Moral Disagreement across Politics is Explained by Different Assumptions about who is Most Vulnerable to Harm

Abstract

Liberals and conservatives disagree about morality, but explaining this disagreement does not require different moral foundations. All people share a common harm-based mind, making moral judgments based on what seems to cause harm—but people make different assumptions of who or what is especially vulnerable to harm. Liberals and conservatives emphasize different victims. Across eight studies, we validate a brief face-valid assessment of assumptions of vulnerability (AoVs) across methodologies and samples, linking AoVs to scenario judgments, implicit attitudes, and charity behaviors. AoVs, especially about the Environment, the Othered, the Powerful, the Divine, help explain political disagreement about hot-button issues surrounding abortion, immigration, sacrilege, gay rights, polluting, race, and policing. Liberals seem to amplify differences in vulnerability, splitting the world into the very vulnerable versus the very invulnerable, while conservatives dampen differences, seeing all people as similarly vulnerable to harm. AoVs reveal common cognition—and potential common ground—among moral disagreement.

Full text

Assumptions of Vulnerability
1
Moral Disagreement across Politics is Explained by Different Assumptions about who is
Most Vulnerable to Harm
Jake Womick1,* Daniela Goya-Tocchetto2,* Nicolas Restrepo Ochoa3, Carlos Rebollar4, Kyra
Kapsaskis4, Samuel Pratt4, B. Keith Payne4, Stephen Vaisey5, Kurt Gray4
1California State University Bakersfield
2University at Buffalo, SUNY
3University of California, Davis
4University of North Carolina at Chapel Hill
5Duke University
*Co-first authors
Author Contributions
Jake Womick contributed to conceptualization, data analysis, and the majority of the writing.
Daniela Goya-Tocchetto contributed to conceptualization, design, data collection, data analysis,
and writing. Nicolas Restrepo Ochoa contributed to design and data analysis. Carlos Rebollar
contributed to data analysis. Kyra Kapsaskis contributed to design and data collection. Samuel
Pratt contributed to writing. B. Keith Payne contributed to conceptualization and writing.
Stephen Vaisey contributed to conceptualization and writing. Kurt Gray contributed to
conceptualization, design, funding, supervision, and writing. All materials, including pre-
registrations are available online:
https://osf.io/gubsv/?view_only=84a4ccb9f65c404d8d16805364887bdb

Assumptions of Vulnerability
2
Abstract
Liberals and conservatives disagree about morality, but explaining this disagreement does
not require different moral foundations. All people share a common harm-based mind, making
moral judgments based on what seems to cause harm—but people make different assumptions of
who or what is especially vulnerable to harm. Liberals and conservatives emphasize different
victims. Across eight studies, we validate a brief face-valid assessment of assumptions of
vulnerability (AoVs) across methodologies and samples, linking AoVs to scenario judgments,
implicit attitudes, and charity behaviors. AoVs, especially about the Environment, the Othered,
the Powerful, the Divine, help explain political disagreement about hot-button issues surrounding
abortion, immigration, sacrilege, gay rights, polluting, race, and policing. Liberals seem to
amplify differences in vulnerability, splitting the world into the very vulnerable versus the very
invulnerable, while conservatives dampen differences, seeing all people as similarly vulnerable
to harm. AoVs reveal common cognition—and potential common ground—among moral
disagreement.

Assumptions of Vulnerability
3
Statement of Limitations
The present research synthesized cross-sectional and multi-wave data, self-report and
implicit measures, and combined correlational with experimental research. This methodological
variety balanced many strengths and weaknesses across studies. Yet, there remain numerous
limitations to consider when evaluating the implications of this research. One limitation on the
inferences that can be drawn from these data is that all participants were located in the United
States. Although we were primarily concerned with understanding political differences in the
United States, and we employed one high-quality national sample, most of our data were not
nationally representative. These limitations leave generalizability to the U.S. population unclear,
and cross-cultural generalizability unknown. Additionally, we focused on how perceptions of the
vulnerability of moral patients can help us understand moral judgement and political differences.
Yet, patients are only one part of the equation of morality. Future research must also consider
perceptions of moral acts and moral patients for a fuller picture. Finally, the current research was
designed to understand liberal-conservative differences in morality broadly. Politics impinge on
a variety of (often intersectional) identities. Further research needs to examine the role of identity
in assumptions of vulnerability and their relationship to politics and moral judgment more
closely.

Assumptions of Vulnerability
4
Moral Disagreement across Politics is Explained by Different Assumptions about who is
Most Vulnerable to Harm
Moral disagreement between liberals and conservatives is obvious, but the drivers of this
moral disagreement are less obvious. Although it sometimes seems that liberals and
conservatives have different moral minds (e.g., Haidt & Joseph, 2004), research finds that moral
judgment revolves around concerns about harm and victimization (e.g., Ochoa, 2022; Schein &
Gray, 2015). Our shared harm-based mind does not rule out the existence of moral pluralism or
political disagreement, as some suggest (Graham et al., 2018), but instead parsimoniously
explains moral differences. Everyone cares about protecting vulnerable entities from harm, but
different people may make different assumptions about who or what is especially vulnerable to
victimization. In this paper, we demonstrate how different assumptions of vulnerability (AoVs)
explain political disagreement on many hot-button issues that are dividing society today. We also
discover overarching differences in AoVs between liberals and conservatives that connect to
broader ideological debates about vulnerability revolving around individual versus group
differences.
Grounding moral disagreement in assumptions of vulnerability helps to bridge a scientific
divide between the moral world and our moral mind. The moral world features clear differences
between liberals and conservatives, with people on the left and the right arguing strongly against
each other’s moral positions on the environment, immigration, the nature of racism, religious
freedoms, etc. Despite these differences, recent research highlights how all moral cognition
seems to rely on a harm-based template, with people condemning acts based on how harmful
they seem (Ochoa, 2022; Schein & Gray, 2018). If liberals and conservatives make different
assumptions about who or what is especially vulnerable to harm, then it could parsimoniously

Assumptions of Vulnerability
5
explain how moral differences can grow out of common cognitive processes, without needing to
posit distinct cognitive foundations.
In the current research, we examine assumptions of vulnerability towards a variety of
targets, explore how these AoVs differ across the political spectrum, and study how AoVs can
help explain political differences in both popular measures of morality (i.e., the Moral
Foundations Questionnaire; Graham et al., 2011) and reactions to many hot-button moral issues.
We first develop brief, face-valid items to measure AoVs and explore contentious debates by
applying them to four ad-hoc themes, the Environment, the Othered, the Powerful, and the
Divine. Our studies show that AoVs both predict important real-world behaviors and can be
manipulated to causally affect moral judgements. Importantly, appreciating the power of
assumptions of vulnerability to drive moral judgments helps create moral understanding across
political divides, because it allows partisans to recognize their common moral currency of harm
across issues. Before presenting these studies, we review the theoretical rationale that guided our
research.
A Popular Account of Political Differences within Moral Judgment
Disagreement between liberals and conservatives is obvious and exists by definition,
because liberalism
1
and conservativism represent different political philosophies and competing
moral commitments. Differences between liberalism and conservativism often revolve around
questions of tradition and hierarchy/inequality (Jost et al., 2003), with conservatives more likely
to support the continuation (i.e., the “conservation”) of traditional social structures, and also
1
Some may characterize this position more as progressivism, to distinguish from the ‘liberalism’ of classical
liberalism, but to stay consistent with past work in social psychology and with lay discussions of political
differences, we use the term of liberal and conservatives

Assumptions of Vulnerability
6
tolerate hierarchy/inequality because it helps promote social order. Psychological studies connect
these political differences to various interpersonal and cognitive dispositions (Womick & King,
2021), which may align more with different values (e.g., Jones et al., 2018; Schwartz et al.,
2010). The underlying roots of these political differences are often complex, ranging from
motivation (e.g., Jost et al., 2003), to identity (Hogg, 2007), existential concerns (e.g., Greenberg
et al., 1990), brain structure (Kanai et al., 2011), and even genetics (Lewis & Bates, 2014).
Moral Foundations Theory (MFT) is a popular theory that documents moral disagreement
between liberals and conservatives through a questionnaire called the Moral Foundations
Questionnaire (Graham et al., 2011). For example, liberals are more likely to condemn income
inequality (“it’s morally wrong that rich children inherit a lot of money while poor children
inherit nothing”) whereas conservative are more likely to support traditional gender roles (“men
and women each have different roles to play in society.”) These moral differences make sense
given the definitional differences between liberalism versus conservatives surrounding
hierarchy/inequality and tradition, but MFT ascribes these differences to deep and distinct
mechanisms in the mind (Haidt, 2012).
Moral foundations theory argues that liberals and conservatives possess different “moral
foundations” (Graham et al., 2009). The inspiration for this theory is a modular model of the
mind, which argues different evolutionary and cultural challenges are solved by our mind having
a toolbox of functionally separate mechanisms (Haidt & Joseph, 2004). In the past decades, the
number of these hypothesized moral mechanisms has increased from three (Rozin et al., 1999),
to four (Haidt & Joseph, 2004), to five (Graham et al., 2009)— care, fairness, loyalty, authority,
and purity—and now 6 (e.g., Atari et al., 2023).

Assumptions of Vulnerability
7
Moral foundation theory argues for a “conservative advantage” in morality (Haidt, 2012),
whereby right-leaning people enjoy a richer set of moral concerns. Liberals narrowly focus on
the individual-centric concerns of care and fairness (Haidt & Graham, 2007), but not the group-
centric concerns about loyalty, authority, and purity. In contrast, conservatives are argued to
focus somewhat less on care about care and fairness, but—unlike liberals—attend much more to
loyalty, authority, and purity (Graham et al., 2009). Calling liberals “individualistic” and
conservatives “group-focused” seems inconsistent with progressive claims for solidary with
labor movements (Gramlich, 2021), or conservative arguments against taxation and for self-
reliance (Fine, 1992), but these labels help explain political differences among the items selected
for the MFQ.
The biggest problem for MFT is that emerging evidence fails to support the idea of moral
foundations in their original “foundational” sense of different mechanism. Moral judgments
about various values are simply too interconnected to be separate modules (Graham et al., 2011),
and—as we will soon see—moral judgments are predicted by perceptions of harm. Leaving aside
claims about cognition, even descriptive claims about political differences are questionable. The
idea of a “conservative advantage” revolves around a set of moral judgments scenarios that were
explicitly designed to showcase right-leaning perspectives (Haidt, 2012): MFT loyalty items
focus on patriotism, MFT authority items focus on obedience to church leaders, and MFT purity
items focus on religious and sexual purity (Graham et al., 2009, 2011).
Although conservatives do care more about obeying conservative-leaning authority
figures (e.g., preachers and police), liberals care more about obeying authority figures when
those authority figures are liberal-leaning (e.g., environmentalists; Frimer et al., 2014). Likewise,
conservatives do care more about the sexual chastity of teenagers, but liberals care more about

Assumptions of Vulnerability
8
the purity of the natural world and health (Frimer et al., 2015, 2017). If you hear someone
discussing how they feel purified after juice cleansing at a yoga retreat, they are unlikely to be
conservative.
Some recent formulations of moral foundations have claimed only that they are
developmentally-prepared constructs subject to the forces of both nature and nurture, but this
formulation seem too broad, neither distinguishing it from other theories nor providing specific
testable hypothesis (Graham et al., 2018). How can one disprove the idea that nature and nurture
are both important to morality? At the same time, proponents of MFT simultaneously persist in
endorsing the “distinct mechanisms” idea (Graham et al., 2018), arguing that the distinctness of
moral concerns is illustrated by moral judgments of purity.
Purity is operationalized by bizarre scenarios including eating pizza off a corpse (Clifford
et al., 2015) and consensual incest (Haidt, 2001). However, a recent review of the literature
additionally found that purity was too messy and ill-defined to support any claims about its
specialness in cognition (Gray, et al., 2022). Other work shows that judgments of purity are
extremely correlated with perceptions of harm (Gray & Keeney, 2015; Ochoa, 2022). Indeed, a
close look at the data even from the original MFT papers casts doubt on claims of distinctness:
the overlap between moral foundations is often higher than overlap of items within moral
foundations (Graham et al., 2011). For example, the loyalty foundation correlates .88 with
authority, which is much higher than .54—the average amount that loyalty items load onto the
loyal foundation (Graham et al., 2011; Figure 3).
Despite the lack of evidence for distinct moral mechanisms in the mind, different political
parties use different keywords in rhetoric (Feinberg & Willer, 2019)—conservatives clamor for

Assumptions of Vulnerability
9
“freedom” and liberals for “fairness.” However, liberals do care about freedom (e.g., from the
effects of slavery; Kendi, 2019) and conservatives do care about fairness (e.g., seeing affirmative
action as unfair; Gramlich, 2023). This complexity suggests that rather than holding
fundamentally different moral minds, people on both ends of the political spectrum care about
freedom and fairness, albeit in different ways. Classic papers in moral psychology also warn us
about confusing people’s explicit moral rhetoric with the basic cognitive processes of moral
judgment (Haidt, 2001).
Not only do liberals and conservatives fundamentally appreciate similar values, research
on social cognition finds the idea of different cognitive foundations untenable (see Cameron et
al., 2015). Drilling down into human social cognitive architecture, it is clear that the mind is not
separated into functionally separate chunks (Barrett, 2009). Evolution does not solve the
challenges of complex social living by developing separate modules—which are costly—but
instead by developing broadly distributed networks that are shared across social judgment tasks
(Barrett & Russell, 2014). In the brain, moral judgments are not themselves distinguishable from
other social-affective judgments and so it is unlikely that different flavors of moral judgments
have separate evolved mechanisms (Barrett, 2013).
The Dyadic Morality Account of Political Differences in Morality
In contrast to Moral Foundations Theory, which argues for deep differences in the minds
between liberals and conservatives, we suggest that we all have a harm-based moral mind. This
idea is formalized as the Theory of Dyadic Morality (TDM; Schein & Gray, 2015, 2018) which
emphasizes similarities in moral judgment among all people. This theory argues that moral
judgments revolve around a common template of harm. Drawing from theories of cognitive

Assumptions of Vulnerability
10
categorization (Murphy, 2004; Rosch, 1978), TDM outlines how moral judgments are made by
comparing a potentially immoral act to a template (e.g., a schema, or exemplar set) involving
two interacting minds—a dyad of an intentional agent causing damage to a vulnerable patient
(A-P). This dyadic template revolves around harm, specifically interpersonal harm, or what you
might call the victimization or mistreatment of someone or something vulnerable.
The theory of dyadic morality—and the related “affective harm account” (Gray et al.,
2022), which connects TDM to affect—are domain-general theories of categorization (McHugh
et al., 2022), arguing that the processes of moral judgment can be understood by looking at other
theories of cognitive categorization. For example, judgments of how much someone is African
American are made by comparing people to a cognitive template of that race—what we often
call a stereotype (Gawronski et al., 2012). With morality, our cognitive template is grounded in
interpersonal harm or victimization. The key prediction of dyadic morality is that acts are
perceived as immoral to the extent that they are perceived as harmful. Importantly, this is not the
random “harm” of stubbing your toe, but the interpersonal harm of victimization. Acts can be
placed along a continuum based on how harmful/victimizing they seem, and a harm-based
template suggests that this continuum should very well predict people’s moral condemnation.
At the maximal end of the dyadic/victimizing continuum are acts where someone with a
deep capacity for intention perpetrates intense suffering on a very vulnerable target, like a
corporate CEO burning a child alive. These “most harmful” acts are typically seen as “most
immoral”—that is, they are most robustly categorized as belonging to the concept of
“immorality”. At the minimal end of the continuum are acts where someone unintentionally
causes mild suffering to an invulnerable agent, like a teething baby biting the calloused hand of
an ultimate fighting champion. These “least harmful” acts are typically seen as “least immoral.”

Assumptions of Vulnerability
11
Even if different kinds of moral violations seem categorically different (e.g., punching a kid vs.
masturbating to bizarre pornography), research reveals that people morally condemn acts based
on perceptions of harm (Ochoa, 2022).
One key claim of TDM is that harm is an intuitive perception, not a reasoned
rationalization (Chalmers, 1997; Epley & Waytz, 2009; Haslam et al., 2008; Waytz et al., 2010):
acts are wrong to the extent that they intuitively seem victimizing (Schein & Gray, 2018). In
people’s judgments, harm is not something that is “objectively present” or “objectively absent,”
but instead something that people “perceive” to be present or absent in varying degrees (Schein
& Gray, 2018). Consistent with an intuitionist view of the moral mind (Haidt, 2001), harm is
perceived rapidly and intuitively, such that acts are evaluated as more or less immoral relative to
the amount of intuitive perceived harm associated with it (Ochoa, 2022; Schein & Gray, 2018).
Evidence for the idea that a continuum of perceived harm underlies the continuum of moral
judgment comes from Ochoa (2022), who asked liberals and conservatives to provide their
intuitions about the harmfulness and immorality of scenarios drawn from the Moral Foundations
Questionnaire. He found that perceptions of harm/victimization predict moral wrongness
extremely well, even for acts that seem “objectively” harmless (Figure 1). In fact, every extant
study where researchers assess perceptions of harm on the same scale format as moral judgment
reveals that moral judgments are extremely well predicted by perceptions of harm (see Figure 1).

