ORG 6499 Week 4 Assignment

Cognition, Language, and Development

Embodying the Police:
The Effects of
Enclothed Cognition
on Shooting Decisions

Saaid A. Mendoza
Providence College, Providence, RI, USA

Elizabeth J. Parks-Stamm
University of Southern Maine, Portland, ME, USA and

Grand Canyon University, Phoenix, AZ, USA

Abstract

The theory of enclothed cognition proposes that wearing physical articles of clothing

can trigger psychological processes and behavioral tendencies connected to their

symbolic meaning. Furthermore, past research has found that increases in power are

associated with greater approach orientation and action tendencies. In this study, we

integrate these two literatures to examine how embodying the role of a police

officer through wearing a uniform would affect responses on a reaction-time mea-

sure known as the Shooter Task. This first-person video game simulation requires

participants to shoot or not shoot targets holding guns or objects. The task typically

elicits a stereotypical pattern of responses, such that unarmed Black versus White

targets are more likely to be mistakenly shot and armed Black versus White targets

are more likely to be correctly shot. Based on the relationship between power and

action, we hypothesized that participants who were randomly assigned to wear a police

uniform would show more shooting errors, particularly false alarms, than control

participants. Consistent with our hypotheses, participants in uniform were more

likely to shoot unarmed targets, regardless of their race. Moreover, this pattern was

partially moderated by attitudes about the police and their abuse of power. Specifically,

uniformed participants who justified police use of power were more likely to shoot

Corresponding Author:

Saaid A. Mendoza, Department of , Providence College, Providence, RI 02918, USA.

Email: [email protected]

Psychological Reports

2020, Vol. 123(6) 2353–2371

! The Author(s) 2019
Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/0033294119860261

journals.sagepub.com/home/prx

https://orcid.org/0000-0002-3052-8653

mailto:[email protected]

http://us.sagepub.com/en-us/journals-permissions

http://dx.doi.org/10.1177/0033294119860261

journals.sagepub.com/home/prx

http://crossmark.crossref.org/dialog/?doi=10.1177%2F0033294119860261&domain=pdf&date_stamp=2019-07-04

innocent targets than those who were wary of it. We discuss implications for police

perceptions and the theory of enclothed cognition more broadly.

Keywords

Stereotyping, enclothed cognition, police shootings, power, social justice

Introduction

The seemingly frequent occurrence of fatal civilian shootings by the police in
recent years has shone a negative spotlight on law enforcement across the coun-
try (Cook, 2015; Nix, Campbell, Byers, & Alpert, 2017). The fact that these
high-profile cases have disproportionately involved unarmed Black men has
brought attention to the racial disparities that have long persisted in the criminal
justice system (Kahn & Martin, 2016). Furthermore, social unrest has been
sparked by the perceived leniency shown to police officers, who are rarely
charged or convicted for their use of lethal force (Stinson, 2017). Researchers
have therefore sought to better understand the factors that influence police
decisions to shoot, with much of the focus centering around the target’s race
or ethnicity (Correll, Park, Judd, & Wittenbrink, 2002; Sadler, Correll, Park, &
Judd, 2012). In this study, we extend this line of psychological inquiry by exam-
ining not only the impact of race but also the effect of power conferred on the
shooter by the police uniform.

Racial bias in weapon perception

Although explicit forms of prejudice undoubtedly contribute to discriminatory
practices and biased policing (Hall, Hall, & Perry, 2016), a considerable amount
of social cognition research has established that race can also implicitly bias
people’s perceptions and behaviors (for a review, see Amodio & Mendoza,
2010). Specifically, negative stereotypes that associate Black individuals with
greater hostility and violence can influence the processing of weapon stimuli.
For instance, following the supraliminal presentation of a Black versus a White
facial prime, individuals are generally faster to identify a gun and are more likely
to misidentify a tool as a gun (Amodio et al., 2004; Klauer & Voss, 2008; Payne,
2001; Payne, Lambert, & Jacoby, 2002). In addition, the subliminal presentation
of a Black, compared to a White, facial prime can facilitate the accurate detec-
tion of degraded weapon images (Eberhardt, Goff, Purdie, & Davies, 2004).

