-012-0112-5

Gray, K., Schein, C., & Ward, A. F. (2014). The myth of harmless wrongs in moral cognition: Automatic

dyadic completion from sin to suffering. Journal of Experimental Psychology: General, 143(4),

1600–1615. https://doi.org/10.1037/a0036149

Gray, K., & Wegner, D. M. (2009). Moral typecasting: Divergent perceptions of moral agents and moral

patients. Journal of Personality and Social Psychology, 96(3), 505–520.

https://doi.org/10.1037/a0013748

Gray, K., & Wegner, D. M. (2011a). Dimensions of moral emotions. Emotion Review, 3(3), 227–229.

Gray, K., & Wegner, D. M. (2011b). Morality takes two: Dyadic morality and mind perception. In M.

Mikulincer & P. R. Shaver (Eds.), The social psychology of morality: Exploring the causes of good and

evil. APA Press.

Gray, K., Young, L., & Waytz, A. (2012). Mind perception is the essence of morality. Psychological Inquiry,

23, 101–124. https://doi.org/10.1080/1047840x.2012.651387

Greene, J. D., & Haidt, J. (2002). How (and where) does moral judgment work? Trends in Cognitive Sciences,

6(12), 517–523.

Hafer, C. L. (2000). Do innocent victims threaten the belief in a just world? Evidence from a modified Stroop

task. Journal of Personality and Social Psychology, 79(2), 165–173. https://doi.org/10.1037/0022-

3514.79.2.165

Haidt, J. (2001). The emotional dog and its rational tail: A social intuitionist approach to moral judgment.

Psychological Review, 108(4), 814–834.

 53

Hammer, J. H., & Vogel, D. L. (2013). Assessing the utility of the willingness/prototype model in predicting

help-seeking decisions. Journal of Counseling Psychology, 60(1), 83–97.

https://doi.org/10.1037/a0030449

Han, H., Kim, J., Jeong, C., & Cohen, G. L. (2017). Attainable and Relevant Moral Exemplars Are More

Effective than Extraordinary Exemplars in Promoting Voluntary Service Engagement. Frontiers in

Psychology, 8. https://doi.org/10.3389/fpsyg.2017.00283

Hardy, S. A., & Carlo, G. (2005). Identity as a Source of Moral Motivation. Human Development, 48(4), 232–

256. https://doi.org/10.1159/000086859

Hartman, R., Blakey, W., & Gray, K. (in press). Deconstructing Moral Character Judgments. Current Opinion

in Psychology.

Harvey, J., Erdos, G., & Turnbull, L. (2009). How do we perceive heroes? Journal of Risk Research, 12(3–4),

313–327. https://doi.org/10.1080/13669870802519430

Heath, C., & Heath, D. (2007). Made to Stick: Why Some Ideas Survive and Others Die. Random House

Publishing Group.

Helzer, E. G., & Critcher, C. R. (2018). What do we evaluate when we evaluate moral character? In K. Gray &

J. Graham (Eds.), Atlas of Moral Psychology (pp. 99–107). Guilford Press.

Hertz, S. G., & Krettenauer, T. (2016). Does Moral Identity Effectively Predict Moral Behavior?: A Meta-

Analysis. Review of General Psychology, 20(2), 129–140. https://doi.org/10.1037/gpr0000062

Hester, N., & Gray, K. (2019). The moral psychology of raceless genderless strangers. Perspectives on

Psychological Science, 47.

Hofmann, W., Wisneski, D. C., Brandt, M. J., & Skitka, L. J. (2014). Morality in everyday life. Science,

345(6202), 1340–1343. https://doi.org/10.1126/science.1251560

Horita, Y., Takezawa, M., Kinjo, T., Nakawake, Y., & Masuda, N. (2016). Transient nature of cooperation by

pay-it-forward reciprocity. Scientific Reports, 6(January), 1–11. https://doi.org/10.1038/srep19471

 54

Hutcheson, F. (1726). An inquiry into the original of our ideas of beauty and virtue: In two treatises. J.