Assumptions of Vulnerability
12
Figure 1. Perceived harm predicts moral ratings independently of moral foundations. Data from
Ochoa (2022); figure reprinted with permission from Gray and colleagues (2022).
That harm is a matter of perception allows it to persist even in scenarios designed by
scientists to be “objectively harmless.” For example, people can—and do (Gray et al., 2014)—
see some amount of harm in a scenario of consensual incest designed to be harmless (Haidt,
2001). The perceptual nature of harm also allows it to vary between liberals and conservatives,
which may help explain their moral disagreement.
Although much emerging evidence supports the idea that moral cognition revolves
around a common question of harm, one limitation with existing work on this theory is how to
explain political moral differences. How can moral disagreement arise from a common cognitive
template (Graham, 2015)? If liberals and conservatives have a common moral mind—generally
relying upon perceptions of harm—how can they make such different judgments toward different
targets? Why are liberals so much more likely than conservatives to support #BlackLivesMatter
and advocate for protecting Black Americans (Horowitz et al., 2023), whereas conservatives are
more likely than liberals to support #BlueLivesMatter and advocate for protecting police
(Brown, 2017)? We suggest that political differences in moral judgment can be explained by

Assumptions of Vulnerability
13
different assumptions of vulnerability, which are assumptions about who or what can be
victimized.
Liberals and conservatives may have different perceptions about how much different
targets are especially susceptible to being victimized, and these underlying perceptions can help
explain different moral judgments about issues involving these targets. In this way, assumptions
of vulnerability can bridge the gap between a common harm-based moral mind and descriptive
differences in moral judgment. Everyone cares about protecting the vulnerable from harm, but
not everyone agrees on who or what is the most vulnerable to harm.
Assumptions of Vulnerability
How much can a CEO be a victim? In 2022, the CEO of the marketing company
Hypersocial posted a selfie of himself crying. He claimed that he was filled with grief because he
had to lay off two employees, but many people online argued that he was not genuinely
suffering—and that if he actually wanted to help them, he could have promoted them in his post,
rather than focusing on himself. The CEO argued it was “the most vulnerable thing” he had done
(Tabahriti, 2022), but was he really vulnerable?
Vulnerability is about being susceptible to victimization or harm, and, on the surface, it
seems easy to roughly determine whether an entity is vulnerable. Someone is vulnerable to
physical harm if they bleed after being struck or cut. Someone is vulnerable to emotional harm if
they cry after betrayals or insults. Yet, even these “objective” signs are only cues to actual
suffering. Traumatized individuals can experience grievous injuries and not feel pain, and people
without visible injuries can still feel chronic pain (e.g., Wall, 1979). Yet, vulnerability often
requires the mental guesswork involved in making assumptions. Because of the problem of other

Assumptions of Vulnerability
14
minds (Dretske, 1973; Putnam, 1975), we can never experience the suffering of others (Scarry,
1987) and so we are left to assume how much people are vulnerable to harm.
Assumptions of vulnerability (AoV) refer to how vulnerable (i.e., susceptible) people
assume (i.e., suppose to be the case) an entity is to harm, victimization, and mistreatment. We
use the word “assumption” because the problem of other minds makes it difficult to have certain
proof of someone’s vulnerability. Even when someone bleeds or shed tears (like our crying
CEO), people may make different assumptions about their true inner state.
AoVs are a subset of what Turiel calls “informational assumptions,” beliefs about the
nature of the world that help to shape differences in moral judgments (Turiel et al., 1987, p. 189).
Informational assumptions can revolve around many things, including any of the three elements
of the moral dyad, whether 1) intentional agents (iA; who is capable of intending harms?) or 2)
the method by which others are damaged (→ d; is witchcraft a viable way to harm someone?) or
3) the moral patients (vP; who is a legitimate target of damage?). From this view, AoVs are most
relevant to how much someone views a given target as a potential moral patient. Research
supporting the Theory of Dyadic Morality (Schein & Gray, 2018) already provides some
evidence of the importance of AoVs: people more harshly condemn acts that harm vulnerable
targets like children and those with developmental disabilities than typical adults (Gray &
Wegner, 2009).
As an informational assumption, AoVs—while being relevant to morality—are not
intrinsically moral or immoral. For example, there’s nothing inherently evil about children being
more or less susceptible to victimization. The conceptual separation between AoVs and morality
is important for scientifically explaining moral differences, especially since it helps escape the

Assumptions of Vulnerability
15
tautology found in other explanations of moral disagreements across politics. Consider the
newer, weaker formulation of MFT, which argues that political differences are explained by
liberals and conservatives relying on different developmentally-prepared moral concerns
(Graham et al., 2018). In plainer terms, this argument is that differences in moral judgments are
driven by differences in moral concerns, which are themselves the tendency to make certain
moral judgments. In even plainer terms, the MFT argument is that “differences in morality” are
driven by “differences in morality.” This tautology may seem unfairly simple, but consider that
the measurement of moral foundations (i.e., the MFQ; Graham et al., 2011) involves assessing
people’s moral judgments towards the specific moral concerns that moral foundations are
thought to explain. For instance, MFT argues that the “purity foundation” explains moral
concerns about sex and religion, but then uses ratings of sex and religion to assess reliance on the
purity foundation (see Schein & Gray, 2018, p. 47). Judgments about sex and religion should
clearly predict judgments about sex and religion.
To move beyond this tautology, the study of political differences in morality must ground
moral differences in constructs that are not intrinsically about morality. These may involve
intention (Cushman, 2008), causation (Hart & Honoré, 1985), foresight (Baldwin, 1979), norms
(Nichols, 2002), emotion (Greene et al., 2001) and identity (Hardy & Carlo, 2005). Again, while
these are all relevant factors that go into moral judgments, they are not themselves intrinsically
moral—instead, they exist at a lower or more basic level of analysis (Cushman & Young, 2011).
AoVs provide another construct that feeds into morality which is not itself ‘moral,’ and which
may therefore help explain political differences across morality. If liberals and conservatives
make different assumptions about who or what is especially vulnerable to victimization and

Assumptions of Vulnerability
16
mistreatment, and perceptions of victimization and mistreatment drive moral judgment, then
AoVs can be said to explain moral differences.
Political Differences in AoV
Whether undocumented immigrants are especially vulnerable to victimization depends on
who you ask. Liberals typically emphasize the vulnerability of people in this group, highlighting
their moral patiency, including the dangers of crossing the border, the pain that children might
face in internment camps, and the suffering that families endure when people are deported back
to their original countries (Buchman, 2022; Lind, 2019; McNulty, 2018). In contrast,
conservative depictions of undocumented immigrants are less likely to emphasize vulnerability,
highlighting how people in this group can be tough and threatening, and include drug-cartel
members who shrug off firefights with police, and potential rapists who are much stronger than
their victims (Miller, 2022; Conservative Zone, 2018; Fox News, 2017). We suggest that these
different AoVs help predict moral disagreement between liberals and conservatives about
immigration policy. If one is trying to protect vulnerable immigrants, opening borders seems like
a moral imperative. If one is trying to protect Americans from the predations of invulnerable
immigrants, then it seems like a moral imperative to close the borders. The importance of
assumptions of vulnerability to moral judgments also seems clear in the abortion debate, where
conservatives are more likely than liberals to see first-trimester fetuses as living souls that are
vulnerable to mistreatment (Akin, 2002).
In addition to disagreeing on the AoVs about fetuses and undocumented immigrants, we
suggest that liberals and conservatives likely make systematically different assumptions of
vulnerability about several targets, and these different AoVs help to predict differences in moral

Assumptions of Vulnerability
17
judgment. For example, liberals and conservatives respond differently to questions within the
Moral Foundations Questionnaire (Graham et al., 2011), and different clusters of AoV judgments
might help us understand these descriptive differences. Conservatives judge it more immoral
than liberals to disrespect authority figures like the police, and these differences may partially
arise because conservatives are more likely to believe that police officers are susceptible to
mistreatment. Anecdotal evidence supports this possibility. For instance, an article by the
conservative-leaning National Police Association, titled “The Disrespect is Getting Dangerous,”
argued that lack of respect for officers leads inevitably to them being harmed (Smith, 2019).
There may be as many political differences in AoVs as there are issues to disagree upon,
but it may be useful to outline some useful themes of AoVs that might cluster together and help
to make sense of current political debates. As with “moral foundations” these themes may
provide some concrete language for discussing political disagreements, as well as a rough
taxonomy that explicitly showcases how a harm-based moral mind allows for moral pluralism—
moral differences across people (Shweder & Haidt, 1993).
However, unlike the moral foundations, we make no claims that any AoV themes carve
nature at its joints to reveal natural moral kinds. Instead, they are merely ad hoc clusters of some
informational assumptions about the vulnerability of potential moral patients. We are not trying
to create an authoritative list of all possible areas of moral difference. Given past work on our
common harm-based moral mind and the scientific untenability of a modular mind, we are
especially not claiming that these themes represent any kind of distinct cognitive mechanism.
Instead, we are teasing out sets of similar and salient targets that might make sense of known
variability in moral judgment.

Assumptions of Vulnerability
18
To help explain political disagreement, we identify four ad-hoc themes: The
Environment, The Other, The Powerful, and The Divine. We expected two (The Environment
and The Othered) to have higher AoVs among liberals, and two (The Powerful and The Divine)
to have higher AoVs among conservatives.
The Environment. Environmental protection is rated as a top governmental priority by
85% of Democrats but only 39% of Republicans (Pew Research Center, 2020). This difference
may be explained by assumptions of vulnerability, with liberals seeing environmental entities
like coral reefs and rainforests to be more vulnerable to harm, as compared to conservatives. Of
course, The Environment may not be as vulnerable as human beings, but pro-environmental
movies (e.g., Fern Gully, Avatar) often depict trees and forests as sentient beings that can suffer
and who are in danger of being victimized. Similarly, children and adults alike
anthropomorphize the environment (e.g., “Mother Nature”; Gebhard et al., 2003) as a vulnerable
entity and ascribe mental capacities and moral worth to animals and ecosystems (Rottman et al.,
2021). Liberals are especially likely to include environmental targets in their circle of moral
concern (Waytz et al., 2019).
The Othered. Sociology frequently understands social identities as relational (i.e., groups
define themselves in relation to other groups) and organized through the lens of power structures
(Callero, 2003; Okolie, 2009). Critical theorists argue that, in America, the dominant group
against which others are defined and judged are White cis-gendered Christian men, and that
those who do not belong in this group are “othered” (e.g., Devos & Banaji, 2005). Whether or
not one endorses critical theories, the term “othered” provides a useful name for groups that are
outside the center of power in society. AoVs about The Othered include perceptions of how
vulnerable to harm are members of groups that are outside the boundaries of the dominant group

Assumptions of Vulnerability
19
(e.g., illegal immigrants, Muslims, and transgender people). Liberals seem to emphasize the
vulnerability of the othered, whereas conservatives are more likely to emphasize how the othered
are not victims, and perhaps even perpetrators (e.g., believing that transgender women would
rape other women if they were allowed to use the same restroom; Prestigiacomo, 2016).
The Powerful. At first blush, people in positions of power, like corporate leaders and state
troopers, are unlikely to be seen as vulnerable. In the dyadic framework of agents and patients,
The Powerful seem more suited to be agents, the intentional doers of morality, than patients who
suffer harm. However, many people—especially conservatives—seem to emphasize the
vulnerability of people in positions of power. For instance, while conservatives might recognize
the powerful status of institutional leaders, defenders, and historically privileged identities, they
might also believe that they are disadvantaged in some ways (Cooper, 2020; Takahashi &
Jefferson, 2021). Liberals on the other hand, might focus on the role of the powerful in the
oppression of marginalized identities and strip them of the capacity for victimhood (Lloyd,
2013).
The Divine. For some, God and Jesus are merely cultural ideals, and the Bible is just a
book. For others, God and Jesus are living beings with rich mental lives, and the Bible is much
more than a book (Altemeyer & Hunsberger, 1992). For instance, it is common practice in the
United State legal system to guarantee the veracity of testimony by swearing to tell the truth with
one’s hand on the bible (Rosefield, 2014), and it is a common Christian teaching that sins hurt
God (e.g., “And do not bring sorrow to God’s Holy Spirit by the way you live.” Ephesians 4:30).
Although it seems harder to victimize supernatural entities than people, clearly many see the
Bible as a living document, and view God as capable of suffering mistreatment. Given links
between politics and religion (Womick et al., 2021), we predict that conservatives will be more

Assumptions of Vulnerability
20
likely than liberals to see The Divine as vulnerable and this perception will help explain their
increased condemnation of violations like desecrating the Bible or blaspheming God.
Overarching Political Differences. In addition to examining political differences in AoVs
about the Environment, the Othered, the Powerful and the Divine, it is also useful to consider
how liberals and conservatives might generally see AoVs. To anticipate the pattern we reveal in
our data, we highlight how committed liberals often divide the world into the highly vulnerable
versus the highly invulnerable. Historically, the Karl Marx painted society as a struggle between
vulnerable workers and the invulnerable ruling overclass (Avineri, 1968), and today progressives
seem to cleave the world into “oppressed” and “oppressors”—true suffering victims and their
invulnerable victimizers (Freire, 2005). From the perspective of extreme liberals, questions of
vulnerability seem to revolve mostly around group membership and identity, with people in
Othered social categorizes being very vulnerable to victimization whereas the Powerful seem
completely invulnerable to victimization (Kendi, 2019).
In contrast, conservatives seem to reject this group-based dichotomy of oppressors versus
oppressed, seeing all people as relatively equally vulnerable to harm. Each of us is “created
equal” (Declaration of Independence, 1776) and—as human beings—can be harmed by violence.
Each of us also has hopes and dreams whose denial can bring disappointment. Consider the
conservative-led opposition to affirmative action, which was argued to unfairly victimize
qualified White college candidates over Black college candidates (Cohen, 1995). This critique
relies upon the idea that both these students—regardless of their race—can suffer (Hughes,
2024). This more individualistic understanding of vulnerability de-emphasizes statistical group-
based differences in potential victimization, suggesting that most anyone can be oppressor or
oppressed depending on people’s circumstances and choices (Krause, 2015). People on the

Assumptions of Vulnerability
21
political right may in fact reject oppressed vs. oppressor framings more broadly in favor of views
that emphasize competition and meritocracy. Within this competition framing, everyone is
vulnerable to losing but is participating in a fair game. Thus, in addition to left-right differences
in who is vulnerable to harm, we may see those on the political right lump different targets
together as equally vulnerable (rather than seeing the oppressed as highly vulnerable and
oppressors as completely invulnerable).
Current Research
The current research is divided into three broad sections. In the first section (Studies 1-3),
we focused on the idea of assumptions of vulnerability generally, measuring them towards a
wide variety of targets, and assessing how they relate to measurements of political ideology,
moral foundations, and moral judgments. In this initial section of this paper, we found that AoVs
generally help explain political differences in moral judgments across political issues. In the
second section (Studies 4-6), we explore whether AoVs about four specific themes—the
Environment, the Othered, the Powerful, the Divine—can help explain political disagreement
surrounding moral foundation questionnaire items and specific hot-button issues. We also
explore the intuitive nature of AoVs with an implicit measure. In the third and final set of studies
(Studies 7-8), we explored whether AoVs could predict specific moral behaviors (i.e., donations
to real charities) and whether they might be impacted by experimental manipulations to change
moral judgements.
Transparency and Openness
Our design, hypotheses, and analysis plan for all studies were preregistered at As
Predicted and can be found in the project’s OSF page:

Assumptions of Vulnerability
22
https://osf.io/gubsv/?view_only=84a4ccb9f65c404d8d16805364887bdb. For all studies, we
report all measures, manipulations, and exclusions. Sample size for all studies was intended to
maximize power and was determined based on availability of resources. Data, analysis code, and
experimental materials are available for download at the same OSF page. Data were analyzed
using R. The studies have received ethical approval from the IRB of the University of North
Carolina at Chapel Hill (“Judgement of Mind and Moral Decision Making”, Protocol 12-11585).
Section 1: Testing whether AoVs Help Explain Moral Differences
In the first set of studies (Pilot and Studies 1-3), we explored how perceptions of
vulnerability can help us understand moral judgments across people and politics. In a Pilot study,
we investigated whether assumptions of vulnerability regarding canonical (i.e., obvious) moral
patients and agents were correlated with moral judgements. We predicted that the more people
perceive a target as vulnerable, the more they will judge it immoral to harm them. Using
canonical patients and agents provides a first face-valid test of this hypothesis, as people likely
have different assumptions of vulnerability about targets that are obvious patients and obvious
agents. Importantly, in the Pilot study, we also developed a face-valid and reliable set of items to
capture general assumptions of vulnerability for different targets. We call these “AoV items.”
They are designed to measure assumptions of vulnerability to harm, mistreatment, and
victimization—related terms that people often use to describe when someone suffers from being
treated unjustly.
In Study 1, we continued to explore the value of using AoVs to predict moral judgments
by examining a wider variety of moral patients. We also examined the association between AoVs
and political orientation by including some targets that stand at the center of highly politicized
issues (e.g., a four-week-old fetus). Given the hypothesized association between AoVs, moral

Assumptions of Vulnerability
23
judgments, and political orientation, Study 2 tested whether AoVs about specific targets (e.g.,
undocumented immigrants) predicts moral judgments related to issues at the center of current
culture wars (e.g., questions of illegal immigration). This study provides the first test of the
explanatory power of AoVs in helping to understand moral differences between liberals and
conservatives regarding hot button issues at the center of politically polarized debates.
Some of the targets that exist at the center of moral political disagreements are less
obviously moral patients (e.g., the Bible and the American flag) than others (e.g., living,
breathing entities). Yet, our framework suggests that people who perceive these to be more
vulnerable to harm also subjectively view them legitimately as entities that are capable of
experiencing victimization and mistreatment. To test this possibility, in Study 3, we assessed
how much people attributed vulnerability, “aliveness,” and other mind-related capacities to these
targets, exploring how these ratings differed across the political spectrum. We expected that
perceiving these types of targets as vulnerable is associated with greater mind-attribution and
perceived aliveness.
Pilot Study
The first goal of this pilot study was developing questions to assess assumptions of
vulnerability. As with the four AoV themes explored above, we are not claiming that these are
the only or ultimate questions to measure AoVs. Rather, our aim was to develop some useful
face-valid questions that tapped the general sense of how much someone or something is
vulnerable to victimization and harm.
We also note that these items are worded to ask whether a target might be especially
vulnerable to mistreatment, victimization, or harm. This was done to avoid potential ceiling

Assumptions of Vulnerability
24
effects. In initial work on mind perception, questionnaires asked whether a variety of targets
were capable of feeling pain or fear (e.g., Gray et al., 2007), but subsequent work often found
ceiling effects for these questions when assessing normal human beings (e.g., Gray et al., 2011).
We can all agree that people are generally capable of feeling pain. To avoid these ceiling
concerns in our measurements of AoVs, we follow a tactic used in other work (e.g., Paap et al.,
2020), making the criterion more stringent. Rather than just asking whether someone or
something might be generally vulnerable, we asked whether targets might be especially
vulnerable.
The second goal of this pilot study was to explore some simple criterion validity of these
questions. Would responses to these AoV questions be higher for targets that generally seem
more vulnerable (e.g., an orphan) versus less vulnerable (e.g., a professional wrestlers)? We also
included some exploratory variables, reported in the Supplementary Materials, pp. 2-6.
Method
Participants. One hundred fifty two American participants completed the online survey
via CloudResearch’s MTurk toolkit. We excluded 13 participants who failed both attention
checks. The final sample (N = 139 participants, 51.8% male, 46.8% female, 1.4% other; Mage =
48.68 years) was 80.6% White, 11.5% Black, 3.6% Latinx/Hispanic, 2.2% American
Indian/Alaska Native, and 2.2% Asian. Education level spanned from no high-school degree
through doctoral degree and income spanned from under $25,000 to over $150,000 per year.
Modal education level was a bachelor’s degree (52.5%), and the median income was $50,000 -
$75,000 per year.