Converging evidence of bias has emerged from psychological studies that
more directly examine how race can affect decisions to shoot criminal suspects
(e.g., Correll et al., 2002; Greenwald, Oaks, & Hoffman, 2003; Plant, Peruche, &
Butz, 2005). In one paradigm known as the first-person Shooter Task,

2354 Psychological Reports 123(6)

participants must quickly decide whether to shoot or not shoot Black and White
male targets who appear on the screen holding handguns or innocuous objects
like wallets and cell phones (Correll et al., 2002). Across numerous studies, a
pattern of racial bias has emerged in both the shoot/don’t shoot decisions and
the speed of such responses, depending on the provided response window (for a
review, see Correll, Hudson, Guillermo, & Ma, 2014; for a meta-analysis, see
Mekawi & Bresin, 2015). When participants are given 630 ms to respond, error
rates are biased such that unarmed Black men are more likely to be mistakenly
shot than unarmed White men, and armed White men are less likely to be
correctly shot than armed Black men. If the response deadline is extended to
850 ms, participants are typically faster to shoot armed Black men, but slower to
not shoot unarmed Black men, compared to White targets. These responses
may, in part, be driven by the activation of cultural stereotypes that link
Black Americans to danger and threat (Correll et al., 2002; Correll, Urland,
& Ito, 2006; but see Mekawi & Bresin, 2015 for further discussion). As such,
various groups of individuals demonstrate a form of shooter bias on the task,
including egalitarian-minded White undergraduates and Black community
members (Correll et al., 2002), as well as actual police officers (particularly in
their reaction times; Correll et al., 2007; Cox, Devine, Plant, & Schwartz, 2014;
Sim, Correll, & Sadler, 2013).

These past findings have established that the automatic processing of the
target’s race can influence the decision to shoot. However, we propose that
the police uniform, and the power associated with it, may be another important
variable that has been previously overlooked in understanding shooting deci-
sions. To explore how physically adopting the role of a police officer may affect
this behavioral outcome, we turned to the literature on the psychological effects
of clothing.

The psychological effects of clothing and power

A plethora of research has investigated how clothing can affect both social and
self-perception (for a review, see K. Johnson, Lennon, & Rudd, 2014).
One research focus within the field has been comparing formal versus casual
styles of dress. Individuals who wear formal clothing are perceived by others to
be more competent across a variety of occupational contexts (Behling &
Williams, 1991; Furnham, Chan, & Wilson, 2013; Gherardi, Cameron, West,
& Crossley, 2009). Adopting formal attire can also influence the extent to which
people feel ambitious, authoritative, trustworthy, and powerful (Hannover &
Kühnen, 2002; Peluchette & Karl, 2007; Slepian, Ferber, Gold, & Rutchick,
2015). Thus, clothing can carry specific associations that guide, and perhaps
bias, perceptions of ourselves and others.

Adam and Galinsky (2012) proposed a psychological model to explain how
clothing can influence one’s mental state and behavior. According to the theory

Mendoza and Parks-Stamm 2355

of enclothed cognition, the physical experience of putting on clothing with sym-
bolic meaning can activate its associated concepts to influence psychological
processes and behavioral tendencies. Much like the effect of physical movements
on how humans feel, process information, or behave (e.g., making a fist can
activate the concept of power; Schubert, 2004), a person may come to embody
(or adopt) the characteristics associated with a specific piece of clothing.
As such, this theory is grounded in the literature on embodied cognition,
which posits that physical experiences can trigger related mental concepts
(Niedenthal, Barsalou, Winkielman, Krauth-Gruber, & Ric, 2005).

To test their enclothed cognition hypothesis, Adam and Galinsky (2012) ran
three studies using a lab coat that had been pretested for its association with
attentiveness (see also Van Stockum & DeCaro, 2014). In their first study,
participants who were assigned to wear, versus not wear, a lab coat while com-
pleting a Stroop Task showed more evidence of selective attention in their per-
formance (i.e., made fewer errors on incongruent trials). Participants in their
follow-up studies who put on a doctor’s lab coat demonstrated greater sustained
attention on a visual search task compared to participants who wore a painter’s
coat, saw a doctor’s coat nearby, or reflected on the personal meaning of the
doctor’s coat. These results therefore suggest that enclothed cognition effects are
dependent on the clothing’s symbolic meaning and are distinct from those that
would be caused by basic priming (Adam & Galinsky, 2012). Indeed, enclothed
cognition requires both the physical experience and the attribution of meaning,
as the same article of clothing can have different influences based on personal
associations with the item (e.g., whether a tunic is identified as a cleaning or
nursing uniform; L�opez-Pérez, Ambrona, Wilson, & Khalil, 2016).