Darby...[and 8 others].

Isen, A. M., Clark, M., & Schwartz, M. F. (1976). Duration of the effect of good mood on helping:" Footprints

on the sands of time.". Journal of Personality and Social Psychology, 34(3), 385.

Janoff-Bulman, R., & Carnes, N. C. (2013). Surveying the moral landscape: Moral motives and group-based

moralities. Personality and Social Psychology Review, 17(3), 219–236.

https://doi.org/10.1177/1088868313480274

Janoff-Bulman, R., & Carnes, N. C. (2018). The model of moral motives. In K. Gray & J. Graham (Eds.), Atlas

of Moral Psychology. Guilford Publications.

Janoff-Bulman, R., Sheikh, S., & Baldacci, K. G. (2008). Mapping moral motives: Approach, avoidance, and

political orientation. Journal of Experimental Social Psychology, 44(4), 1091–1099.

https://doi.org/10.1016/j.jesp.2007.11.003

Jayawickreme, E., & Di Stefano, P. (2012). How Can We Study Heroism? Integrating Persons, Situations and

Communities. Political Psychology, 33(1), 165–178. https://doi.org/10.1111/j.1467-

9221.2011.00861.x

John, O. P., & Robins, R. W. (1994). Accuracy and bias in self-perception: Individual differences in self-

enhancement and the role of narcissism. Journal of Personality and Social Psychology, 66(1), 206.

Kahane, G., Everett, J. A. C., Earp, B. D., Caviola, L., Faber, N. S., Crockett, M. J., & Savulescu, J. (2018).

Beyond sacrificial harm: A two-dimensional model of utilitarian psychology. Psychological Review,

125(2), 131–164. https://doi.org/10.1037/rev0000093

Kidd, D. C., & Castano, E. (2013). Reading Literary Fiction Improves Theory of Mind. Science, 342(6156),

377–380. https://doi.org/10.1126/science.1239918
 55

Kohen, A., Langdon, M., & Riches, B. R. (2019). The Making of a Hero: Cultivating Empathy, Altruism, and

Heroic Imagination. Journal of Humanistic Psychology, 59(4), 617–633.

https://doi.org/10.1177/0022167817708064

Kohlberg, L. (1964). Development of moral character and moral ideology. Review of Child Development

Research, 1, 381–431.

Kruskal, J. B. (1964). Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis.

Psychometrika, 29(1), 1–27. https://doi.org/10.1007/BF02289565

Lee, S. (1963). X-Men. 1.

Lilgendahl, J. P. (2015). The dynamic role of identity processes in personality development: Theories,

patterns, and new directions. In The Oxford handbook of identity development (pp. 490–507).

Oxford University Press.

Luyckx, K., Teppers, E., Klimstra, T. A., & Rassart, J. (2014). Identity processes and personality traits and

types in adolescence: Directionality of effects and developmental trajectories. Developmental

Psychology, 50(8), 2144–2153. https://doi.org/10.1037/a0037256

Malle, B. F., & Horowitz, L. M. (1995). The puzzle of negative self-views: An exploration using the schema

concept. Journal of Personality and Social Psychology, 68(3), 470.

Mark Young, S., & Pinsky, D. (2006). Narcissism and celebrity. Journal of Research in Personality, 40(5), 463–

471. https://doi.org/10.1016/j.jrp.2006.05.005

Maurer, T. W., Pleck, J. H., & Rane, T. R. (2001). Parental Identity and Reflected-Appraisals: Measurement

and Gender Dynamics. Journal of Marriage and Family, 63(2), 309–321.

https://doi.org/10.1111/j.1741-3737.2001.00309.x

McAdams, D. P. (2013). The Redemptive Self: Stories Americans Live By-Revised and Expanded Edition.

Oxford University Press.