Assumptions of Vulnerability
25
Materials and Procedure. In this within-subjects study, participants rated AoVs for six
targets, three generally more vulnerable (Orphan Girl, Child with Down's Syndrome, and
Newborn Puppy) and three generally less vulnerable (Fortune 500 CEO, Professional Wrestler,
and Certified General Accountant).
AoV Items. Assumptions of vulnerability were assessed with three face valid items
assessing a target’s susceptibility to harm (henceforth AoV items). Each item was answered on a
5-point scale from (1) not at all vulnerable to (5) completely vulnerable. The items were,
“I believe that the following are especially vulnerable to being harmed”
“I think that the following are especially vulnerable to mistreatment”
“I feel that the following are especially vulnerable to victimization.”
Descriptive statistics and reliabilities for each target are shown in Table 1. Note again that
Cronbach’s alpha hinges on the number of items within a scale, and our measure of AoVs
includes only 3 items.
Results and Discussion
A dependent t-test was conducted to evaluate whether canonical patients were rated as
higher on AoV items than canonical agents. AoV scores were aggregated across patients and
across agents. As we predicted, participants rated canonical patients (M = 4.02, SD = 0.80) as
more vulnerable than canonical agents (M = 2.12, SD = 0.79), t(135) = 19.29, d = 2.39, p < .001.

Assumptions of Vulnerability
26
Table 1. Means, standard deviations, and Cronbach alpha for assumptions of vulnerability for
each canonical patient and agent (Pilot Study).
Assumptions of Vulnerability
α
M
SD
Orphan Girl
.86
4.13
0.82
Child with Down’s Syndrome
.84
4.15
0.82
Newborn Puppy
.85
3.80
1.07
Fortune 500 CEO
.85
1.85
0.91
Professional Wrestler
.77
2.42
1.03
Certified Accountant
.83
2.06
0.88
These initial results suggest that items used to assess AoV are indeed higher for targets
people generally assume to be more (vs. less) vulnerable to harm, demonstrating some
straightforward criterion validity. We next examined a broader range of targets and tested the
ability of AoV items to predict judgments of moral concern.
Study 1: AoVs and Moral Concern Toward Different Targets
A key claim about AoVs is that they should predict moral judgments. This first study
explored whether AoVs across a diverse set of targets predicted judgments of moral status. Of
course, moral judgments can vary across political affiliation, and so this study also investigated
whether AoVs differ across politics and whether these differences help predict differing moral
judgment. For example, would conservativism predict both higher AoVs and higher ratings of
moral status for fetuses? Importantly, despite liberals and conservatives assigning different AoV
ratings across targets, we predict that AoVs will similarly predict moral judgments. In other

Assumptions of Vulnerability
27
words, the predictive power of AoVs for morality should be similar across liberals and
conservatives.
Method
Participants
There were 483 American participants who completed the survey online via
CloudResearch. Our survey contained 7 attention checks and we only included in our final
sample participants who passed at least six of these checks (n = 102 excluded before analyses)
2
.
The final sample (N = 381 participants, 225 male, 154 female, 2 other; Mage = 35.96 years) was
78.0% White, 7.6% Latinx/Hispanic, 7.1% Black, 6.0% Asian, and 1.0% American
Indian/Alaska Native, 0.3% Native Hawaiian or other Pacific Islander. Education level spanned
from no high-school degree through doctoral degree and modal education level was a bachelor’s
degree (53.3%. Income ranged from under $25,000 to over $150,000 per year and the median
income was $50,000 - $75,000 per year. Sample size was determined in advance of the data
collection, per preregistration.
Measures
Assumptions of vulnerability. Participants rated the perceived vulnerability of a total of
twenty targets used in previous research on mind perception (Babies, chimpanzees, dogs, 4-
week-old fetuses, frogs, God, robots, adult women, corporations, trees, spirits of the dead,
patients in a persistent vegetative state (PVS), nuns, convicted criminals, civil rights leaders,
teachers, terrorists, 5-year-olds, dictators, famous singers) using the three AoV items from the
2
In all studies, we pre-registered more attention checks than reported. We realized that some of the attention checks
we included might be tapping more into education than attention and decided to exclude these from consideration
and analyze those who passed either all or all but one of the remaining attention checks.

Assumptions of Vulnerability
28
pilot study, which were averaged to give an overall AoV score for each target. See Supplemental
Materials (p. 7) for descriptive statistics and reliabilities for each target.
Judgments of moral concern. Participants evaluated the moral status of each of the
twenty targets with two items (5-point scale from 1 = Not at all to 5 = Completely): “I believe
that it is especially immoral to hurt the following” and “I think that the following are especially
deserving of moral protection.” For each target, we created an average composite using these two
items – see the Supplement (p. 7) for descriptive and reliabilities.
Political orientation. Participants answered the question “How would you describe your
political views overall?” on a 7-point scale from 1 = Extremely Liberal to 7 = Extremely
Conservative
3
(M = 3.57, SD = 1.75).
Results
Assumptions of vulnerability and judgments of moral concern
Overall AoV and judgments of moral concern are plotted in Figure 2. Consistent with
expectations and with the pilot study, canonical moral patients (e.g., babies and dogs) were seen
as more vulnerable than other entities (e.g., robots and corporations), and AoVs appeared to track
judgments of moral concern. A cross-classified, multi-level model that included random
intercepts for participants and for moral targets, and random slopes for moral targets, found that
AoVs predicted judgments of moral concern, β=.48, p < .001).
Do AoVs predict judgments of moral concern for both liberals and conservatives?
Although it can be difficult to argue for a null hypothesis, we explored this idea by conducting a
similar multi-level model, with an interaction effect between vulnerability and political
3
In all Studies, ideology was measured on a 1 (Extremely Conservative) to 7 (Extremely Liberal) scale. In all cases,
we reverse scored our ideology measure so that higher values indicate greater conservatism.

Assumptions of Vulnerability
29
orientation. This interaction effect was not significant, 90% CI = (-.02, .07), suggesting that
AoVs similarly predict judgments of moral concern across politics.
Figure 2. Relationship between vulnerability ratings and moral concern for each target (Study
1).
Differences in AoV across political orientation
We predicted liberals and conservatives would attribute different levels of vulnerability
to certain targets. We found significant differences for almost all targets (full results reported in
the Supplement, p. 11). Those on the left assigned more vulnerability (see Figure 3) (and moral
status, see Figure 4) to animals, robots, criminals, teachers, and civil rights leaders (r’s ranged

Assumptions of Vulnerability
30
from -.12 to -.28, all p’s < .02). Conversely, conservatives provided higher ratings on these
dimensions for fetuses, God, and corporations (r’s ranged from .13 to .41, all p’s <.003).
Figure 3. Assumptions of Vulnerability Across the Political Spectrum (Study 1).
Figure 4. Moral Concern Across the Political Spectrum (Study 1).

Assumptions of Vulnerability
31
Discussion
As predicted, AoVs were sensibly related to moral judgment. They explained significant
variance with overall judgments of moral status across a diverse set of targets and accounted for
significant variance in political differences in moral judgment, including toward fetuses and
animals. This study revealed that AoVs accounted for significant variance in the general
judgments of moral status toward targets, but would AoVs also help make sense of hot-button
issues?
Study 2: AoVs Predict Morality above Ideology: Hot Button Issues
Americans are divided on many contentious moral issues, and Study 2 explored whether
AoVs help explain this division. We examined six controversial issues and predicted that the
different moral judgments provided by liberals and conservatives on these issues would be
predicted by AoVs toward moral patients within these scenarios. More specifically, for issues
ranging from gay marriage and flag burning, we expected AoVs would predict moral judgments
made by liberals and conservatives, above and beyond political ideology. For example, increased
liberal (vs. conservative) AoVs towards undocumented immigrants should predict liberals
increased condemnation of detaining those illegally entering the country, above and beyond
generally being liberal.
Method
Participants
There were 482 American participants who completed the survey online via
CloudResearch. Our survey contained 6 attention checks. We only included in our final sample

Assumptions of Vulnerability
32
participants who passed at least five of these checks (n = 82 excluded before analyses). The final
sample (N = 400 participants, 215 female, 182 male, 3 other; Mage = 36.71 years) was 71.0%
White, 11.5% Black, 8.8% Asian, 7.8% Latinx/Hispanic, and 0.8% American Indian/Alaska
Native, 0.3% Native Hawaiian or other Pacific Islander. Education level spanned from no high-
school degree through doctoral degree and Modal education level was a bachelor’s degree
(50.8%.) Income spanned from under $25,000 to over $150,000 per year. Median income was
$50,000 - $75,000 per year. Sample size was determined in advance of the data collection, per
preregistration.
Design
Participants read six scenarios, three designed to tap hot-button issues typically judged to
be more immoral to liberals (detaining undocumented immigrants at the Mexican border, a florist
refusing to help gay people, and Congress voting to remove environment protections); and, three
designed to be more immoral to conservatives (a teenager disrespecting state troopers, burning
the American flag, and defacing a bible). Participants rated AoVs toward the potential moral
patients in each scenario and rated the immorality of the scenario before completing
demographics.
Measures
Assumptions of vulnerability. Using the same AoV items as the previous studies,
participants rated the perceived vulnerability of six targets: illegal immigrants, gay people, the
environment, state troopers, the American flag, and the Bible. Descriptives and reliabilities are
shown in the Supplement (p. 12).
Moral judgments. Participants evaluated the immorality of each scenario (see Table 2)
on a 5-point scale, ranging from (1) not at all immoral to (5) extremely immoral.

Assumptions of Vulnerability
33
Table 2. Moral Scenarios Used in Study 2.
Right-Leaning Topic
Left-Leaning Topic
Two state troopers are fired by a judge for using
excessive force on a teenager. The teenager was
acting aggressively and being verbally
disrespectful.
Illegal immigrants are forcibly detained for
long periods at the US-Mexico border.
A group of college students set fire to the
American flag.
A florist refuses to work with a gay couple
who is planning their wedding.
A modern art gallery allows people to publicly
deface the Bible by writing graffiti and swear
words on its pages.
Congress votes to remove environmental
protections.
Political Orientation. Participants completed the same measure of conservatism.
Results
Figure 5 shows AoV for all 6 targets across the political spectrum. Liberal participants
rated “typically liberal” targets as highly vulnerable to harm, and conservative participants rated
these as relatively less vulnerable. Liberal participants rated “typically conservative” targets as
highly invulnerable to harm, and conservative participants rated these as relatively more
vulnerable.

Assumptions of Vulnerability
34
Figure 5. Relationship between political ideology and average scores on AoV for each target
(Study 2).
To examine the link between AoV and moral judgment we first ran models with
conservatism as a predictor. Conservatism was negatively associated with the immorality of left-
leaning scenarios (β = -.47, R2 = .31, p < .001), and positively associated with the immorality of
right-leaning scenarios (β = .39, R2 = .21, p < .001). We then ran the same models, including
AoV scores as an additional predictor. As expected, conservatism still predicted the immorality
of left-leaning (β = -.29, p < .001) and right-leaning scenarios (β = .26, p < .001), but AoVs also
predicted the immorality of left-leaning (β = .51, p < .001) and right-leaning scenarios, (β = .41,
p < .001). AoV therefore provided additional explanatory power in predicting moral judgments

Assumptions of Vulnerability
35
of scenarios involving harm within left-leaning (∆R2 = .11) and right-leaning (∆R2 = .10) hot-
button issues.
Discussion
This study reveals that people’s AoVs—assumptions about who and what is vulnerable to
harm—help to explain moral differences across politics. Liberals see gay couples as more
vulnerable and so see it as more immoral to mistreat them. Likewise, conservatives see the
American flag as more vulnerable and so see it as more immoral to mistreat it. Of course, AoVs
represent only one construct that feeds into moral judgment, and so political affiliation—which
has connections to the constructs of compassion (Hirsh et al., 2010; Womick & King, 2021),
empathy (Morris, 2020), cognitive rigidity (Choma et al., 2014; Zmigrod et al., 2021), preference
for hierarchy (Ho et al., 2015), attitudes towards marginalized groups (Crawford et al., 2017),
group alliances (Pinsof et al., 2023), and beliefs about the nature of systems, society, and human
nature (Jost et al., 2003)—also continues to explain people’s judgments on these hot button
issues. Some of the targets included in Study 2 (i.e., the Bible) were inanimate objects. A
reasonable person may question whether assumptions of vulnerability apply to such objects. The
purpose of the next study was to address whether people legitimately view sacred objects like
these as alive and in turn vulnerable to harm.
Study 3: The Vulnerability of Sacred Objects: the Bible and the Flag
Assumptions of vulnerability help to explain political differences in moral judgment,
including (as we saw in Study 2) toward actions like burning the American flag or defacing the
Bible. One could argue that, as inanimate objects, these entities are objectively invulnerable to
harm and so people are somehow “wrong” about their AoVs. However, consistent with ample
evidence about the subjectivity of harm (Schein & Gray, 2018), we emphasize that AoVs are

Assumptions of Vulnerability
36
more subjective than one might expect, and so people can see at least some kind of authentic
vulnerability in these objects.
To explore this possibility, we assessed AoVs toward the sacred objects of the American
flag and the Bible, and also examined how much people perceived these objects as having the
qualities of living being with a mind—qualities that provide for both vulnerability to harm and
worthy of moral protection. Cultural evidence supports the idea of the US flag and the Bible are
seen as—at least somewhat—legitimately alive. The United States Flag Code—United States
Code, Title 4, Chapter 1, Section 8, Subsection j, notes “The flag represents a living country and
is itself considered a living thing.” Hebrews 4:12 says that the Word of God is “alive and active,
sharper than any two-edged sword.”
We expect conservatives endorse that these objects were alive more strongly than
liberals, but did not expect that right-leaning folks would see the flag and the Bible as alive as a
much as a human being, but instead as intermediately alive. AoVs form a continuum (as with the
severity of moral judgments), and we expect that AoV toward these targets for right leaning
participants will be somewhere between standard inanimate objects (e.g., a piece of concrete)
and canonical moral patient (e.g., babies). Importantly, those high on conservatism should view
these as more alive than those who are relatively liberal. Further, perceptions of the flag and
bible as being alive should explain moral disagreement about the treatment of these sacred
entities.
Method
Participants
We recruited 400 American participants via CloudResearch. As preregistered, we only
analyzed data from participants who passed more than 1 of three total attention checks (4

Assumptions of Vulnerability
37
excluded before analyses). The final sample (N = 396; 204 female, 191 male 2 other; Mage =
50.03 years) was 77.3% White, 12.4% Black, 5.6% Asian, 2.0% American Indian/Alaska Native,
1.8% other, and 1.0% Native Hawaiian or other Pacific Islander. Education level spanned from
no high-school degree through doctoral degree and modal education level was a bachelor’s
degree (51.2%). Income ranged from under $25,000 to over $150,000 per year. Modal education
level was a bachelor’s degree (51.2%) and the median income was $50,000 - $75,000 per year.
Sample size was determined in advance of the data collection, per preregistration.
Measures
Assumptions of vulnerability. As in previous studies, participants rated the 3 AoV items
for each of six targets, two highly vulnerable targets (a 5-year old child, a newborn baby), two
sacred objects (the American flag, the Bible) and two standard inanimate objects (a frying pan,
and a block of concrete). We aggregated items within these three categories.
Aliveness. The perceived aliveness of each target was measured by averaging three items:
“I feel that the following entities are, in some shape or form, alive” (1 = Not at all alive, 5 =
Completely alive), “I feel that the following entities are, in some shape or form, a living thing” (1
= Not at all a living thing, 5 = Completely a living thing), and “I feel that the following entities,
in some shape or form, have a mind” (1 = Does not have a mind at all, 5 = Completely has a
mind). Again, we created three composites across the three categories.
Political ideology. Participants completed the same measure of conservatism, (M = 3.70,
SD = 1.85).
Results
As predicted, people rated AoVs higher for obviously vulnerable targets of a baby and a
child M(SD) = 4.60 (0.72), than for standard inanimate objects like a frying pan and a block of

Assumptions of Vulnerability
38
concrete, M(SD) = 1.36 (0.61), with sacred objects of the American flag and the Bible sitting in
between, M(SD) = 2.38 (1.20). Following AoV ratings, ratings of aliveness were highest for
obviously vulnerable targets, M (SD) = 4.86 (0.40), lowest for standard inanimate objects, M
(SD) = 1.06 (0.34), with sacred objects in between, M (SD) = 1.40 (0.79).
Although ratings of AoVs and aliveness for obviously vulnerable targets (e.g., babies)
and standard inanimate objects (e.g., a block of concrete) did not vary by political affiliation,
ratings of the sacred objects did, with conservates seeing sacred objects higher in AoVs, r = .41,
p < .001, and higher in aliveness, r = .33, p < .001. Figure 6 shows perceptions of aliveness
plotted across the political spectrum. These results suggest that conservatives indeed see these
objects as at least somewhat alive, and supports the possibility that these ratings of aliveness help
to explain their higher AoVs toward sacred objects.
To examine whether the link between conservatism and vulnerability was due to their
mutual overlap with aliveness, we ran a mediation model. Conservatism positively predicted
aliveness, and aliveness in turn predicted vulnerability. The indirect effect of conservatism on
vulnerability was significant, b(SE) = 0.07 (0.01), 95% CI = [0.04, 0.10], and so was the direct
effect, b(SE) = 0.19(0.03), 95% CI = [0.14, 0.25]. Of course, we can not draw causal inferences
from these data, but these results show perceived vulnerability is not simply metaphorical or
symbolic. Although AoVs may connect with some broader cultural or symbolic considerations
(Gutierrez & Giner-Sorolla, 2007), perceptions of the Bible and flag as alive are not only
consistent with legal and biblical statements, but also help us understand why conservatism
promotes seeing the Bible and flag as vulnerable.