The theory of enclothed cognition provides a framework for understanding
how putting on symbolic clothing, such as a police uniform, can influence
behavior. Past research has found that the police uniform can represent
safety, honesty, and helpfulness (Mauro, 1984), but is also linked to authority
(Bell, 1982; Lowenstein, Blank, & Sauer, 2010), legitimacy (Bushman, 1988),
and power (R. R. Johnson, 2005). These latter effects of the uniform are evident
in the pattern of abusive behavior displayed by role-playing police guards in the
classic Stanford Prison Experiment (Haney, Banks, & Zimbardo, 1973), as well
as findings that citizens are more compliant with requests from individuals
wearing police-like uniforms (Bickman, 1974; Bushman, 1988). Furthermore,
police officers have reported the psychological effects of the uniform on self-
perceptions, including feeling a near-celebrity status when wearing it (De
Camargo, 2012). As others have suggested, a police officer is a socially dominant
role (Sidanius, Liu, Pratto, & Shaw, 1994), and the uniform serves as a symbol
of this power (Vrij & Winkel, 1991).

The theory of enclothed cognition would therefore posit that wearing a police
uniform can psychologically impact the wearer in line with the clothing’s asso-
ciations. Recently, Civile and Obhi (2017) examined this possibility by having

2356 Psychological Reports 123(6)

participants wear a police uniform, mechanic overalls, or their own clothing
while completing a series of visual attention tasks. Their first two studies
found that those in a police uniform showed an attentional bias (assessed by
reaction times) toward stimuli involving low-socioeconomic status targets wear-
ing hoodies compared with high-socioeconomic status targets wearing business
suits. Similar effects emerged in a third study that compared participants who
wore a police uniform or simply saw one placed next to them.

Thus, the extant research on enclothed cognition has demonstrated that
attention can be biased by wearing a symbolic article of clothing like a police
uniform. However, research on the social priming of power suggests police
uniforms should also have visible effects for action tendencies related to policing
behavior. For example, power increases approach orientation (Keltner,
Gruenfeld, & Anderson, 2003; Smith & Bargh, 2008) and action (Galinsky,
Gruenfeld, & Magee, 2003; Smith & Bargh, 2008). Anderson and Galinsky
(2006) found that this tendency to act was mediated by optimistic risk percep-
tions, suggesting that the sense of power associated with a police uniform could
increase individuals’ tolerance for risk-taking behaviors, such as shooting poten-
tially innocent targets. In addition, those with power are more likely to make
errors because they are overconfident (See, Morrison, Rothman, & Soll, 2011),
as well as show less distress and compassion toward the plight of others (Van
Kleef et al., 2008).

Together, past studies examining enclothed cognition and power effects sug-
gest that wearing a police uniform should increase the propensity to shoot, even
at the cost of accuracy. We propose this overall power effect on action
initiation should occur independent of the target’s race. Although powerful
individuals have been shown to attend more to stereotype-consistent informa-
tion (e.g., Civile & Obhi, 2017; Fiske & Dépret, 1996), research has generally
demonstrated that power does not increase stereotyping and prejudice except in
perceiving others in instrumental ways (i.e., as a means to achieve one’s goal;
Chen, Lee-Chai, & Bargh, 2001; Fiske, 1993; Guinote, 2007; Overbeck & Park,
2006; Vescio, Snyder, & Butz, 2003). We therefore expected that wearing a
uniform would lower participants’ threshold for shooting targets, regardless
of their race.

Current research

In this study, we examined the effect of wearing a police uniform on Shooter
Task performance using a short response window (i.e., in which bias should be
observed in error rates but not reaction times; Mekawi & Bresin, 2015).
The original point structure by Correll et al. (2002) differentially rewards
correct or incorrect responses based on type of response (i.e., 5 points for
correct rejection, 10 points for hit, �20 points for a false alarm, and �40
points for a miss), producing an average payoff of 25 points for consistently

Mendoza and Parks-Stamm 2357

choosing a shoot response. D. J. Johnson, Cesario, and Pleskac (2018) demon-

strated that giving higher payoffs for shoot decisions significantly affects mul-

tiple aspects of the shooting decision, by increasing participants’ initial bias to

favor a shoot response and creating a learning curve toward shooting over time.

To avoid this biasing influence and more cleanly test our hypothesis, we altered

the point system to equally reward correct responses (þ10 points) and penalize
incorrect responses (�10 points), as well as discourage nonresponsiveness
(�5 points).