 56

McAdams, D. P., Diamond, A., de St. Aubin, E., & Mansfield, E. (1997). Stories of commitment: The

psychosocial construction of generative lives. Journal of Personality and Social Psychology, 72(3),

678–694. https://doi.org/10.1037/0022-3514.72.3.678

McAdams, D. P., Reynolds, J., Lewis, M., Patten, A. H., & Bowman, P. J. (2001). When bad things turn good

and good things turn bad: Sequences of redemption and contamination in life narrative and their

relation to psychosocial adaptation in midlife adults and in students. Personality and Social

Psychology Bulletin, 27(4), 474–485.

McKiernan, P. S., & Gary, L. (2009). Some HBS students adopt ethical code. Harvard Gazette.

https://news.harvard.edu/gazette/story/2009/06/some-hbs-students-adopt-ethical-code/

Midlarsky, E., Jones, S. F., & Corley, R. P. (2005). Personality Correlates of Heroic Rescue During the

Holocaust. Journal of Personality, 73(4), 907–934. https://doi.org/10.1111/j.1467-

6494.2005.00333.x

Miller, D., & Xu, X. (2019). MBA CEOs, Short-Term Management and Performance. Journal of Business

Ethics: JBE, 154(2), 285–300. http://dx.doi.org/10.1007/s10551-017-3450-5

Nam, W., Kim, J.-W., Kim, Y.-K., Koo, J.-W., & Park, C.-Y. (2010). The Reliability and Validity of the Negative

Acts Questionnaire-Revised (NAQ-R) for Nurses for the Assessment of Workplace Bullying. Korean

Journal of Occupational and Environmental Medicine, 22(2), 129–139.

https://doi.org/10.35371/kjoem.2010.22.2.129

Notelaers, G., & Einarsen, S. (2013). The world turns at 33 and 45: Defining simple cutoff scores for the

Negative Acts Questionnaire–Revised in a representative sample. European Journal of Work and

Organizational Psychology, 22(6), 670–682. https://doi.org/10.1080/1359432X.2012.690558

Nunnally, J. C. (1978). Psychometric Theory: 2d Ed. McGraw-Hill.

 57

Ok, E., Qian, Y., Strejcek, B., & Aquino, K. (2020). Signaling Virtuous Victimhood as Indicators of Dark Triad

Personalities. Journal of Personality and Social Psychology, 2(999).

https://doi.org/10.1037/pspp0000329

Olin, J. (2013). The public and the profession’s perception of social work. Columbia Social Work Review,

11(1), 92–102.

Pacini, R., Muir, F., & Epstein, S. (1998). Depressive Realism From the Perspective of Cognitive-Experiential

Self-Theory. Journal of Personality and Social Psychology, 74(4), 1056–1068.

Phillips, M. C., Meek, S. W., & Vendemia, J. M. C. (2011). Understanding the underlying structure of

deceptive behaviors. Personality and Individual Differences, 50(6), 783–789.

https://doi.org/10.1016/j.paid.2010.12.031

Pizarro, D., & Tannenbaum, D. (2011). Bringing character back: How the motivation to evaluate character

influences judgments of moral blame. In M. Mikulincer & P. R. Shaver (Eds.), The social psychology

of morality: Exploring the causes of good and evil (pp. 91–108). APA Press.

Pizarro, D., Uhlmann, E., & Bloom, P. (2003). Causal deviance and the attribution of moral responsibility.

Journal of Experimental Social Psychology, 39(6), 653–660. https://doi.org/10.1016/S0022-

1031(03)00041-6

Prentice, M., Jayawickreme, E., Hawkins, A., Hartley, A., Furr, R. M., & Fleeson, W. (2019). Morality as a

Basic Psychological Need. Social Psychological and Personality Science, 10(4), 449–460.

https://doi.org/10.1177/1948550618772011

Puleston, J. (2011). Online Research: Now & Next 2011 (Warc), Kings Fund, London, 1 March 2011.