Assumptions of Vulnerability
39
Figure 6. Perceptions of Aliveness across the Political Spectrum.
Discussion
Conservatives saw the Bible and the flag as being somewhat alive, and this perception
helped explain ratings of AoVs toward them. We acknowledge that people—even
conservatives—viewed obviously vulnerable babies as more alive and vulnerable to harm than
the US flag and Bible. These results are reasonable and were expected, given that the flag is
made of fabric and the bible of paper, rather than of flesh. Nevertheless, that these ratings are not
at floor, and were even higher than standard inanimate objects, buttresses the ideas that people
truly can perceive the flag and Bible as vulnerable and alive, that these can legitimately
experience harm, and that these perceptions can help explain political moral differences. These
results also add to past work suggesting that violations of sacredness and purity are not
completely distinct from considerations of harm (Gray et al., 2022).

Assumptions of Vulnerability
40
Section 2: Four Themes of AoV Help Explain Moral Differences Across Politics
The second set of studies (Studies 4-6) explored four specific political themes that
repeatedly surface in debates: The Environment (e.g., coral reefs), The Othered (e.g., illegal
immigrants), The Powerful (e.g., business leaders), and The Divine (e.g., God). As reviewed
above, these themes were motivated by the liberal emphasis on protection of the environment
and of marginalized groups and past work on tying conservativism to respect for authority and
religion.
In Studies 4a and 4b (convenience and nationally representative samples), we combined
the AoV items with a subset of targets from the four political themes (The Environment, The
Other, The Powerful, and The Divine) and used factor analysis to develop a reliable scale of
potential differences. The scale included ratings of vulnerability to harm, mistreatment, and
victimization (as in previous studies) of twelve targets (three targets per each of the four AoV
political themes), which we expected to form a four-factor scale structure. To test the criterion
validity of these AoV ratings across these four themes, we also explored the connection between
AoV ratings and the Moral Foundations Questionnaire (MFQ; Graham et al., 2013). For
additional converging evidence (via method variance), we also developed a series of scenarios
involving moral transgressions that fall into each of the AoV themes—the AoV scenarios—and
tested whether AoV theme ratings predicted moral judgments on these AoV scenarios.
In Study 5, we further tested the validity of the AoV measure across these four themes.
We administered the scale across two time points to assess test-retest reliability. We included a
battery of other measures (e.g., moral foundations questionnaire, moral expansiveness scale, etc)
to assess convergent and divergent validity. Study 6 tested whether explicit assumptions of
vulnerability would to be reliably associated with implicit assumptions of vulnerability, given

Assumptions of Vulnerability
41
typical concerns about whether self-report measures connect with more intuitive judgments (e.g.,
Dovidio et al., 2001, p. 182), We use an adapted version of the Affect Misattribution Procedure
(AMP; Payne et al., 2005) to test this prediction.
Study 4a: Exploring Four Themes of AoV Targets
Developing Themes
American society is filled with disagreement, and we suggest that some of this
disagreement revolves around AoVs regarding some specific clusters of targets: the
Environment, the Othered, the Powerful, the Divine. As mentioned before, we argue strongly
against the idea of distinct mental modules and so do not suggest that these themes are natural
kinds or reflect some special mental mechanism. Instead, we suggest that AoVs surrounding
each theme might represent general informational assumptions about sets of similar entities,
which may help explain descriptive moral differences.
Connection to MFT
After first exploring the factor structure of the four themes, we then explore whether
these AoV themes have criterion validity. Do AoVs predict differences in ratings of moral
foundations? Although we take issue with the unique claims of MFT—especially its roots in a
modular mind (Haidt & Joseph, 2004)—we acknowledge that ratings of the MFQ do show some
differences across liberals and conservatives. Although it can be hard to empirically disentangle
all five foundations, and the foundations are operationalized in a way that leans conservative
(e.g., asking about loyalty towards preachers rather than union leaders; Frimer et al., 2014), it is
clear ratings of the MFQ—whatever those suggest—show differences across politics. Liberals
seem to give higher ratings of the “individualizing” factor combining “care and fairness” while

Assumptions of Vulnerability
42
conservatives prioritize the "binding” factor combining “loyalty, authority, and purity” (e.g.,
Graham et al., 2011). Here we explore whether AoVs toward the four themes help explain these
differences in MFQ above and beyond political affiliation.
More specifically, we test whether the Othered and the Environment predict judgments
about care/fairness, while the Powerful and the Divine predict judgments about
loyalty/authority/purity. Although there may be some broader conceptual links between these
themes and these sets of moral values, we think an important connection lies in the specific items
used in the measurement of MFT. Conservative-leaning MFQ items measuring
loyalty/authority/purity focus on targets likely to be higher in AoVs for conservatives (e.g., the
Bible), and the liberal-leaning care/fairness items focus on targets likely to be higher in AoVs for
liberals (e.g., endangered species; Graham et al., 2009).
Moral Scenarios
Finally, this study also explores whether AoV ratings across four themes also predict
moral judgments of scenarios that contain the entities from each of these themes (e.g., coral reefs
in the Environment) as victims of harm. This methodology allows us to connect AoV ratings to
moral judgments of specific and relevant actions.
Method
Participants
There were 1008 American participants who completed the online survey via
CloudResearch. We only analyzed data from participants who passed more than four of 6 total
attention checks (76 excluded before analyses). The final sample (N = 932; 496 female, 430
male, 5 other; Mage = 45.16 years) was 80.0% White, 8.8% Black, 5% Hispanic, 4.6% Asian,

Assumptions of Vulnerability
43
2.1% American Indian/Alaska Native, and 2.1% Native Hawaiian or other Pacific Islander.
Modal education level was a bachelor’s degree. Income ranged from under $25,000 to over
$150,000 per year. Median income was $50,000 - $75,000 per year. Sample size was determined
in advance of the data collection, per preregistration.
Measures
Assumptions of Vulnerability. For each of the four AoV groups, participants rated the 3
AoV items (“I believe that the following are especially vulnerable to being harmed,” “I think that
the following are especially vulnerable to mistreatment,” and “I feel that the following are
especially vulnerable to victimization”) for three targets on a scale from 1 (Not at all vulnerable)
to 5 (completely vulnerable). Targets can be found in Table 4. Descriptive statistics and
reliabilities for each AoV group are shown in the Supplement (p. 13). For analyses, we first
averaged across the 3-items for each target, creating a mean composite variable representing
vulnerability for each target. We then created a mean-composite for each of the four categories
by aggregating these across all targets.
Moral Foundations. Participants completed the 20 item Moral Foundations
Questionnaire (Graham et al., 2008), which consists of 10 moral relevance items (e.g., “When
you decide something is right or wrong, to what extent are the following considerations relevant
to your thinking?”) and 10 moral judgments items (e.g., “Compassion for those who are
suffering is the most crucial virtue.”) on a scale from 1 (not at all relevant / strongly disagree) to
7 (extremely relevant, strongly agree). For analyses, we aggregated all items pertaining to care
and fairness, M (SD) = 4.83 (0.70), α = .77 and all items pertaining to loyalty, authority, and
purity, M (SD) = 3.51 (1.02), α = .90.

Assumptions of Vulnerability
44
AoV Target Moral Scenarios. For each of the hypothesized factors, we created two
scenarios (shown in Table 3) and asked respondents to rate how immoral each was on a scale
from 1 = Not at all immoral to 5 = Extremely immoral. The two scenarios associated with each
factor were aggregated to form measures of perceived wrongness for acts involving harm to the
Othered (M = 2.66, SD = 1.27, inter-item r = .75), the Environment (M = 3.59, SD = 1.07, inter-
item r = .50), the Powerful (M = 2.80, SD = 1.17, inter-item r = .73), and the Divine (M = 2.86,
SD = 1.45, inter-item r = .92). These were used to assess convergent validity of the AoV scale.
Table 3. Scenarios used to Assess Criterion Validity
The Othered
The Environment
The Powerful
The Divine
Someone makes a
transgender person
use the bathroom of
the sex they were
born, rather than their
current sex.
Some people teach
their kids that global
warming is a myth.
Someone distributes
anarchist guides to
teach children to rise
against the
authorities.
Someone burns a
bible for fun.
Someone reports a
family of illegal
immigrants to the
police.
A recycling company
discards all the
plastic that should
have been recycled in
a forest nearby.
Someone shows zero
respect when talking
to a police officer.
Someone uses a
Christian cross for
firewood.

Assumptions of Vulnerability
45
Political Ideology. Participants completed the same measure of conservatism, M (SD) =
3.55 (1.74).
Results
Factor Analyses
We first employed exploratory factor analysis to empirically assess whether four, fewer,
or more factors were warranted to explain variation across the items used to assess our four ad
hoc themes. We explored solutions ranging from one to six factors using maximum likelihood
and an oblimin rotation. Results provided support for the four-factor solution. Only the first four
factors had eigenvalues > 1.00 (ranging from 1.27 to 4.08), cumulatively explaining 80% of the
variance in these items. Figure 7 shows the fit measures and BIC for exploratory analyses
ranging from one to six factors. The RMSEA and TLI only reach adequate levels at four factors
and beyond, suggesting that simpler solutions are not acceptable. The BIC value improves up to
four factors and suggests that a five or six-factor solution overfits the sample data. Overall, all
indicators suggest that the four-factor solution is optimal. Primary and cross-loadings from the
four-factor solution are presented in Table 4. All primary loadings were above .73 and no cross-
loadings were above .10. Table 5 shows relationships among these factors.

Assumptions of Vulnerability
46
Figure 7. Exploring Fit for 1 to 6 Factor Solutions.

Assumptions of Vulnerability
47
Table 4. Exploratory Factor Analysis of the AoV Scale, Study 4a
Factor 1
Environment
Factor 2
Divine
Factor 3
Othered
Factor 4
Authority
Rainforest
.97
.00
-.01
.00
Reef
.91
-.04
-.03
.01
Earth
.88
.04
.06
-.01
Jesus
.01
.95
.01
-.02
God
-.01
.95
.00
-.02
Bible
.00
.81
-.03
.07
Undoc. Immigrant
.00
-.02
.93
-.03
Muslim
.00
.03
.89
.05
Transgender
.02
.02
.87
-.02
Authority
.02
-.02
-.01
.98
Trooper
.03
.03
-.05
.81
Corporate Leader
-.06
.02
.07
.73
M (SD) α
4.06 (0.97) .95
1.78 (1.04) .93
3.87 (0.96) .93
2.45 (0.92) .87
Note. Analysis was conducted using a composite of 3 AoV items for each of the 12 targets.
Descriptive and reliabilities are for the aggregate composite of the factor in each column.

Assumptions of Vulnerability
48
Political Ideology and Assumptions of Vulnerability
Correlations between conservatism and AoVs are shown in Table 5 and conform with
previous findings: Liberals view the Othered and the Environment as more vulnerable than do
conservatives, and conservatives rated the Powerful and the Divine as more vulnerable than do
liberals. As in Study 3, we plotted assumptions of vulnerability across the political spectrum,
shown in Figure 8. Replicating those findings, political centrists showed a moderate distinction
between how vulnerable they viewed the Othered/Environment vs. the Powerful/Divine. By
contrast, extreme liberals showed a substantial level of separation regarding how vulnerable they
perceived the Othered and Environment compared to the Powerful and Divine. On the far-right
end of the Figure, those who strongly endorsed conservatism showed a much smaller distinction
between how much they viewed all clusters as vulnerable to harm.
Figure 8. Assumptions of Vulnerability Across the Political Spectrum (Study 4a)

Assumptions of Vulnerability
49
Political Ideology and Moral Foundations
We next examined the relationships between these four factors, care/fairness,
loyalty/authority/purity, and political ideology. Correlations are shown in Table 5. Consistent
with past work, conservatism was negatively associated with care/fairness, and was positively
related to loyalty/authority/purity. As we expected, care/fairness was positively related to AoVs
for the Othered and the Environment. Loyalty/authority/purity was positively related to higher
AoVs for the Powerful and the Divine. One asymmetrical finding was that care/fairness were
unrelated to AoVs for the Powerful and Divine, but loyalty/authority/purity were negatively
associated with AoVs for the Environment and the Othered. Future research centering on MFT
might explore these differences.
Table 5. Correlations Among Variables, Study 4a
Variable
1
2
3
4
5
6
7
1. MFQ Care/Fairness
--
2. MFQ LAP
.02
--
3. Conservatism
-.32**
.56**
--
4. AOV Environment
.34**
-.20**
-.36**
--
5. AOV Other
.41**
-.26**
-.47**
.54**
--
6. AOV Powerful
-.03
.38**
.29**
.03
.10*
--
7. AOV Divine
.02
.41**
.26**
.01
.02
.31**
--
Note. N = 871. ** p < .001, * p < .01. LAP = Loyalty/Authority/Purity
We next conducted a hierarchical regression analysis to examine whether AoV scores add
additional explanatory power in predicting the two groupings from the MFQ, over and above

Assumptions of Vulnerability
50
political ideology. All predictor and outcome variables were standardized. Conservatism was
entered as the first step in each hierarchical regression model. AoV factors were added in a
second step to examine their predictive value after controlling for ideology. AoVs for the
Environment (β = .16, p < .01) and the Othered (β = .24, p < .01) explained additional significant
variance in care/fairness (second step ∆R2 = .11, p < .001); and AoVs for the Powerful (β = .20, p
< .01) and the Divine (β = .26, p < .01) explained additional variance in loyalty/authority/purity
(second step ∆R2 = .12, p < .001), above and beyond political ideology. Full results for these
models can be found in the Supplement (Table S12, p 15).
Moral Scenarios
We next examined if the AoV factors predict moral judgements for vulnerable targets
related to each factor. We computed four regression models. The dependent variables were
immorality ratings for the scenarios involving canonical patients associated with each factor. We
hierarchically regressed these immorality ratings on the vulnerability scores for the
corresponding factor (on the second step), after controlling for conservatism, as well as
care/fairness and loyalty/authority/purity on the first step. As predicted, the vulnerability scores
for the corresponding factor are predictive of the wrongness attributed to scenarios: AoVs for the
Environment predicted moral judgements in the environmental scenario, β = .35, p < .001,
second step ∆R2 = .10, p < .001; AoVs for the Othered, β = .28, p < .001, predicted moral
judgements in the othered scenario, second step ∆R2 = .05, p < .001; AoVs for the Powerful, β =
.15, p < .001, second step ∆R2 = .02, p < .001 explained significant variance in the powerful
scenario; and, AoVs for the Divine, β = .20, p < .001, second step ∆R2 = .04, p < .001
significantly predicted judgements in the divine scenario. Again, these effects were significant,
even after accounting for political orientation and moral foundations on the first step. For a more

Assumptions of Vulnerability
51
detailed presentation of these results, including information for the first step variables and model
R2s, see Supplemental Table S13, p 16.
Discussion
In this study, we explored the idea of AoV themes. Exploratory factor analysis provided
support for the clustering of our ad hoc categories. As predicted, liberals were more likely to rate
AoVs higher for the Othered and the Environment, and conservatives were more likely to rate
AoVs higher for the Powerful and the Divine. These different ratings in AoVs across the four
clusters helps to explain differences in moral judgment between liberals and conservatives across
the MFQ and specific scenarios, even when controlling for political ideology.
Interestingly, the pattern of these AoVs results suggested a general trend between liberals
and conservatives. Liberals amplify differences between the more and less powerful, whereas
conservatives diminish differences between the more and less powerful. Although an
authoritative explanation behind this effect may itself be a whole other paper, one can speculate.
One dominant narrative of the far left is the tension between invulnerable oppressors and the
extremely vulnerable oppressed (Pluckrose & Lindsay, 2020), with liberals emphasizing how
some groups of people are much more likely to suffer than others (Pinsof et al., 2023). This is
consistent with the amplification of differences in vulnerability we observed. In contrast,
conservatives often emphasize the narrative that all individuals are created fundamentally equal,
no matter what group to which they are born (e.g., Burke, 2017). All people can ultimately be
harmed and suffer—if both a rich white and a poor Black person are cut, both will bleed and feel
pain.
Study 4b: Four AoV Themes with a Unique National Sample

Assumptions of Vulnerability
52
In this study, we sought to replicate the key results from Study 4a using a high-quality
national sample, exploring whether we obtained the same patterns of AoVs of the four themes
and political orientation. We compared how this nationally representative sample differed from
the sample in 4a on each of the four AoV clusters. This sample was obtained as part of a
Templeton funded-project examining the cognitions of religious people, and so consistent of a
nationally represented sample of theists. This allowed us to perform an interesting test across
samples, exploring whether this sample scored higher on AoV for the Divine than the less
religious sample in 4a.
Method
Participants
We advertised for 2000 American participants using the Qualtrics panels service. We
determined sample size by recruiting as many participants as possible given the cost constraints
of the panel service. Participants were all religious and recruited to be nationally representative
on the key dimensions of age, political party affiliation, race, and region of the country (South,
Northeast, Midwest, West). In total, 2010 participants completed the online survey. We only
analyzed data from participants who passed a single attention check. (178 excluded before
analyses). The final sample (N = 1,832; 1,379 women; 447 men; 6 other) was 73.2% White,
12.1% Black, 10.2% Other, and 4.6% Asian. Education level spanned from grammar school
through doctoral degree and income spanned from under $25,000 to over $150,000 per year.
Modal education level was some college (30.9%) and the median income was $25,000 -$50,000
per year.
Measures

Assumptions of Vulnerability
53
Assumptions of Vulnerability. Participants rated the vulnerability of each patient (twelve
total) using the same AoV items: For the Othered, M (SD) = 3.64 (0.97), α = .91; the
Environment, M (SD) = 3.77 (1.03), α = .91; the Powerful, M (SD) = 2.72 (0.84), α = .83; and,
the Divine, M (SD) = 2.46 (1.19), α = .91.
Political Orientation. Participants answered the question “How would you classify
yourself on the political spectrum?” on a 9-point scale from 1 = Very Liberal to 9 = Very
Conservative, M (SD) = 5.41 (2.21). Participants also completed other measures as part of
another study.
Results
We conducted a confirmatory factor analysis to evaluate whether the four-factor solution
showed acceptable model fit. All fit statistics surpassed typical thresholds (χ2(48) = 449.53, p <
.001, SRMR = 0.05, RMSEA = 0.07, TLI = .96, CFI = .97), indicating the four-factor model
adequately captures the structure of the data. Correlations among the factors were similar to
Study 4a (see the Supplement, p. 17). As in previous studies, conservatism related negatively to
the Othered, r = -.41, and the Environment, r = -.25, and was positively related to the powerful, r
= .15, and the divine, r = .08, all p’s < .001.
We next compared Study 4b participants to Study 4a participants on averages for each
AoV theme. Compared to the sample from Study 4a, participants rated the AoV of the
Environment lower (mDiff = 0.29), t(1,974) = -7.28, p < .001, rated the AoV of the Othered
lower (mDiff = 0.23), t(1,890) = -5.93, p < .00, rated the AoV of the Powerful higher (mDiff =
0.27), t(1,729) = 7.51, p < .001, and rated the AoV of the Divine higher (mDiff = 0.68), t(2,108)
= 15.46, p < .001. These results are consistent with a more religious sample.