Given the aforementioned links between power and action, we hypothesized

that uniformed participants would be more likely to commit errors

overall, especially false alarms. However, because recent research has shown

that behavior may be impacted by personal beliefs about power (Hays &

Goldstein, 2015), we also examined whether participants’ attitudes about the

use of power by police would moderate the tendency to shoot unarmed targets.

Thus, we predicted that individual perceptions regarding the legitimacy of

police power would moderate the predicted effect of wearing a uniform on

shooting decisions.

Method

Participants and design

A total of 191 undergraduates aged 18–22 years (M ¼ 19.19, SD ¼ 1.13) partic-
ipated in our study on video game vigilance and received either partial course

credit or $5 as compensation (there were no performance differences based on

compensation). Our sample closely reflected the racial demographics of

Providence College, with 81% of participants identifying as White, 10% as

Hispanic/Latino, 4% as Black, 2% as Asian, and 3% as other. The study

used a between-subjects design with participants randomly assigned to either

the control or uniform condition.

Measures

Shooter task. The Shooter Task was adapted from Correll et al. (2002) and con-

sisted of 20 practice trials and 80 critical trials presented in the form of a shooter

video game through Inquisit 4 Lab (Millisecond Software LLC, 2015). Each

trial began with a fixation point, followed by a series of one to four randomly

presented background images lasting 500–1000 ms. These images depicted var-

ious real-life settings (e.g., parks, courtyards, offices), with the final image show-

ing a superimposed Black or White male target who was holding a gun or object.

The participants’ goal was to shoot armed targets and not shoot unarmed

targets as quickly as possible by pressing one of two computer keys that were

counterbalanced across participants. A 630 ms response deadline was

2358 Psychological Reports 123(6)

implemented to preclude deliberative responses (as in Correll et al., 2002;

Mendoza, Gollwitzer, & Amodio, 2010), and parallel rewards and penalties

were allotted to correct (þ10) and incorrect (�10) responses, with a smaller
penalty for nonresponse (�5).

Police power attitudes. As part of an online battery that included other sociopo-

litical questions and was administered at the beginning of the semester, all

eligible participants responded to a bipolar item asking how much they

thought the police did not abuse (1) or abused (6) their power, with higher

scores reflecting more negative law enforcement perceptions (M ¼ 3.51, SD¼ 1.25).

Procedure

All consenting participants were run individually and informed that they would

be playing the role of a police officer in a shooter video game. As in prior

Shooter Task studies, participants in the control condition wore their own cloth-

ing while completing the game in a private computer area. Participants in the

uniform condition were told that they would wear a police uniform in order to

add realism to the game. To avoid having the police uniforms feel like themed

costumes, we purchased several sizes of black long sleeve tops and pants from a

company that provides professional gear to law enforcement agencies (www.

511tactical.com). Black uniforms were chosen because they are the most com-

monly used in U.S. police departments (R. R. Johnson, 2013), including the city

where the study took place. Participants were instructed to wear the uniform

over their own clothing along with a police badge on their upper left chest

pocket. Once the participants had changed, the experimenter directed them to

a private computer to read the task instructions and start the game. At the end

of the task, all participants completed a basic demographic questionnaire and

were carefully debriefed verbally and in writing.

Results

Data reduction

Because the task specifically examines stereotyping against African Americans,

we made an a priori decision to exclude data from Black participants (n ¼ 8) due
to the variability in their implicit in-group evaluations (Ashburn-Nardo,

Knowles, & Monteith, 2003; Livingston, 2002). We included data from other

non-White participants for statistical power (n ¼ 28), but the small representa-
tion of students of color in the sample did not allow us to appropriately test for

the possible moderating effect of participant race in our study. The pattern and

significance of the reported results did not change when we only analyzed data

from White participants or included data from Black participants.