International Journal of Market Research, 53(4), 557–562. https://doi.org/10.2501/IJMR-53-4-557-

562

Puleston, J. (2013). Gamification of Market Research. In Social Media, Sociality, and Survey Research (pp.

253–293). John Wiley & Sons, Ltd. https://doi.org/10.1002/9781118751534.ch11

 58

Rai, T. S., & Fiske, A. P. (2011). Moral psychology is relationship regulation: Moral motives for unity,

hierarchy, equality, and proportionality. Psychological Review, 118(1), 57–75.

https://doi.org/10.1037/a0021867

Reed, A., Aquino, K., & Levy, E. (2007). Moral Identity and Judgments of Charitable Behaviors. Journal of

Marketing, 71(1), 178–193. https://doi.org/10.1509/jmkg.71.1.178

Reeder, G. D. (2009). Mindreading: Judgments About Intentionality and Motives in Dispositional Inference.

Psychological Inquiry, 20(1), 1–18. https://doi.org/10.1080/10478400802615744

Reimer, K., & Wade-Stein, D. (2004). Moral Identity in Adolescence: Self and Other in Semantic Space.

Identity, 4(3), 229–249. https://doi.org/10.1207/s1532706xid0403_2

Reynolds, S. J., & Ceranic, T. L. (2007). The effects of moral judgment and moral identity on moral behavior:

An empirical examination of the moral individual. Journal of Applied Psychology, 92(6), 1610–1624.

https://doi.org/10.1037/0021-9010.92.6.1610

Reynolds, T., Howard, C., Sjåstad, H., Zhu, L., Okimoto, T. G., Baumeister, R. F., Aquino, K., & Kim, J. (2020).

Man up and take it: Gender bias in moral typecasting. Organizational Behavior and Human Decision

Processes, 161, 120–141. https://doi.org/10.1016/j.obhdp.2020.05.002

Rigby, K., & Slee, P. T. (1993). Dimensions of Interpersonal Relation Among Australian Children and

Implications for Psychological Well-Being. The Journal of Social Psychology, 133(1), 33–42.

https://doi.org/10.1080/00224545.1993.9712116

Rosenberg, M. (1965). Rosenberg self-esteem scale (RSE). Acceptance and Commitment Therapy. Measures

Package, 61(52), 18.

Rotter, J. B. (1966). Generalized expectancies for internal versus external control of reinforcement.

Psychological Monogrpahs: General and Applied, 80(1).

https://pdfs.semanticscholar.org/161c/b7ac92d7571042bb11ebdaaa1175be8079f8.pdf
 59

Rozin, P., Lowery, L., Imada, S., & Haidt, J. (1999). The CAD triad hypothesis: A mapping between three

moral emotions (contempt, anger, disgust) and three moral codes (community, autonomy,

divinity). Journal of Personality and Social Psychology, 76(4), 574–586.

https://doi.org/10.1037/0022-3514.76.4.574

Schein, C., & Gray, K. (2015). The unifying moral dyad: Liberals and conservatives share the same harm-

based moral template. Personality and Social Psychology Bulletin, 41(8), 1147–1163.

https://doi.org/10.1177/0146167215591501

Schein, C., & Gray, K. (2018). The Theory of Dyadic Morality: Reinventing moral judgment by redefining

harm. Personality and Social Psychology Review, 22(1), 32–70.

Sheldon, K. M., Elliot, A. J., Kim, Y., & Kasser, T. (2001). What is satisfying about satisfying events? Testing

10 candidate psychological needs. Journal of Personality and Social Psychology, 80(2), 325.