Assumptions of Vulnerability
54
AoV scores plotted across the political spectrum as shown in Figure 9, and demonstrate
the same pattern obtained in previous studies: Liberals see extreme differences in levels of
vulnerability between the Othered/Environment vs. Powerful/Divine and conservatives do not.
Figure 9. Assumptions of Vulnerability Across the Political Spectrum (Study 4b).
Discussion
Study 4b replicated the results of the previous studies in a large, nationally representative
sample, providing strong support for our hypotheses. We again obtained a four-factor solution
with sound structure for the AoV scale. We also replicated the relationships between
conservatism and AoV themes. Importantly, in this bigger sample, we again observed the key
pattern that, when considering differences in levels of vulnerability for the Othered/Environment
vs. Powerful/Divine, extreme liberals distinguish to a high degree, centrists distinguish to a
moderate degree, and conservatives distinguish very little. Even though conservatism was

Assumptions of Vulnerability
55
positively correlated with AoVs for the Powerful and Divine, they still provided higher ratings
for the vulnerability level of the Othered and Environment.
Study 5: AoVs in the Nomological Net
Assumptions of vulnerability track political orientation and help explain moral
disagreement across politics. In this study, we explored the nomological net surrounding AoVs
to allow scientists to better situate this construct among other constructs. We also used this
opportunity to conduct a confirmatory factor analysis on then four themes of the Othered, the
Environment, the Powerful, and the Divine, and collected data over two time points to assess
test-retest reliability.
In addition to replicating results for political affiliation, MFQ associations, and
judgments of moral scenarios from the previous studies, we included at least one measure that
we expected should be related to each AoV theme due to conceptual overlap:
The Othered. We expected these AoVs would be associated with more strongly viewing
people who are stigmatized as legitimate moral patients (measured by the moral expansiveness
scale). Because universalism values are centrally about tolerance, appreciation of differences,
and supporting welfare for all people (S. Schwartz, 2012), we also expected to find a positive
relationship between universalism and AoVs for the othered.
The Environment. We predicted AoVs for the environment would be associated with
endorsing the view that animals, plants, and the environment are legitimate moral patients (moral
expansiveness scale).
The Powerful. Particularly due to the component that involves submission to authorities
(Manganelli Rattazzi et al., 2007), we expected AoVs for the powerful to be positively

Assumptions of Vulnerability
56
associated with right-wing authoritarianism. Likewise, we tested the hypotheses that viewing the
powerful as more vulnerable to harm would be associated with valuing power (status, prestige,
control, dominance) as a guiding life principle (measured by the Schwarz value scale), and with
endorsing deference to authorities as a moral imperative (operationalized by the morality as
cooperation scale). Because submitting to the powerful and authorities also requires a degree of
conformity (e.g., Steiner & Johnson, 1963), we also expected AoVs for the powerful would also
be positively related to basic values for conformity (avoiding conflict with others and following
social norms).
The Divine. Due to their overlap with religious views, AoVs for the divine should be
associated with more endorsement of supernatural beliefs.
Uncorrelated Measures. We had no reason to expect AoVs for any of the four themes
would be significantly associated with openness to experience or need for cognitive closure
outside of their mutual overlap with political ideology.
Method
Participants
We recruited participants through CloudResearch and had them complete two surveys
with a one-week gap in between. There were 1011 American participants who completed the
online survey. We only analyzed data from participants who passed more than 6 of 8 total
attention checks (91 excluded before analyses). The final sample (N = 920; 512 women, 405
men, 3 other; Mage = 41.51 years) was 81.7% White, 9.9% Black/African American, 5.0 %
Latinx/Hispanic, 2.6% Asian, 0.8% American Indian/Alaska Native, 0.1% Native
Hawaiian/Other Pacific Islander. Modal education level was a bachelor’s degree. Income ranged

Assumptions of Vulnerability
57
from under $25,000 to over $150,000 per year. and median income was $50,000 - $75,000 per
year. For the second wave, after excluding for attention checks, we had a total of 758
participants, an 82.4% retention rate (426 women, 330 men, 2 other; Mage = 39.72).
Procedure
Unless otherwise noted, responses were provided on a 1 (low endorsement) to 7 (strong
endorsement) scale. At both waves, participants completed the AoV scale and a measure of
conservatism. During the first wave only, participants completed all of the following measures:
Descriptive statistics and reliabilities for all wave 1 measures are shown in Supplemental Table
S14 (p. 19).
Measures
Assumptions of Vulnerability. Participants completed the same measures of the othered,
(Time 2, M (SD) = 3.84 (0.97), a = .93); the environment, (Time 2, M (SD) = 3.85 (1.11), a = .95;
the powerful, (Time 2, M (SD) = 2.39 (0.84), a = .88), and the divine, (Time 2, M (SD) = 1.63
(0.95), a = .92), as in Study 4.
Moral Foundations. Participants completed the same measure of moral foundations as in
Study 4, and we used the same procedure to create composites for care/fairness and
loyalty/authority/purity.
Factor-specific Scenarios. We use the same factor-specific scenarios as in study 3.
Again, the participants rated each scenario on a 5-point Likert scale ranging from 1 = “Not at all
immoral” to 5 = “Extremely Immoral”. For the othered, M (SD) = 2.70 (1.23), inter-item r = .52;
the environment, M (SD) = 3.45 (1.14), inter-item r = .44; the powerful, M (SD) = 2.85 (1.19),
inter-item r = .58; and, the divine, M (SD) = 2.87 (1.47), inter-item r = .83.

Assumptions of Vulnerability
58
Political Orientation. Participants completed the same conservatism measure from
Studies 1-4, M (SD) = 3.53 (1.79).
Other Variables. We also measured moral concern for the stigmatized, animals, and the
environment Moral Expansiveness (Crimston et al., 2016; responses provided from 1 = Outside;
4 = Inner circle). Participants also completed a measure of deference to authority as a moral
imperative (the deference subscale as the Morality as Cooperation Questionnaire (Curry et al.,
2019)) on a scale from 0 (Not at all relevant) to 100 (Extremely relevant). To measure basic
values for universalism, power, and conformity, we administered the Schwarz Value Survey
(Lindeman & Verkasalo, 2005), using a scale from 0 (Opposed to my principles) to 8 (Of
supreme importance). We measured right-wing authoritarianism using a 15-item scale
(Manganelli Rattazzi et al., 2007). Participants also completed the 10-item Supernatural Belief
Scale (Jong et al., 2013) on a scale from 1 = “Strongly disagree” to 9 = “Strongly agree.” We
measured openness to experience using the 10-item Personality Inventory (Gosling et al., 2003).
We also asked participants to respond to the 15-item Need for Cognitive Closure Scale (Roets &
Van Hiel, 2011) providing responses from 1 = “Completely disagree” to 6 = “Completely
agree.”
Additional exploratory measures—without preregistered hypotheses—included
utilitarianism, anxiety, faith in intuition, moral identity, belief in a dangerous world, social
dominance orientation, interpersonal reactivity, and self-efficacy.
Results
Confirmatory Factor Analysis
Our factor analyses show that the four-factor solution is still the most appropriate in both
waves. For the first wave, the fit measures for the confirmatory factor analysis are appropriate

Assumptions of Vulnerability
59
with RMSEA = 0.066, CFI = 0.978, TLI = 0.969. For the second wave, these values are similar:
RMSEA = 0.061, CFI = 0.984, TLI = 0.977. Explanatory analyses for both waves show that the
fit statistics only reach acceptable values after four factors are included, and that five or six-
factor solutions overfit the data. The mean within person reliability across time for the whole
scale was .88, indicating a good degree of stability in responses across time. Test-retest
reliability was similarly high for each of the four themes: The Environment, .85; The Othered,
.83; The Powerful, .91, and The Divine, .93.
Replications
The full correlation matrix can be found in the Supplement (Table S14, p. 19). These both
replicated findings from previous studies and provided initial support for our predictions (more
conservative tests of these are reported below). Figure 10 shows that again we obtained the same
pattern of AoV ratings across the political spectrum (Wave 1 data shown in the first panel and
Wave 2 data shown in the second panel), with liberals showing extreme separation in levels of
vulnerability, and conservatives showing small differences in these perceptions.

Assumptions of Vulnerability
60
Figure 10. Assumptions of Vulnerability Across the Political Spectrum (Study 5).
We used the same regression procedure as in previous studies to test the relationships
between assumptions of vulnerability and care/fairness and loyalty/authority/purity, controlling
for conservatism; as well as judgement of moral scenarios. Results replicated our previous
findings and are presented in the Supplement (pp. 20-21).
Psychological Concomitants
We pre-registered our intent to control for conservatism to assess the unique contribution
of assumptions of vulnerability to the outcome variables included in this Study. For all models,
we hierarchically regressed each outcome variable on mean-centered conservatism on a first
step, and the four AoV factors (mean centered) on a second step. Below we present results
specific to our pre-registered hypotheses. Full results are shown in the Supplement, pp 22-26.
The Othered. Results for viewing the Othered as vulnerable to harm conformed to our
expectations. AoVs for the othered significantly and positively predicted the stigma facet of the

Assumptions of Vulnerability
61
moral expansiveness scale, β = .24, p < .001; and universalism values, β = .14, p < .001. The
more people endorsed AoVs for the Othered, the more they showed concern for moral
transgressions against people who belong to stigmatized groups, and enduring values reflecting
care for the dissimilar, even after accounting for conservatism.
The Environment. Viewing the environment as vulnerable to harm was associated with
moral concern in the way we predicted. AoVs for the Environment were significantly and
positively associated with endorsements of plants, β = .34, animals, β = .34, and the
environment, β = .44, all’s p < .001, as subjects of moral concern. Even after adjusting for
political ideology, these results show that the more people viewed the environment as vulnerable
to harm, the more concern they show about potential harm done to environmentally-relevant
targets.
The Powerful. AoVs for the powerful were associated with authoritarianism, deference,
and basic values as expected. Specifically, viewing the Powerful as more vulnerable to harm
significantly predicted higher endorsement of right-wing authoritarianism, β = .14, p < .001,
viewing deference as a moral requirement, β = .24, p < .001, as well as basic values for power, β
= .14, p < .03 and conformity, β = .09, p < .03. These results demonstrate consistency between
viewing powerful figures as vulnerable to harm and endorsing similar ideologies and values.
The Divine. As we expected, those who viewed the Divine as more vulnerable to harm
also more strongly believed in the supernatural, β = .31, p < .001, likely due to the fact that key
tenants of many different religious belief systems involve supernatural elements.
Finally, in contrast to our pre-registered predictions, AoVs for the environment were a
significant and positive predictor of openness to experience. For need for cognitive closure, only
AoVs for the divine were a significant (and positive) predictor. Thus, people who view the

Assumptions of Vulnerability
62
environment as more vulnerable to harm also tend to be higher on trait openness, and those who
believe divine targets are more vulnerable to harm show more cognitive rigidity.
Discussion
Study 5 replicated previous findings for the structure of our AoV scale; the relationships
between assumptions of vulnerability, political ideology, and moral foundations; and their
relevance to judgements of everyday moral scenarios. Additionally, Study 5 results generally
provided support for our pre-registered predictions. Each of the four AoVs each showed the
kinds of associations we expected with other moral constructs, values, beliefs, and cognition,
even when controlling for conservatism and the remaining assumptions of vulnerability. First,
AoVs for the othered were associated with viewing people who belong to stigmatized groups as
deserving of moral protection, and universalism values for tolerance, appreciation, and welfare
for all. Second, those who viewed the environment as more vulnerable to harm rated animals,
plants, and the environment as more worthy of moral concern. Third, the more that people
believed the powerful are vulnerable to harm, the more strongly they endorsed right-wing
authoritarianism, as well as power and conformity values, and viewed deference to authorities as
a moral imperative. Fourth, participants who provided higher ratings for AoVs of the Divine also
showed higher endorsement of supernatural beliefs. These results add to our confidence in the
validity of this AoV measure and support the argument that assumptions of vulnerability are core
to understanding morality.
We also observed a handful of relationships that contrasted with out pre-registered
expectations. AoVs for the environment were significantly associated with openness to
experience, and AoVs for the divine were significantly associated with need for cognitive
closure. These findings cannot be explained by the overlap of these AoVs with conservatism, as

Assumptions of Vulnerability
63
we controlled for it in these models. It could be that these relationships are driven by more
nuanced aspects of political and religious ideology. Better understanding these relationships is a
valid direction for future research.
Overall, the evidence provided by Study 5 suggests these AoV items constitute a valid
measure that adds an important missing piece to the puzzle of morality and politics.
Study 6: Assessing AoVs with Implicit Measures
Study 6 addressed important concerns with the self-report methodology that we relied
upon in previous studies. With any self-report measure, it is possible that responses reflect some
self-monitoring or post-hoc altering of beliefs. We note that all other measures of morality
involve the same self-reported Likert-type responses as used by the AoV scale, but nevertheless,
the methodology used in Study 6 allows us to directly address these concerns. We used the
Affect Misattribution Procedure (AMP) to investigate whether self-reported explicit AoVs were
associated with implicit AoVs across all four categories captured by the AoV scale (The Other,
The Environment, The Powerful, and The Divine). The AMP uses the logic of projective tests,
where people implicitly ascribe meaning to ambiguous stimuli. Past research supports its validity
and shows that it strongly predicts explicit attitudes, such as racism (Payne et al., 2005; Miles et
al., 2019). We expected to find high consistency between explicit and implicit assumptions of
vulnerability.
Method
Participants
There were 299 American participants recruited via CloudResearch who completed the
online survey. We only analyzed data from participants who passed more than 3 of 5 total
attention checks (18 excluded before analysis). After completing the survey, three additional

Assumptions of Vulnerability
64
participants did not consent to analyzing their data. The final sample (N = 278; 138 men, 139
women, 1 other; Mage = 37.55) was 75.5% White, 9.0% Black or African American, 7.2% Asian,
6.8% Latinx/Hispanic, 1.4% American Indian or Alaska Native. Education level spanned from
no high-school degree through doctoral degree and income spanned from under $25,000 to over
$150,000 per year. Modal education level was a college degree (54.0%) and the median income
was between $50,000 and $75,000 per year.
Measures
Assumptions of Vulnerability. Participants completed the same measures of the othered,
M (SD) = 3.92, (0.93), α = .92; the environment, M (SD) = 4.20, (0.91), α = .92; the powerful, M
(SD) = 2.42, (0.91), α = .85, and the divine, M (SD) = 1.82, (1.00), α = .93, as in Study 4 and
Study 5.
Implicit Measures of Vulnerability. We adapted the Affect Misattribution Procedure
(AMP) developed by Payne and colleagues (2005) to measure perceptions of vulnerability.
Participants were flashed with two images in quick succession: a primed stimulus (one of the
twelve AoV words; e.g., The Rainforest) followed by an ambiguous target stimulus (a Chinese
symbol). Following the target stimulus, participants were shown a visual mask image to prevent
them from viewing the Chinese symbol while they made an evaluative judgment about it (see
Figure 9). Participants were explicitly told to ignore the English word and select one of two
options indicating whether the Chinese symbol referred to something vulnerable (coded as 1) or
not vulnerable (coded as 0). The logic of the AMP is that evaluative judgments about an
ambiguous stimulus (the Chinese character) may be unintentionally influenced by affective
reactions to the primed stimulus (the AoV target word). The extent to which the judged
vulnerability of the Chinese symbol is influenced positively or negatively by the AoV prime

Assumptions of Vulnerability
65
allows us to determine the extent to which the AoV prime is implicitly seen as vulnerable. The
AMP is validated by over a decade of research as a measure of implicit attitudes (see Payne &
Lundberg, 2014 for a review) that predict both explicit attitudes (e.g., racial attitudes; political
preferences; Payne et al., 2005) and behaviors (Cameron et al., 2012).
Participants went through a total of three blocks of trials of this procedure. In each block
they evaluated all twelve AoV targets (flashed one at a time paired with a Chinese symbol) on
one of the three dimensions of vulnerability: vulnerability to harm, mistreatment, and
victimization. The order of targets was randomized within each block. We generated AMP
scores by averaging across these twenty-seven ratings: the Othered, M (SD) = 0.59, (0.25); the
Environment, M (SD) = 0.60, (0.24); the Powerful, M (SD) = 0.45 (0.23); and the Divine, M (SD)
= 0.39 (0.24). Figure 11 displays a representative set of stimuli from the AMP procedure.
Political Orientation. Participants completed the same conservatism measure from
Studies 1-4, M (SD) = 3.48 (1.63).

Assumptions of Vulnerability
66
Figure 11. Schematic Diagram of the Affect Misattribution ProcedureUused in Study 6.
Results
We first used correlational analyses to test the association between explicit ratings of
vulnerability provided by AoV and the implicit ratings of vulnerability provided by the AMP.
Results, shown in Table 6, revealed that explicit and implicit measures were, as predicted, highly
correlated for all four dimensions. These results show that ratings of vulnerability for these
targets at the conscious corresponded to automatic responses at the more intuitive level. These
relationships indicate responses to the AoV scale we have developed are valid, rather than
representing a set of response biases or post-hoc justifications.