Sirgy, M. J., Gurel-Atay, E., Webb, D., Cicic, M., Husic, M., Ekici, A., Herrmann, A., Hegazy, I., Lee, D.-J., &

Johar, J. S. (2012). Linking Advertising, Materialism, and Life Satisfaction. Social Indicators Research,

107(1), 79–101. https://doi.org/10.1007/s11205-011-9829-2

Stanley, M. L., & De Brigard, F. (2019). Moral Memories and the Belief in the Good Self. Current Directions

in Psychological Science, 28(4), 387–391. https://doi.org/10.1177/0963721419847990

Stets, J. E., & Carter, M. J. (2011). The Moral Self: Applying Identity Theory. Social Psychology Quarterly,

74(2), 192–215. https://doi.org/10.1177/0190272511407621

Stewart, N., Ungemach, C., Harris, A. J. L., Bartels, D. M., Newell, B. R., Paolacci, G., & Chandler, J. (2015).

The Average Laboratory Samples a Population of 7,300 Amazon Mechanical Turk Workers.

Judgment and Decision Making, 16.

Strohminger, N., & Nichols, S. (2014). The essential moral self. Cognition, 131(1), 159–171.

https://doi.org/10.1016/j.cognition.2013.12.005
 60

Swearer, S. M., Song, S. Y., Cary, P. T., Eagle, J. W., & Mickelson, W. T. (2001). Psychosocial correlates in

bullying and victimization: The relationship between depression, anxiety, and bully/victim status.

Journal of Emotional Abuse, 2(2–3), 95–121.

Tang, W., Wu, D., Yang, F., Wang, C., Gong, W., Gray, K., & Tucker, J. D. (2021). How kindness can be

contagious in healthcare. Nature Medicine, 27(7), 1142–1144. https://doi.org/10.1038/s41591-021-

01401-x

Tannenbaum, D., Uhlmann, E. L., & Diermeier, D. (2011). Moral signals, public outrage, and immaterial

harms. Journal of Experimental Social Psychology, 47, 1249–1254.

Tappin, B. M., & McKay, R. T. (2016). The illusion of moral superiority. Social Psychological and Personality

Science, 1948550616673878.

Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: A social psychological perspective on mental

health. Psychological Bulletin, 103(2), 193–210. https://doi.org/10.1037/0033-2909.103.2.193

Tobacyk, J. J. (2004). A Revised Paranormal Belief Scale. International Journal of Transpersonal Studies,

23(1), 94–98. https://doi.org/10.24972/ijts.2004.23.1.94

Truman, J. L., & Morgan, R. E. (2014). Nonfatal Domestic Violence, 2003-2012. Bureau of Justice Statistics.

Tsai, C.-L., Chaichanasakul, A., Zhao, R., Flores, L. Y., & Lopez, S. J. (2014). Development and Validation of

the Strengths Self-Efficacy Scale (SSES). Journal of Career Assessment, 22(2), 221–232.

https://doi.org/10.1177/1069072713493761

Tsvetkova, M., & Macy, M. W. (2014). The social contagion of generosity. PLoS ONE, 9(2).

https://doi.org/10.1371/journal.pone.0087275

Van Lange, P. A. M., Otten, W., De Bruin, E. M. N., & Joireman, J. A. (1997). Development of Prosocial,

Individualistic, and Competitive Orientations: Theory and Preliminary Evidence. Journal of

Personality and Social Psychology, 73(4), 733–746.

 61

Vest, B. M. (2013). Citizen, Soldier, or Citizen-Soldier? Negotiating Identity in the US National Guard. Armed

Forces & Society, 39(4), 602–627. https://doi.org/10.1177/0095327X12457725

Victims of Sexual Violence: Statistics. (2020). RAINN. https://www.rainn.org/statistics/victims-sexual-

violence

Watkins, H. M., & Laham, S. M. (2020). The principle of discrimination: Investigating perceptions of soldiers.

Group Processes & Intergroup Relations, 23(1), 3–23. https://doi.org/10.1177/1368430218796277

Wojciszke, B., Baryla, W., Parzuchowski, M., Szymkow, A., & Abele, A. E. (2011). Self-esteem is dominated

by agentic over communal information. European Journal of Social Psychology, 41(5), 617–627.

https://doi.org/10.1002/ejsp.791
 62

Appendices for the Moral Identity Picture Scale (MIPS)

Appendix A: The Full MIPS

Instructions:
Next, you will view a series of photos and read some statements. Please indicate your
opinion about each statement by filling in the box that corresponds to the answer you would like
to give. Please be honest. Try not to let your response to one statement influence your response
to other statements. There are no correct or incorrect answers. Answer according to your own
feelings, rather than how you think "most people" would answer.

Number 1:

How much do you identify with each person above?

Not at all (1) Slightly (2) Moderately (3) Extremely (4)

Person on left    
Person on right    
 63

Number 2:

How much do you identify with each person above?

Not at all (1) Slightly (2) Moderately (3) Extremely (4)

Person on left    
Person on right    
 64

Number 3:

How much do you identify with each person above?

Not at all (1) Slightly (2) Moderately (3) Extremely (4)

Person on left    
Person on right    
 65

Number 4:

How much do you identify with each person above?

Not at all (1) Slightly (2) Moderately (3) Extremely (4)

Person on left    
Person on right    
 66

Number 5:

How much do you identify with each person above?

Not at all (1) Slightly (2) Moderately (3) Extremely (4)

Person on left    
Person on right    
 67

Number 6:

How much do you identify with each person above?

Not at all (1) Slightly (2) Moderately (3) Extremely (4)

Person on left    
Person on right    
 68

Number 7:

How much do you identify with each person above?

Not at all (1) Slightly (2) Moderately (3) Extremely (4)

Person on left    
Person on right    
 69

Number 8:

How much do you identify with each person above?

Not at all (1) Slightly (2) Moderately (3) Extremely (4)

Person on left    
Person on right    
 70

Number 9:

How much do you identify with each person above?

Not at all (1) Slightly (2) Moderately (3) Extremely (4)

Person on left    
Person on right    
 71

Number 10:

How much do you identify with each person above?

Not at all (1) Slightly (2) Moderately (3) Extremely (4)

Person on left    
Person on right    
 72

Number 11:

How much do you identify with each person above?

Not at all (1) Slightly (2) Moderately (3) Extremely (4)

Person on left    
Person on right    
 73

Number 12:

How much do you identify with each person above?

Not at all (1) Slightly (2) Moderately (3) Extremely (4)

Person on left    
Person on right    
 74

Number 13:

How much do you identify with each person above?

Not at all (1) Slightly (2) Moderately (3) Extremely (4)

Person on left    
Person on right    
 75

Number 14:

How much do you identify with each person above?

Not at all (1) Slightly (2) Moderately (3) Extremely (4)

Person on left    
Person on right    
 76

Number 15:

How much do you identify with each person above?

Not at all (1) Slightly (2) Moderately (3) Extremely (4)

Person on left    
Person on right    
 77

Number 16:

How much do you identify with each person above?

Not at all (1) Slightly (2) Moderately (3) Extremely (4)

Person on left    
Person on right    
 78

Appendix B: Scoring the MIPS

Hero: Average of (1/Left, 2/Right, 5/Right, 6/Left, 9/Right, 10/Right, 13/Left, 16/Right)

Villain: Average of (1/Right, 3/Left, 5/Left, 7/Left, 9/Left, 11/Left, 13/Right, 14/Left)
Victim: Average of (3/Right, 4/Left, 7/Right, 8/Right, 11/Right, 12/Right, 14/Right, 15/Right)
Beneficiary: Average of (2/Left, 4/Right, 6/Right, 8/Left, 10/Left, 12/Left, 15/Left, 16/Left)
 79

Appendix C: Summary Table of the MIPS

Picture File Name Brief Scene Description Identity of Left Identity of

right

MIPS_1_Hero_Vill Two women discuss intentions Hero Villain

to save or leave people in a fire.