Assumptions of Vulnerability
67
Table 6. Correlations between scores on AOV and AMP for the Othered, Environment,
Powerful, and Divine in Study 6.
Implicit AoVs: AMP Scores
Othered
Environment
Powerful
Divine
Explicit AoVs
Othered
.32***
.12*
-.07
-.04
Environment
.16**
.31***
-.11
.03
Powerful
.05
.00
.45***
.12
Divine
-.06
-.03
.21***
.39***
Note. N = 278. ***p < .001, **p < .01, and *p < .05.
We next examined relationships between conservatism and both sets of measures. For
explicit AoV variables, results replicated previous studies, with conservatism being negatively
associated with AoVs of the Othered (r = -.43) and the Environment (r = -.36), and positively
associated with AoVs of the Powerful (r = .25) and the Divine (r = .39, all p’s < .001). The same
pattern emerged for implicit AoVs as assessed by the AMP, although the magnitudes of these
relationships were expectedly weaker, which is not uncommon when there exist methodological
differences across measures (with conservatism being assessed at the explicit level and AoVs
being assessed at the implicit level; Dovidio et al., 2001). Specifically, conservatism was
negatively associated with AMP scores for the Othered (r = -.15, p = .01) and the Environment (r
= -.12, p = .049) and positively associated with AMP for the Powerful (r = .18, p = .002).
Conservatism was not significantly associated with AMP for the Divine (r = .09, p = .15), but the
relationship was in the expected direction.
Discussion
Study 6 showed explicit and implicit measures of AoVs correlated positively with each
other and showed the same pattern of relationships with political ideology. Of course, due to

Assumptions of Vulnerability
68
differences in whether measurement methods overlapped, associations among variables at the
explicit level were of stronger magnitude than associations for variables across the implicit and
explicit levels (Dovidio et al., 2001). As in previous studies, more liberal people showed higher
implicit and explicit assumptions of vulnerability for the Othered and the Environment; and
those higher on conservatism viewed the Powerful and the Divine as more vulnerable. These
results indicate that the explicit AoV measure indeed captures genuine judgements of these
various targets as susceptible to harm, mistreatment, and victimization, rather than some other
process of post-hoc reasoning and justification.
Section 3
In the third and final set of studies, we extended self-reported AoVs to behaviors with
real-world implications. In Study 7, we tested whether the four AoV themes predicted decisions
regarding which types of charity to which they would like to donate money. We expected that
higher scores on each AoV theme would predict likelihood of donating to a charity with a
corresponding theme. For instance, those who see the Environment as more vulnerable to harm
should be more likely to donate to an environmental charity, rather than charities supporting the
Othered, the Powerful, or the Divine.
Although we suggest that AoVs are relatively stable perceptions of people, we suggest
that they can be experimentally influenced. In Study 8, we provide a test of the contextual
malleability of assumptions of vulnerability of one paradigmatically “Othered” and “Powerful”
targets: Homeless people and CEOs. We manipulated perceptions of vulnerability of a CEO and
a homeless person, and then measured moral assessments of the CEO refusing to donate money
to the homeless person. Our key prediction was that describing the CEO as more vulnerable
would reduce moral condemnation of their decision to not help the homeless.

Assumptions of Vulnerability
69
Study 7: AoVs Predict Charity Donations
Thus far, we have shown the AoV scale has sound structure, reliability, and validity.
However, all preceding studies have relied on self-report variables (and, in one case, an implicit
measure). In Study 7, we assessed the relevance of AoV to behavior in the moral domain.
Participants completed AoV towards the Othered, the Environment, the Powerful, and the
Divine. Adopting methodology from previous research (e.g., Goenka & Van Osselaer, 2019),
participants were forced to choose whether to donate to a charity relevant to one domain (e.g.,
the NAACP for the Othered) vs. another (e.g., Law Enforcement Charitable Foundation for the
Powerful). We purposefully chose real charities for this study and ensured that each charity
exemplified one of the four entities represented in the AoV scale. We expected those who
perceive relatively high vulnerability in each category to be more likely to donate to the charity
corresponding to that domain. For instance, those who perceive the Othered as highly vulnerable
should be more likely to donate to the NAACP.
We noted that binary measures of real behavior are typically noisier than continuous self-
report Likert scales but nevertheless suggest that they will—overall—map onto people’s AoVs.
Method
Participants
There were 200 American participants who completed the survey online via Amazon
Mechanical Turk. We only analyzed data from participants who passed more than 3 of 5 total
attention checks (14 excluded before analyses). The final sample (N = 186 participants, 93
female, 92 male, 1 other; Mage = 41.91 years) was 86.0% White, 5.9 % Asian, 4.8% Black or
African American, 2.7% Latinx/Hispanic, 0.5% American Indian or Alaska Native. Income

Assumptions of Vulnerability
70
ranged from under $25,000 to over $150,000 per year. Modal education level was a bachelor’s
degree (57.5%) and the median income was $50,000 - $75,000 per year. Sample size was
determined in advance of data collection, per preregistration.
Measures
Assumptions of Vulnerability. We administered the same measure for assumptions of
vulnerability as in Studies 4 through 6. The othered, M (SD) = 3.70, (1.09), α = .93; the
environment, M (SD) = 3.83, (1.04), α = .92; the powerful, M (SD) = 2.41, (0.93), α = .89, and
the divine, M (SD) = 1.70, (1.08), α = .93.
Charity Donations. The prompt preceding the charity donation questions read as follows:
“Imagine that you have been given an amount of money to donate to charity. We have a list of 8
charities from which we will be choosing 2 to donate to, but first we want to see which charities
people care about the most. For each pair of charities presented to you, select the charity that you
would most prefer your money going to.” Participants were then tasked with choosing between
pairs charities, each of which represented one of the four AoV factors (Combinations presented
to participants shown in Table 7). This measure was binary, with the first charity was coded as 0
and the second charity coded as 1. We paired the charities in such a way that participants were
eventually presented with all six possible combinations between these charities.

Assumptions of Vulnerability
71
Table 7. Forced-Choice Donation Pairings
Charity 1 (coded as 0)
Charity 2 (coded as 1)
The Environment
vs.
The Othered
National Association for the
Advancement of Colored People:
works to eliminate race-based
discrimination
Clean Air Task Force: supports
zero-emissions policies for
cleaner air
The Divine vs.
The Powerful
Concerns of Police Survivors:
supports the families of deceased
police officers
The Voice of the Martyrs:
supports persecuted Christians
The Powerful vs.
The Othered
International Rescue Committee:
supports refugees
Law Enforcement Charitable
Foundation: supports police
officers
The Divine vs.
The Environment
Clean Air Task Force: supports
zero-emissions policies for
cleaner air
The Voice of the Martyrs:
supports persecuted Christians
The Divine vs.
The Othered
National Association for the
Advancement of Colored People:
works to eliminate race-based
discrimination
Help The Persecuted: protects
Christians from radical groups
The Environment
vs.
The Powerful
Law Enforcement Charitable
Foundation: supports police
officers
Clean Air Task Force: supports
zero-emissions policies for
cleaner air
Political Orientation. Participants completed the same conservatism measure from
Studies 1-4, M (SD) = 3.64 (1.92).
Results
Correlation analyses replicated previous results for conservatism and AoVs (see the
Supplement, p. 27). To test our main hypotheses regarding charity donation, we ran a series of
logistic regressions. In 6 models, we regressed each binary charity donation variable onto
conservatism, and each of the four AoV variables. Full results (including standardized beta
weights) are shown in the Supplement, pp. 27 – 29.

Assumptions of Vulnerability
72
The Environment. In all cases, AoVs for the environment were associated with higher
likelihood of donating to the environmental charity: For donating to the environment (coded 1)
vs. the othered (coded 0), OR [95% CI] = 1.68 [1.16, 2.50]; the environment (0) vs. the divine
(1), OR [95% CI] = 0.63 [0.37, 1.05]; and, the environment (1) vs. the powerful (0), OR [95%
CI] = 1.28 [0.82, 2.03]. Note, the 95% confidence intervals are relatively large, likely due to the
small sample size in this study Although all results were in the direction consistent with our
predictions, not all of these confidence intervals do not meet significance thresholds, likely
because of relatively low power with a binary behavioral measure.
The Othered. Consistent with our predictions, endorsing higher AoVs for the Othered
was significantly associated with a higher likelihood of donating to a charity representing people
who are othered (0) vs. charities representing the environment (1), OR [95% CI] = 1.68 [1.16,
2.50], the powerful (1), OR [95% CI] = 1.68 [1.16, 2.50], and the divine (1), OR [95% CI] =
1.68 [1.16, 2.50]. These results add evidence to the correspondence between our self-report
measure of AoVs and relevant real-world behaviors.
The Powerful. Again supporting our pre-registered expectations, AoVs for the powerful
were associated with a higher likelihood of donating to a charity that benefits the powerful vs.
other charities: For donating to the powerful (1) vs. the othered (0), OR [95% CI] = 1.42 [0.93,
2.20]; the environment (1) vs. the powerful (0), OR [95% CI] = 0.50 [0.30, 0.81]; and, the divine
(1) vs. the powerful, OR [95% CI] = 0.80 [0.53, 1.17]. Again, despite being in the predicted
direction, this last confidence intervals overlapped with 1, likely due to limited power.
Nevertheless, this overall patten of results demonstrates the expected consistency between self-
reported perceptions of vulnerability for powerful figures, and the likelihood of engaging in
charity behaviors that helps those figures.

Assumptions of Vulnerability
73
The Divine. Finally, in every forced-choice pairing, AoVs for the Divine were
significantly associated with a higher likelihood of donating to a charity representing the divine
(coded 1) vs. the othered (0), OR [95% CI] = 1.80 [1.10, 3.00], the environment (0), OR [95%
CI] = 1.70 [1.08, 2.73], and the powerful (0), OR [95% CI] = 1.51 [1.08, 2.14]. These results
show a high degree of overlap between self-reported views of divine targets as vulnerable to
harm and actually engaging in behaviors that provides charity support to them.
For more details about these results, including standardized beta weights, see the
Supplement pp. 27-28 and Table S18, p. 29.
Discussion
Study 7 provided evidence that self-reported assumptions of vulnerability predict
important real-world behaviors. Those who more strongly endorsed vulnerability of the
Environment, the Othered, the Powerful, and the Divine were more likely to choose to donate to
charities representing those groups. This pattern emerged even when using relatively
conservative models (controlling for conservatism and all other AoVs to assess the unique effect
of each), and when those assumptions of vulnerability that are more closely related to each other
(the environment with the othered; and the powerful with the divine) were pitted against each
other. In all cases, effects were in the expected direction, but a handful of findings for the
environment and the powerful failed to reach. But the overall pattern supports for the relevance
of assumptions of vulnerability for charity behaviors.
Study 8: Experimentally Manipulating AoVs
In this study, we attempt to experimentally manipulate AoVs and see if this causally
impacts people’s moral judgments. Participants passed judgment about a CEO who refused to

Assumptions of Vulnerability
74
donate money to a homeless person, after emphasizing the vulnerability of the CEO or the
homeless person (or neither). In line with the idea of moral typecasting (K. Gray & Wegner,
2009), we expected that this emphasis would impact AoVs towards theses targets and make the
failure to donate either less immoral (because the CEO is more of a victim) or more immoral
(because the homeless person is more of a victim).
Method
Participants
There were 507 American participants who completed the survey online via Amazon
Mechanical Turk. We only analyzed data from participants who passed more at least one of two
attention checks (1 excluded before analyses). The final sample (N = 506 participants, 266
women, 237 men, 3 other; Mage = 43.0 years) was 78.5%, 9.1% Black or African American,
6.9% Asian, 4.5% Latinx/Hispanic, 0.8% American Indian or Alaska Native, and 0.2% Native
Hawaiian or Other Pacific Islander. Education level spanned from no high-school degree through
doctoral degree and income spanned from under $25,000 to over $150,000 per year. Modal
education level was a bachelor’s degree (52.2%) and the median income was $50,000 - $75,000
per year. Sample size was determined in advance of the data collection, per preregistration.
Design
After providing informed consent, participants read the following scenario:
“It's almost dark and James Smith, a homeless person, is standing at the entrance
to the garage as Nicole French, the CEO of an investment banking firm, is
walking to her car after working late. James asks for some money and Nicole,
despite having a lot of money, hurries by.”

Assumptions of Vulnerability
75
Then, participants were randomly assigned to one of three conditions. In the Control
condition (n = 183), participants were directed to the dependent variables after reading the
scenario. In the CEO Vulnerability condition (n = 166), participants were instructed to think
about ways in which the CEO could be vulnerable to harm and mistreatment, and to write a few
sentences about that. In the Homeless Vulnerability condition (n = 157), participants were
instructed to think about ways in which the homeless person could be vulnerable to harm and
mistreatment, and to write a few sentences about that. Finally, participants completed the same
dependent measures as in the Control condition.
Measures
AoV CEO. We asked participants to rate CEOs on the following three items (5-point
scale from 1 = Not at all vulnerable to 5 = Completely vulnerable): “I believe that the [CEO is]
especially vulnerable to being harmed,” “I think that the [CEO is] especially vulnerable to
mistreatment,” and “I feel that the [CEO is] especially vulnerable to victimization.” We averaged
these three items to create a composite reflecting the belief that CEOs are vulnerable, M (SD) =
2.45 (0.96), α = .82.
AoV Homeless. Participants responded to three items on a (5-point scale from 1 = Not at
all vulnerable to 5 = Completely vulnerable): “I believe that [homeless people are] especially
vulnerable to being harmed,” “I think that [homeless people are] especially vulnerable to
mistreatment,” and “I feel that [homeless people are] especially vulnerable to victimization.”
These were used to create a mean composite reflecting perceptions of people who are homeless
as vulnerable, M (SD) = 4.11 (0.84), α = .81.
Moral Judgements of CEO’s Behavior. On a scale from 1 (not at all) to 7 (very much),
participants responded to the following three items: “I believe that the CEO did the right thing.”

Assumptions of Vulnerability
76
[reverse scored], “I believe that the CEO should have given some money to the homeless
person.”, and “I believe that the CEO’s behavior was wrong,” M (SD) = 4.19, (1.51), α = .90.
Political Orientation. Participants completed the same conservatism measure from
Studies 1-4, M (SD) = 3.76 (1.83).
Results
Manipulation Check
The manipulation significantly affected perceived vulnerability of the CEO, F(2, 503) =
10.86, d = 0.41, p < .001. Tukey HSD adjusted pairwise-comparisons revealed that perceived
CEO AoV was higher for participants in the CEO vulnerability condition (M = 2.71, SD = 1.00)
than participants in the Control (M = 2.40, SD = 0.93; p = .007) or Homeless vulnerability (M =
2.23, SD = 0.89; p < .001) conditions. Likewise, homeless AoVs were significantly higher in
homeless vulnerability condition (M = 4.26, SD = 0.72) vs. the CEO vulnerability condition (M =
3.99, SD = 0.89; p = .009), F(2, 503) = 4.47, d = 0.29, p = .01. Perceived vulnerability of the
homeless did not differ between other conditions, ps > .16, likely because homeless people are
usually perceived as highly vulnerable).
Moral Judgments
The manipulation also significantly affected how wrong participants viewed the CEO’s
behavior, F(2, 503) = 6.47, d = .35, p = .002. Tukey HSD adjusted pairwise-comparisons
revealed the CEO’s behavior was less wrong (more acceptable) in the CEO vulnerability
condition (M = 3.85, SD = 1.49), relative to the Control (M = 4.34, SD = 1.45; p = .007) and
Homeless vulnerability (M = 4.38, SD = 1.54; p = .004) conditions. Perceived wrongness of

Assumptions of Vulnerability
77
CEO’s behavior was not significantly different between the homeless vulnerability and control
conditions (p = .96).
Lastly, results of mediational analyses (controlling for political orientation) showed a
significant indirect effect of condition (from Homeless Vulnerability to CEO Vulnerability
condition) on perceived wrongfulness of the CEO’s behavior via assumptions of vulnerability of
the CEO (b = 0.17, SE = 0.06, p < .05, 95% CI [0.08, 0.30]), and a significant indirect effect via
assumptions of vulnerability of the homeless (b = 0.08, SE = 0.04, 95% CI [0.01, 0.17]). Thus,
all hypotheses were supported. Experimentally enhancing a target’s perceived vulnerability
caused participants to view transgressions against them as more morally wrong.
Discussion
Study 8 demonstrated that assumptions of vulnerability are at the crux of understanding
moral judgements. Participants exposed to a moral dyad (a homeless person and a CEO) rated a
moral act (refusing to help the homeless person) as more permissible when the typical agent (the
CEO) was experimentally manipulated to be perceived as more vulnerable; and as more
condemnable when the typical patient (the homeless person) was more vulnerable. This pattern
of results shows that AoVs are a mechanism that can be leveraged to change moral perceptions
in the context of any given dyad. Moral judgements are driven by perceptions of harm, and these
harm perceptions boil down to who and how much targets are believed to be vulnerable.
General Discussion
The current research addressed an ongoing debate in moral psychology about whether
descriptive political differences in moral judgments require the existence of different foundations
rooted in different mental moral modules (e.g., Haidt & Joseph, 2004); or whether moral

Assumptions of Vulnerability
78
disagreement can stem from a common harm-based moral template that we all seem to share
(e.g., Ochoa, 2022; Schein & Gray, 2015). Consistent with the idea of a harm-based moral mind,
and the Theory of Dyadic Morality (Schein & Gray, 2015, 2018), we found that different
understandings of harm can give rise to moral disagreement. Even though people ground their
moral judgments in concerns about harm, different people make different assumptions about
which targets are more vulnerable to victimization or mistreatment. These different assumptions
of vulnerability (AoVs) parsimoniously explain moral differences without needing to posit
separate moral mechanisms for every domain of political disagreement.
The first section of studies explored the basic concept of assumptions of vulnerability
(AoVs). In the Pilot Study, we developed some face-valid AoV items, and found that canonical
moral patients, typically seen by people as more vulnerable, indeed had higher AoVs. Study 1
showed that AoVs predicted judgments of moral status across different targets, and mirrored
obvious differences between liberals and conservatives. For example, conservatives had higher
AoVs about fetuses, and this explained why conservatives imbued them with moral rights. Study
2 revealed the connection between AoVs and moral judgment for politically contentious issues
like illegal immigration, even controlling for political ideology. When people disagree about the
vulnerability of an entity, they disagree about its moral treatment. Study 3 found that
conservatives (vs. liberals) viewing certain inanimate objects as more vulnerable to harm, like
the Bible and the US Flag, was grounded in viewing these objects as more alive, suggesting that
AoVs are not merely metaphorical but instead revolve around actual vulnerability to harm.
In the second set of studies, we applied AoVs to four themes in ongoing political debates,
The Environment, The Othered, The Powerful, and The Divine. In Studies 4a (convenience) and
4b (high quality national sample), exploratory and confirmatory factor analyses supported the

Assumptions of Vulnerability
79
structure of these themes and found illuminating political differences: liberals view the
Environment and the Othered as more vulnerable to harm, and conservatives tend to view the
Powerful and the Divine as more vulnerable to harm.
Across these four clusters, a broader pattern emerged: committed liberals see the Othered
and Environment as very vulnerable to harm and the Powerful and the Divine as very
invulnerable to harm. Committed conservatives view these four of these groups as similarly
vulnerable to harm. Studies 4a and 4b also showed that AoVs predicted political differences in
moral foundations. AoVs also predicted immorality ratings for transgressions against
thematically related targets (e.g., AoVs for the Othered predicted immorality of detaining
immigrants), above and beyond moral foundation items and political ideology, showing their
unique ability to explain moral judgments.
Study 5 added to our confidence in our measure of AoVs by demonstrating high test-
retest reliability and by linking AoVs to a nomological net, including moral expansiveness, basic
values, authoritarianism, supernatural beliefs, and cognitive dispositions. People who more
strongly endorsed AoVs for the Environment, the Othered, the Powerful, and the Divine also
endorse sets of values, beliefs, and dispositions that are thematically related to these. Study 6
further added legitimacy to our conceptualization of AoVs by showing the correspondence of
AoVs measured implicitly and explicitly.
The third set of studies expanded on our understanding of AoVs. In Study 7, we
demonstrated that AoVs for the four themes predicted higher likelihood of contributing real
donations to a thematically relevant charity vs. unrelated charities (e.g., higher AoVs for the
Othered predicted donating to the NAACP vs. charities thematically linked to the other themes),
even when controlling for conservatism. Even though AoVs were often measured via self-report,

Assumptions of Vulnerability
80
they correspond to every-day, morally-relevant, and conceptually-consistent behaviors. Study 8
found that AoVs can be experimentally manipulated, and this allowed us to causally demonstrate
that AoVs predicted moral judgments.
Across these three sets of studies, AoVs seem to meaningfully explain both moral
judgments and political disagreement within the framework of a common harm-based moral
mind. These results are consistent with ideas proposed years ago by Turiel, who suggested that
different “informational assumptions” about harm help give rise to moral differences across
people and cultures (Turiel et al., 1987). Explaining differences in moral judgment with
informational assumptions—which are intrinsically amoral—helps avoid the tautology found
with moral foundations theory, which explains moral differences via moral differences (i.e.,
conservatives are more likely to moralize certain concerns because a questionnaire reveals that
they moralize those concerns).
Political Differences
When comparing those on the political left to the political right, moral differences are
bound to emerge because these ideological views are tied up in factors such as demographics and
identity (Hogg, 2007; Pew Research Center, 2018), attitudes about systems and power (Jost et
al., 2003), different motivational concerns (Greenberg et al., 1990) personality traits (e.g., Hirsh
et al., 2010), cognitive dispositions (Choma et al., 2014; Womick & King, 2021), interpersonal
orientation (Morris, 2020), worldviews (Koltko-Rivera, 2004), and even genetics (Lewis &
Bates, 2014). The current paper expands upon past research by revealing a concrete ideological
difference in AoVs and demonstrating that it predicts moral judgment.

Assumptions of Vulnerability
81
AoVs also help explain why liberals and conservatives respond differently to specific
scenarios and items in the MFQ. Rather than possessing different overall values or foundations,
partisans systematically view different entities are seen as vulnerable to harm. For instance, the
fact that conservatives see general authority figures as more vulnerable than liberals provides
insight into why they report valuing “authority” more than liberals (and, as the current work
shows, provides additional explanatory value in values for deference, conformity, and right-wing
authoritarianism).
Although we do not believe that the moral differences between ideological groups are
essential, immutable characteristics, liberals and conservatives consistently depart from each
other on AoVs in ways that connect with their political ideologies. These different perspectives
help make sense of real-world conflict about topics like race and policing. Liberals emphasize
how black men are generally more vulnerable, emphasizing the number of unarmed black men
killed from unjustified force and police brutality. Conservatives typically highlight the
underemphasized vulnerability of police officers, who risk their lives every day to help enforce
laws.
Likewise, debates about immigration seem to focus on different narratives of
vulnerability. Liberals typically emphasize the vulnerability of illegal immigrants, whose
children may be detained of fleeing violence and instability in their home country. Conservatives
seem to emphasize the vulnerability of local business owners or Americans who they believe
may be harmed by illegal immigrants who they feel threaten the American economy. Our goal
here is not to weigh in on the truth of these perceptions—many other social psychology scholars
have done illuminating work on race and power—but merely to illustrate the connection between
AoVs and ongoing battles about politics and morality. Still, there remains work to be done

Assumptions of Vulnerability
82
investigating the psychological underpinnings responsible for liberals and conservatives view
specific groups vs. others as more and less vulnerable. These are likely linked to broader
tendencies around topics like traditional power hierarchies, and should be investigated in future
research.
In the current research, we found that not only do liberals and conservatives differ in
which entities they perceive as relatively vulnerable to harm, they also differ in the extent to
which they distinguish between the vulnerable and invulnerable. In the current set of studies,
those on the far left divided the world into the extremely vulnerable (The Othered and The
Environment) and the extremely invulnerable (The Powerful and The Divine). Conservatives on
the other hand, minimized the differences in AoV between entities, seeing them as more similar
in levels vulnerability. The strong distinction shown by liberals is consistent with the emphasis
of progressives on the importance of systemic power differences, and the tendency to typecast
social groups as oppressors and oppressed (e.g., Freire, 2005). These data suggest, to those on
the far left, the world is a dichotomy between the powerful who are undeserving of protection,
and the marginalized who are in urgent need of protection. We suspect these narratives of
equality versus inequality in vulnerability arise from basic political assumptions, and look
forward to studying this in future work.
Limitations & Future Directions
The present research synthesized many studies using cross-sectional and multi-wave data,
self-report and implicit measures, convenience sampling and nationally representative data, and
combined correlational with experimental research. This methodological variety balances
strengths and weaknesses across studies and provides a rich set of consistent and replicable

Assumptions of Vulnerability
83
evidence for our hypotheses. Table 8 summarizes key limitations to these data. One limitation on
the inferences that can be drawn from these data is that all participants located in the United
States. We were primarily concerned with understanding political differences in the United
States, which in many ways is an outlier on politics globally. Still, for those interested in
applying these results to other contexts, the reliance on U.S. participants leaves unclear the cross-
cultural generalizability of these results. Researchers should probe generalizability in subsequent
studies.
Table 8. Table of Limitations
Category
Description
1. Generalizability
Use of non-representative data
2. Generalizability
Use of U.S. based samples
3. Conceptual
Focus on perceptions of moral patients and not moral acts and
moral agents
4. Conceptual
Focus on politics broadly, rather than the intersectional identities
and related processes underlying political dispositions
Based on our theoretical framework, we focused on how perceptions of the vulnerability
of moral patients can help us understand moral judgement and political differences. Yet, patients
are only one part of the equation of morality. The Theory of Dyadic Morality posits that harm is
perceived when 1) an intentional agent 2) acts on 3) a vulnerable patient. We expect that moral
disagreement can also be understood in terms of perceptions of moral agents and the extent to

Assumptions of Vulnerability
84
which different categories of agentic entities are perceived to be likely to and capable of causing
harm, and what kinds of acts are perceived to be valid deliveries of harm. For instance, people
may differ on what kinds of actors they view as intentional agents capable of causing harm (e.g.,
men vs. women; corporations vs. people; systems vs. individuals), as well as what kind of acts
are legitimate means of causing harm (e.g., verbal vs. physical; direct vs. indirect; prayer vs.
witchcraft, etc). Exploring these issues may further enlighten how moral differences in politics
and other domains can emerge from a common harm-based template.
Finally, the current research was designed to understand liberal-conservative differences
in morality broadly. Yet, political attitudes, beliefs, and behaviors are varied and connect with a
variety of identities (which in the case of politics, are often intersectional). Future research
should explore how assumptions of vulnerability help explain moral differences that emerge
from other politically relevant attitudes, such as social dominance orientation, system
justification, just world beliefs, right-wing authoritarianism, and left-wing authoritarianism.
Additionally, further research needs to examine the role of identity in assumptions of
vulnerability and their relationship to politics and moral judgment more closely (see Hester &
Gray, 2020).
In the United States, there are systematic demographic differences between those on the
political left and the political right. Those on the right tend to be more demographically
homogenous and are more likely to be older, white, wealthy, and male Americans (Pew Research
Center, 2018). In contrast, the left tends to be comprised of greater demographic heterogeneity,
including more Americans who belong to marginalized groups, and a wider variety of religious
diversity (Pew Research Centers, 2018). Understanding how these factors impinge on
assumptions of vulnerability and their relationship to politics and morality is an urgent goal for

Assumptions of Vulnerability
85
future science. Even in light of these considerations, assumptions of vulnerability represent a
novel piece of the puzzle of morality and politics and provide a key to understanding
commonalities and differences across these.
Conclusion
Political differences in moral judgments are obvious, but the explanation for these
differences are less obvious. Across eight studies, we find that people’s assumptions of
vulnerability (AoVs) predict their moral judgments and give shape to political disagreement. We
all share the same harm-based mind, grounding our moral judgments in concerns about
protecting vulnerable entities from suffering. But there are many different potentially vulnerable
entities, from the children who make up the future of society, to the powerful people who enforce
social order. How much people view these various entities as vulnerable to victimization—
including the Othered, the Environment, the Powerful, and the Divine—helps explain political
disagreement about hot-button issues.
AoVs not only parsimoniously explain moral differences, they also provide some
optimism. Although they disagree, liberals and conservatives both care about victimization and
mistreatment. Grounding differences in the common currency of assumptions of vulnerability
provides a shared reality to foster understanding. When seeking to understand someone with a
different moral position, ask yourself a simple question: what do they see as vulnerable to
victimization, mistreatment, or harm?

Assumptions of Vulnerability
86
References
Akin, J. (2002, March 1). When babies get their souls. Catholic Answers.
https://www.catholic.com/magazine/print-edition/when-babies-get-their-souls
Altemeyer, B., & Hunsberger, B. (1992). Authoritarianism, Religious Fundamentalism, Quest,
and Prejudice. International Journal for the Psychology of Religion, 2(2), 113–133.
https://doi.org/10.1207/s15327582ijpr0202_5
As Economic Concerns Recede, Environmental Protection Rises on the Public’s Policy Agenda.
(2020, February 13). Pew Research Center.
https://www.pewresearch.org/politics/2020/02/13/as-economic-concerns-recede-
environmental-protection-rises-on-the-publics-policy-agenda/
Atari, M., Haidt, J., Graham, J., Koleva, S., Stevens, S. T., & Dehghani, M. (2023). Morality
Beyond the WEIRD: How the Nomological Network of Morality Varies Across Cultures.
Journal of Personality and Social Psychology: Personality Processes and Individual
Differences, 125(5), 1157–1188. https://doi.org/10.1037/pspp0000470
Avineri, S. (1968). The Social and Political Thought of Karl Marx. Cambridge University Press.
https://books.google.com/books?id=UG2bcaExDGkC
Baldwin, T. (1979). Foresight and Responsibility. Philosophy, 54(209), 347–360.
https://doi.org/10.1017/S0031819100048750
Barrett, L. F. (2009). The future of psychology: Connecting mind to brain. Perspectives on
Psychological Science, 4(4), 326–339. https://doi.org/10.1111/j.1745-6924.2009.01134.x
Barrett, L. F. (2013). Psychological Construction: The Darwinian approach to the science of
emotion. Emotion Review, 5(4), 379–389. https://doi.org/10.1177/1754073913489753

Assumptions of Vulnerability
87
Barrett, L. F., & Russell, J. A. (Eds.). (2014). The Psychological Construction of Emotion (1
edition). The Guilford Press.
Brown, A. (2017, January 13). Republicans more likely than Democrats to have confidence in
police. https://www.pewresearch.org/short-reads/2017/01/13/republicans-more-likely-
than-democrats-to-have-confidence-in-police/
Buchman, E. (2022, September 5). The dangers of crossing the U.S./Mexico border. March On
Foundation. https://mo-foundation.org/blog-dangers-of-crossing-the-border/
Burke, M. A. (2017). Racing Left and Right: Color-Blind Racism’s Dominance across the U.S.
Political Spectrum. The Sociological Quarterly, 58(2), 277–294.
https://doi.org/10.1080/00380253.2017.1296335
Callero, P., L. (2003). The sociology of the self. Annual Review of Sociology, 29, 115–133.
https://doi.org/10.1146/annurev.soc.29.010202.100057
Cameron, C. D., Brown-Iannuzzi, J. L., & Payne, B. K. (2012). Sequential Priming Measures of
Implicit Social Cognition A Meta-Analysis of Associations With Behavior and Explicit
Attitudes. Personality and Social Psychology Review, 16(4), 330–350.
Cameron, C. D., Lindquist, K. A., & Gray, K. (2015). A Constructionist Review of Morality and
Emotions. Personality and Social Psychology Review, 19(4), 371–394.
https://doi.org/10.1177/1088868314566683
Chalmers, D. J. (1997). The conscious mind: In search of a fundamental theory (1st ed.). Oxford
University Press, USA.
Choma, B., Hodson, G., Hoffarth, M. R., Charlesford, J. J., & Hafer, C. L. (2014). Reasoning
Ability and Ideology: Inaccuracies in Hierarchical Category Relations (But Not

Assumptions of Vulnerability
88
Numerical Ability) Are Associated With Right-Wing Authoritarianism. Journal of
Individual Differences, 35(3), 177–183. https://doi.org/10.1027/1614-0001/a000142
Clifford, S., Iyengar, V., Cabeza, R., & Sinnott-Armstrong, W. (2015). Moral foundations
vignettes: A standardized stimulus database of scenarios based on moral foundations
theory. Behavior Research Methods, 47(4), 1–21. https://doi.org/10.3758/s13428-014-
0551-2
Cooper, F. (2020, April 1). Cop fragility and blue lives matter.
https://www.semanticscholar.org/paper/Cop-Fragility-and-Blue-Lives-Matter-
Cooper/cf3d2466890849f5b967387645fa8587c0080e15
Crawford, J. T., Brandt, M. J., Inbar, Y., Chambers, J. R., & Motyl, M. (2017). Social and
Economic Ideologies Differentially Predict Prejudice Across the Political Spectrum, but
Social Issues Are Most Divisive. Journal of Personality and Social Psychology, 112(3),
383–412. https://doi.org/10.1037/pspa0000074
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. http://psycnet.apa.org/psycinfo/2016-01230-001/
Curry, O. S., Jones Chesters, M., & Van Lissa, C. J. (2019). Mapping morality with a compass:
Testing the theory of ‘morality-as-cooperation’ with a new questionnaire. Journal of
Research in Personality, 78, 106–124. https://doi.org/10.1016/j.jrp.2018.10.008
Cushman, F. (2008). Crime and punishment: Distinguishing the roles of causal and intentional
analyses in moral judgment. Cognition, 108(2), 353–380.
https://doi.org/10.1016/j.cognition.2008.03.006

Assumptions of Vulnerability
89
Cushman, F., & Young, L. (2011). Patterns of moral judgment derive from nonmoral
psychological representations. Cognitive Science, 35(6), 1052–1075.
https://doi.org/10.1111/j.1551-6709.2010.01167.x
Devos, T., & Banaji, M. R. (2005). American = White? Journal of Personality and Social
Psychology, 88(3), 447–466. https://doi.org/10.1037/0022-3514.88.3.447
Dovidio, J. F., Kawakami, K., & Beach, K. R. (2001). Implicit and Explicit Attitudes:
Examination of the Relationship between Measures of Intergroup Bias. In Blackwell
handbook of social psychology: Intergroup Processes (pp. 175–197).
Dretske, F. I. (1973). Perception and other minds. Noûs, 7(1), 34–44.
https://doi.org/10.2307/2216182
Epley, N., & Waytz, A. (2009). Mind perception. In S. T. Fiske, D. T. Gilbert, & G. Lindzey
(Eds.), The Handbook of Social Psychology (5th ed., pp. 498–541). Wiley.
Feinberg, M., & Willer, R. (2019). Moral reframing: A technique for effective and persuasive
communication across political divides. Social and Personality Psychology Compass,
13(12). https://doi.org/10.1111/spc3.12501
Fine, T. S. (1992). Individualism & Liberalism/Conservatism: Broadening Dimensions of Policy
Support. Polity, 25(2), 315–327. https://doi.org/10.2307/3235115
Freire, P. (2005). Pedagogy of the oppressed (30th anniversary ed). Continuum.
Frimer, J. A., Gaucher, D., & Schaefer, N. K. (2014). Political conservatives’ affinity for
obedience to authority is loyal, not blind. Personality and Social Psychology Bulletin,
40(9), 1205–1214. https://doi.org/10.1177/0146167214538672

Assumptions of Vulnerability
90
Frimer, J. A., Tell, C. E., & Haidt, J. (2015). Liberals Condemn Sacrilege Too The Harmless
Desecration of Cerro Torre. Social Psychological and Personality Science,
1948550615597974.
Frimer, J. A., Tell, C. E., & Motyl, M. (2017). Sacralizing liberals and fair-minded
conservatives: Ideological symmetry in the moral motives in the culture war. Analyses of
Social Issues and Public Policy, 17(1), 33–59. https://doi.org/10.1111/asap.12127
Gawronski, B., Strack, F., & Bodenhausen, G. V. (2012). Attitudes and cognitive consistency:
The role of associative and propositional processes. In In R. E. Petty, R. H. Fazio, & P.
Brinol (Eds), Attitudes: Insights from the new implicit measures (pp. 83–116). Taylor and
Francis.
Gebhard, U., Nevers, P., & Billmann-Mahecha, E. (2003). Moralizing trees: Anthropomorphism
and identity in children’s relationships to nature. In Identity and the Natural
Environment: The Psychological Significance of Nature (pp. 91–111). MIT Press.
https://books.google.com/books?id=Wwf0oVe2rHIC
Goenka, S., & Van Osselaer, S. M. J. (2019). Charities Can Increase the Effectiveness of
Donation Appeals by Using a Morally Congruent Positive Emotion. Journal of Consumer
Research, 46(4), 774–790. https://doi.org/10.1093/jcr/ucz012
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
Graham, J. (2015). Explaining away differences in moral judgment: Comment on Gray and
Keeney (2015). Social Psychological and Personality Science, 6(8), 869–873.
https://doi.org/10.1177/1948550615592242

Assumptions of Vulnerability
91
Graham, J., Haidt, J., Motyl, M., Meindl, P., Iskiwitch, C., & Mooijman, M. (2018). Moral
Foundations Theory: On the advantages of moral pluralism over moral monism. In Atlas
of Moral Psychology (pp. 211–222). Guilford Press.
Graham, J., Haidt, J., & Nosek, B. A. (2008). The Moral Foundations Questionnaire (short
version). https://moralfoundations.org/questionnaires/
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
Graham, J., Nosek, B. A., Haidt, J., Iyer, R., Koleva, S., & Ditto, P. H. (2011). Mapping the
moral domain. Journal of Personality and Social Psychology, 101(2), 366–385.
https://doi.org/10.1037/a0021847
Gramlich, J. (2021, September 3). Majorities of Americans say unions have a positive effect on
U.S. and that decline in union membership is bad. Pew Research Center.
https://www.pewresearch.org/short-reads/2021/09/03/majorities-of-americans-say-
unions-have-a-positive-effect-on-u-s-and-that-decline-in-union-membership-is-bad/
Gramlich, J. (2023, June 16). Americans and affirmative action: How the public sees the
consideration of race in college admissions, hiring. Pew Research Center.
https://www.pewresearch.org/short-reads/2023/06/16/americans-and-affirmative-action-
how-the-public-sees-the-consideration-of-race-in-college-admissions-hiring/
Gray, H. M., Gray, K., & Wegner, D. M. (2007). Dimensions of mind perception. Science,
315(5812), 619.
Gray, K., DiMaggio, N., Schein, C., & Kachanoff, F. (2022). The problem of purity in moral
psychology. Personality and Social Psychology Review, 1–37.