MIPS_2_Bene_Hero Two women lost in the Beneficiary Hero

wilderness. One helps the other
find a way out.

MIPS_3_Vill_Vict One woman indicating taking Villain Victim

revenge on another woman.

MIPS_4_Vict_Bene One woman is trapped under a Victim Beneficiary

rock while the other woman is
being rescued from a cave.

MIPS_5_Vill_Hero One woman wants more Villain Hero

profits, so she wants to fire
someone. The other woman
says to fire her if anyone.

MIPS_6_Hero_Bene One woman listens and Hero Beneficiary

comforts another woman.

MIPS_7_Vill_Vict A boss tells another woman Villain Victim

that her idea is terrible.

MIPS_8_Bene_Vict Two young girls discuss Beneficiary Victim

whether or not they’ve been
adopted.

MIPS_9_Vill_Hero One man is about to press a Villain Hero

button to kill other people.
Another man states he’s going
to save them.

MIPS_10_Bene_Hero One soldier is pulling another Beneficiary Hero

soldier to safety.

MIPS_11_Vill_Vict One man seems like he’s going Villain Victim

to hurt another man.

MIPS_12_Bene_Vict One man is rescued from the Beneficiary Victim

top of a burning building while
the other one is left there.
 80

MIPS_13_Hero_Vill Two men on a park bench. One Hero Villain

wants the other to be kinder,
the other man says a different
man deserves to suffer.

MIPS_14_Vill_Vict One athlete taunts another Villain Victim

athlete after beating him in a
competition.

MIPS_15_Bene_Vill One man tells another that his Beneficiary Victim

wife has taken him back. The
other man says that his wife
just left him.

MIPS_16_Bene_Hero One man listens to another Beneficiary Hero

man’s problems.
 81

Appendix D: supplemental analyses for Study 4

Repeated Measures ANOVAs and Contrasts for Each Source Type in Study 4

UNC MSW ANOVA

Sum Sq Mean Sq NumDF DenDF F value Pr(>F)

125.40 41.80 3 279.00 354.72 0.00

UNC MSW Contrasts

profile Mean SE df lower.CL upper.CL Group

hero 2.99 0.05 236.15 2.89 3.08 A

villain 1.43 0.05 236.15 1.33 1.52 B

victim 2.01 0.05 236.15 1.92 2.10 C

beneficiary 2.49 0.05 236.15 2.40 2.58 D

Duke MBA ANOVA

Sum Sq Mean Sq NumDF DenDF F value Pr(>F)

27.50 9.17 3 111.00 56.42 0.00

Duke MBA Contrasts

profile Mean SE df lower.CL upper.CL Group

hero 2.92 0.08 101.81 2.75 3.08 A

villain 1.95 0.08 101.81 1.78 2.11 B

victim 1.84 0.08 101.81 1.68 2.01 B

beneficiary 2.41 0.08 101.81 2.25 2.58 C

Workplace Bullied ANOVA

Sum Sq Mean Sq NumDF DenDF F value Pr(>F)

26.20 8.73 3 147.00 30.57 0.00

 82

Workplace Bullied Contrasts

profile Mean SE df lower.CL upper.CL Group

hero 2.54 0.08 179.65 2.37 2.70 A

villain 1.54 0.08 179.65 1.37 1.70 B

victim 2.23 0.08 179.65 2.07 2.40 C

beneficiary 2.14 0.08 179.65 1.97 2.30 C

MTurk ANOVA
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)

66.34 22.11 3 300.00 73.45 0.00

MTurk Contrasts
profile Mean SE df lower.CL upper.CL Group

hero 2.70 0.06 339.80 2.58 2.83 A

villain 1.58 0.06 339.80 1.45 1.70 B

victim 2.11 0.06 339.80 1.99 2.23 C

beneficiary 2.30 0.06 339.80 2.18 2.42 C

 83

Appendix E: MDS results for Study 4

UNC MSW

Duke MBA
 84

Workplace Bullied