Assumptions of Vulnerability
92
Gray, K., Jenkins, A. C., Heberlein, A. S., & Wegner, D. M. (2011). Distortions of mind
perception in psychopathology. Proceedings of the National Academy of Sciences,
108(2), 477–479. https://doi.org/10.1073/pnas.1015493108
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., MacCormack, J. K., Henry, T., Banks, E., Schein, C., Armstrong-Carter, E., Abrams,
S., & Muscatell, K. A. (2022). The Affective Harm Account (AHA) of Moral Judgment:
Reconciling Cognition and Affect, Dyadic Morality and Disgust, Harm and Purity.
Journal of Personality and Social Psychology, 123(6), 1199–1222.
https://doi.org/10.1037/pspa0000310
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, 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
Greenberg, J., Pyszczynski, T., Solomon, S., Rosenblatt, A., Veeder, M., Kirkland, S., & Lyon,
D. (1990). Evidence for terror management theory II: The effects of mortality salience on
reactions to those who threaten or bolster the cultural worldview. Journal of Personality
and Social Psychology, 58(2), 308–318. https://doi.org/10.1037/0022-3514.58.2.308

Assumptions of Vulnerability
93
Greene, J. D., Sommerville, R. B., Nystrom, L. E., Darley, J. M., & Cohen, J. D. (2001). An
fMRI investigation of emotional engagement in moral judgment. Science, 293(5537),
2105–2108. https://doi.org/10.1126/science.1062872
Gutierrez, R., & Giner-Sorolla, R. (2007). Anger, disgust, and presumption of harm as reactions
to taboo-breaking behaviors. Emotion, 7(4), 853–868. https://doi.org/10.1037/1528-
3542.7.4.853
Haidt, J. (2001). The emotional dog and its rational tail: A social intuitionist approach to moral
judgment. Psychological Review, 108(4), 814–834.
Haidt, J. (2012). The Righteous Mind: Why Good People Are Divided by Politics and Religion.
Knopf Doubleday Publishing Group.
Haidt, J., & Graham, J. (2007). When morality opposes justice: Conservatives have moral
intuitions that liberals may not recognize. Social Justice Research, 20(1), 98–116.
https://doi.org/10.1007/s11211-007-0034-z
Haidt, J., & Joseph, C. (2004). Intuitive ethics: How innately prepared intuitions generate
culturally variable virtues. Daedalus, 133(4), 55–66.
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
Hart, H. L. A., & Honoré, T. (1985). Causation in the law (2nd ed.). Oxford University Press,
USA.
Haslam, N., Loughnan, S., Kashima, Y., & Bain, P. (2008). Attributing and denying humanness
to others. European Review of Social Psychology, 19(1), 55–85.
https://doi.org/10.1080/10463280801981645

Assumptions of Vulnerability
94
Hester, N., & Gray, K. (2020). The moral psychology of raceless genderless strangers.
Perspectives on Psychological Science, 15(2), 216–230.
https://doi.org/10.1177/1745691619885840
Hirsh, J. B., DeYoung, C. G., Xu, X., & Peterson, J. B. (2010). Compassionate Liberals and
Polite Conservatives: Associations of Agreeableness With Political Ideology and Moral
Values. Personality and Social Psychology Bulletin, 36(5), 655–664.
https://doi.org/10.1177/0146167210366854
Ho, A. K., Sidanius, J., Kteily, N., Sheehy-Skeffington, J., Pratto, F., Henkel, K. E., Foels, R., &
Stewart, A. L. (2015). The Nature of Social Dominance Orientation: Theorizing and
Measuring Preferences for Intergroup Inequality Using the New SDO7 Scale. Journal of
Personality and Social Psychology, 109(6), 1003–1028.
https://doi.org/10.1037/pspi0000033
Hogg, M. A. (2007). Uncertainty–Identity Theory (Vol. 39, pp. 69–126). Academic Press.
https://doi.org/10.1016/S0065-2601(06)39002-8
Horowitz, J. M., Hurst, K., & Braga, D. (2023, June 14). Support for the Black Lives Matter
movement has dropped considerably from its peak in 2020. Pew Research Center.
https://www.pewresearch.org/social-trends/2023/06/14/views-on-the-black-lives-matter-
movement/
Jones, K. L., Noorbaloochi, S., Jost, J. T., Bonneau, R., Nagler, J., & Tucker, J. A. (2018).
Liberal and Conservative Values: What We Can Learn From Congressional Tweets.
Political Psychology, 39(2), 423–443. https://doi.org/10.1111/pops.12415

Assumptions of Vulnerability
95
Jong, J., Bluemke, M., & Halberstadt, J. (2013). Fear of death and supernatural beliefs:
Developing a new supernatural belief scale to test the relationship. European Journal of
Personality, 27(5), 495–506. https://doi.org/10.1002/per.1898
Jost, J. T., Glaser, J., Kruglanski, A. W., & Sulloway, F. J. (2003). Political conservatism as
motivated social cognition. Psychological Bulletin;Psychological Bulletin, 129(3), 339–
375. https://doi.org/10.1037/0033-2909.129.3.339
Kanai, R., Feilden, T., Firth, C., & Rees, G. (2011). Political Orientations Are Correlated with
Brain Structure in Young Adults. Current Biology, 21(8), 677–680.
https://doi.org/10.1016/j.cub.2011.03.017
Kendi, I. X. (2019). How to Be an Antiracist. Random House Publishing Group.
https://books.google.com/books?id=6pNbDwAAQBAJ
Koltko-Rivera, M. E. (2004). The Psychology of Worldviews. Review of General Psychology,
8(1), 3–58. https://doi.org/10.1037/1089-2680.8.1.3
Krause, S. R. (2015). Freedom Beyond Sovereignty: Reconstructing Liberal Individualism.
University of Chicago Press. https://books.google.com/books?id=U_GQoAEACAAJ
Lewis, G. J., & Bates, T. C. (2014). Common Heritable Effects Underpin Concerns Over Norm
Maintenance and In‐Group Favoritism: Evidence From Genetic Analyses of Right‐Wing
Authoritarianism and Traditionalism. Journal of Personality, 82(4), 297–309.
https://doi.org/10.1111/jopy.12055
Lind, D. (2019, June 25). The horrifying conditions facing kids in border detention, explained.
Vox. https://www.vox.com/policy-and-politics/2019/6/25/18715725/children-border-
detention-kids-cages-immigration

Assumptions of Vulnerability
96
Lindeman, M., & Verkasalo, M. (2005). Measuring values with the Short Schwartz’s Value
Survey. Journal of Personality Assessment, 85(2), 170–178.
https://doi.org/10.1207/s15327752jpa8502_09
Lloyd, M. (2013). Power, politics, domination, and oppression. In The Oxford Handbook of
Gender and Politics (pp. 111–134). Oxford University Press.
https://books.google.com/books?id=GrNoAgAAQBAJ
Manganelli Rattazzi, A. M., Bobbio, A., & Canova, L. (2007). A short version of the Right-Wing
Authoritarianism (RWA) Scale. Personality and Individual Differences, 43(5), 1223–
1234. https://doi.org/10.1016/j.paid.2007.03.013
McHugh, C., McGann, M., Igou, E. R., & Kinsella, E. L. (2022). Moral Judgment as
Categorization (MJAC). Perspectives on Psychological Science, 17(1), 131–152.
https://doi.org/10.1177/1745691621990636
McNulty, J. (2018, August 1). Deportation and family separation impact entire communities,
researchers say. UC Santa Cruz News. https://news.ucsc.edu/2018/08/langhout-
immigration.html
Miles, A., Charron-Chenier, C., & Schleifer, C. (2019). Measuring automatic cognition:
Advancing dual-process research in sociology. American Sociological Review, 84, 308-
333.
Miller, A. (2022, July 7). Crimes committed by illegal immigrants surged in 2021 after declining
in previous years [Text.Article]. Fox News; Fox News.
https://www.foxnews.com/us/crimes-committed-illegal-immigrants-surged-2021-
declining-previous-years

Assumptions of Vulnerability
97
Morris, S. G. (2020). Empathy and the liberal-conservative political divide in the U.S. Journal of
Social and Political Psychology, 8(1), 08–24. https://doi.org/10.5964/jspp.v8i1.1102
Murphy, G. L. (2004). The big book of concepts. MIT Press.
Nichols, S. (2002). Norms with feeling: Towards a psychological account of moral judgment.
Cognition, 84(2), 221–236.
Ochoa, N. R. (2022). Template matching and moral judgment: A new method and empirical test.
Poetics, 101643. https://doi.org/10.1016/j.poetic.2021.101643
Okolie, A. C. (2009). Introduction to the Special Issue -- Identity: Now You Don’t See It; Now
You Do. Identity, 3(1), 1–7. https://doi.org/10.1207/S1532706XID0301_01
Paap, D., Schepers, M., & Dijkstra, P. U. (2020). Reducing ceiling effects in the Working
Alliance Inventory-Rehabilitation Dutch Version. Disability and Rehabilitation, 42(20),
2944. https://doi.org/10.1080/09638288.2018.1563833
Payne, B. K., Cheng, C. M., Govorun, O., & Stewart, B. D. (2005). An inkblot for attitudes:
Affect misattribution as implicit measurement. Journal of Personality and Social
Psychology, 89(3), 277–293. https://doi.org/10.1037/0022-3514.89.3.277
Payne, K., & Lundberg, K. (2014). The Affect Misattribution Procedure: Ten Years of Evidence
on Reliability, Validity, and Mechanisms. Social and Personality Psychology Compass,
8(12), 672–686. https://doi.org/10.1111/spc3.12148
Pinsof, D., Sears, D. O., & Haselton, M. G. (2023). Strange Bedfellows: The Alliance Theory of
Political Belief Systems. Psychological Inquiry, 34(3), 139–160.
https://doi.org/10.1080/1047840X.2023.2274433

Assumptions of Vulnerability
98
Pluckrose, H., & Lindsay, J. (2020). Cynical theories: How activist scholarship made everything
about race, gender, and identity―and why this harms everybody (None edition).
Pitchstone Publishing.
Prestigiacomo, A. (2016, April 22). 5 times ‘transgender’ men abused women and children in
bathrooms. The Daily Wire. https://www.dailywire.com/news/5-times-transgender-men-
abused-women-and-children-amanda-prestigiacomo
Putnam, H. (1975). Brains and behavior. In H. Putnam (Ed.), Philosophical Papers, Volume 2:
Mind, Language and Reality (pp. 325–341). Cambridge University Press.
Roets, A., & Van Hiel, A. (2011). Item selection and validation of a brief, 15-item version of the
Need for Closure Scale. Personality and Individual Differences, 50(1), 90–94.
https://doi.org/10.1016/j.paid.2010.09.004
Rosch, E. (1978). Principles of categorization. In E. Rosch & B. B. Lloyd (Eds.), Cognition and
Categorization. Erlbaum.
Rosefield, H. (2014, June 20). A brief history of oaths and books. The New Yorker.
https://www.newyorker.com/books/page-turner/a-brief-history-of-oaths-and-books
Rottman, J., Crimston, C. R., & Syropoulos, S. (2021). Tree‐Huggers Versus Human‐Lovers:
Anthropomorphism and Dehumanization Predict Valuing Nature Over Outgroups.
Cognitive Science, 45(4), e12967. https://doi.org/10.1111/cogs.12967
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

Assumptions of Vulnerability
99
Scarry, E. (1987). The body in pain: The making and unmaking of the world. Oxford University
Press.
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.
https://doi.org/10.1177/1088868317698288
Schwartz, S. (2012). An Overview of the Schwartz Theory of Basic Values. Online Readings in
Psychology and Culture, 2(1). https://doi.org/10.9707/2307-0919.1116
Schwartz, S. H., Caprara, G. V., & Vecchione, M. (2010). Basic Personal Values, Core Political
Values, and Voting: A Longitudinal Analysis: Basic Personal Values, Political Values
and Voting. Political Psychology, 31(3), 421–452. https://doi.org/10.1111/j.1467-
9221.2010.00764.x
Shweder, R., & Haidt, J. (1993). The Future of Moral Psychology: Truth, Intuition, and the
Pluralist Way. Psychological Science. https://doi.org/10.1111/j.1467-
9280.1993.tb00582.x
Smith, B. B. (2019, September 5). The disrespect is getting dangerous. National Police
Association. https://nationalpolice.org/the-disrespect-is-getting-dangerous/
Steiner, I. D., & Johnson, H. H. (1963). Authoritarianism and Conformity. Sociometry, 26(1), 21.
https://doi.org/10.2307/2785722
Tabahriti, S. (2022). The startup CEO who went viral for crying about laying off two employees
says “a lot of good” has come from his LinkedIn post. Business Insider.

Assumptions of Vulnerability
100
https://www.businessinsider.com/ceo-hypersocial-linkedin-crying-selfie-says-good-
came-from-it-2022-8
Taber, K. S. (2018). The Use of Cronbach’s Alpha When Developing and Reporting Research
Instruments in Science Education. Research in Science Education, 48(6), 1273–1296.
https://doi.org/10.1007/s11165-016-9602-2
Takahashi, K., & Jefferson, H. (2021). When the powerful feel voiceless: White identity and
feelings of racial voicelessness. PsyArXiv. https://doi.org/10.31234/osf.io/ry97q
The 6 worst crimes committed by illegal immigrants in recent history. (2018). Conservative
Zone. https://www.conservativenewszone.com/articles/the-6-worst-crimes-committed-by-
illegal-immigrants-in-recent-history/
Trends in party affiliation among demographic groups. (2018, March 20). Pew Research Center.
https://www.pewresearch.org/politics/2018/03/20/1-trends-in-party-affiliation-among-
demographic-groups/
Trump says Mexico sending “rapists” across border; he’d make country pay for border wall.
(2017, January 1). Fox News. https://www.foxnews.com/politics/trump-says-mexico-
sending-rapists-across-border-hed-make-country-pay-for-border-wall
Turiel, E., Killen, M., & Helwig, C. C. (1987). Morality: Its structure, functions, and vagaries. In
J. Kagan & S. Lamb (Eds.), The emergence of morality in young children (pp. 155–243).
University of Chicago Press.
Wall, P. D. (1979). On the relation of injury to pain the John J. Bonica Lecture. PAIN, 6(3), 253–
264. https://doi.org/10.1016/0304-3959(79)90047-2

Assumptions of Vulnerability
101
Waytz, A., Cacioppo, J., & Epley, N. (2010). Who sees human? The stability and importance of
individual differences in anthropomorphism. Perspectives on Psychological Science,
5(3), 219–232. https://doi.org/10.1177/1745691610369336
Waytz, A., Iyer, R., Young, L., Haidt, J., & Graham, J. (2019). Ideological differences in the
expanse of the moral circle. Nature Communications, 10(1), 4389.
https://doi.org/10.1038/s41467-019-12227-0
Womick, J., & King, L. A. (2021). Testing cognitive and interpersonal asymmetry vs. Symmetry
among voters in the 2020 presidential primaries. Journal of Social and Political
Psychology, 9(2), Article 2. https://doi.org/10.5964/jspp.7771
Womick, J., Woody, B., & King, L. A. (2021). Religious fundamentalism, right‐wing
authoritarianism, and meaning in life. Journal of Personality, 90(2), 277–293.
https://doi.org/10.1111/jopy.12665
Zmigrod, L., Eisenberg, I. W., Bissett, P. G., Robbins, T. W., & Poldrack, R. A. (2021). The
cognitive and perceptual correlates of ideological attitudes: A data-driven approach.
Philosophical Transactions of the Royal Society B: Biological Sciences, 376(1822),
20200424. https://doi.org/10.1098/rstb.2020.0424