Mendoza and Parks-Stamm 2359

http://www.511tactical.com

http://www.511tactical.com

Task performance was primarily assessed through error rates given our use of
the 630 ms respond deadline (as in Correll et al., 2002 and Mendoza et al., 2010).
Data were excluded from two participants who failed to follow the uniform
instructions, as well as from one participant who reported during the debriefing
process learning about the study’s hypotheses from a friend. In addition, two

participants were excluded for having a mean time-out rate that exceeded three
standard deviations. Our analyses were therefore conducted using a final sample
of 178 participants (84% White, 65% female) who were comparably distributed
into the two conditions (ncontrol ¼ 91, nuniform ¼ 87).1

Police uniform effects on task performance

To examine our predicted uniform effects on task performance, we submitted
error rates to a 2 (Race: White vs. Black) � 2 (Object: gun vs. no gun)
� 2 (Condition: no uniform vs. uniform) mixed-model analysis of variance.2
As seen in Figure 1, we found the predicted condition � object interaction,
F(1, 176) ¼ 6.10, p ¼ .014, g2p ¼ .033, as well as the expected main effect of con-
dition, F(1, 176) ¼ 4.43, p ¼ .037, g2p¼ .025. Deconstructing the interaction
revealed that participants in the uniform condition were more likely to
commit a false alarm by erroneously shooting unarmed targets (M ¼ .32,
SD ¼ .15) than those in the control condition (M ¼ .26, SD ¼ .12), F(1, 176) ¼
7.62, p ¼ .006, g2p ¼ .042, but uniform participants (M ¼ .23, SD ¼ .11) did not
differ from control participants in their gun trial errors (i.e., failing to shoot an
armed target; M ¼ .22, SD ¼ .10), F < 1. Given that we altered the original point system, it was also important for us to validate the task by examining whether it elicited the stereotypical pattern of shooter bias across conditions. The critical race � object interaction was observed, F(1, 176) ¼ 16.39, p < .001, g2p¼ .085, along with a main effect of Figure 1. Mean error rates by object type and target race for the control and uniform conditions. 2360 Psychological Reports 123(6) object, F(1, 176) ¼ 32.41, p < .001, g2p¼ .156, reflecting a tendency to make more errors on no-gun versus gun trials. Simple effect tests following the significant interaction revealed that participants were significantly more likely to not shoot armed White (M ¼ .25, SD ¼ .12) versus Black (M ¼.21, SD ¼ .13) targets, F(1, 176) ¼ 19.81, p < .001, g2p¼ .101, and had a greater tendency to shoot unarmed Black (M ¼ .30, SD ¼ .17) versus White (M ¼ .27, SD ¼ .16) targets, F(1, 176) ¼ 3.43, p ¼ .066, g2p¼ .019. Thus, the task produced the typical pattern of shooter bias against Black targets. No other significant main effects or interactions emerged (all ps > .17).

To further investigate the impact of the police uniform on task responses, we
applied signal detection theory by following the recommendations of Stanislaw
and Todorov (1999). Specifically, we used the z-scored hit rate (i.e., correct
responses on gun trials) and false alarm (i.e., errors on no-gun trials) averages
to calculate the decision criterion c, which reflects the participants’ threshold for
shooting versus not shooting a target. Deviations from zero in a positive direc-
tion indicate a bias favoring the don’t shoot response and deviations in the
negative direction indicate a shoot bias (Mekawi & Bresin, 2015). Consistent
with our expectation that wearing a uniform would produce a relative shooting
bias, there was a significant difference between the two groups, t(176) ¼ 2.08,
p ¼ .039, such that uniformed participants (M ¼ �.09, SD ¼ .60) demonstrated a
more liberal shooting threshold than control participants (M ¼ .08, SD ¼ .47).

Police power attitudes

Finally, we tested whether the observed uniform effects on participants’ tenden-
cy to shoot unarmed targets would be moderated by their attitudes about police
power. The no-gun error rate (i.e., shooting unarmed targets) was entered as the
outcome in a regression that included the effect-coded condition (control ¼ �1,
uniform ¼ 1), the mean-centered police power item, and the interaction between
these two predictors. The regression model explained a significant amount
of variance in no-gun error rates, F(3, 172) ¼ 4.25, p ¼ .006, R2 ¼ .069.
The expected interaction between uniform condition and attitude about police
power was marginally significant, b ¼ �.02, t(174) ¼ 1.76, p ¼ .081, 95%
confidence interval (CI) [�.031, .002], suggesting that attitudes about power
moderated the effect of wearing a uniform on shooting unarmed targets.
The effect of condition remained significant, b ¼ .03, t(173) ¼ 2.78, p ¼ .003,
95% CI [.01, .051], but the item measuring participants’ attitudes about police
power was not a significant predictor on its own, b ¼ �.01, t(174) ¼ 1.27,
p ¼ .207, 95% CI [�.027, .006].

To better understand how the relationship between participant attitudes and
task behavior differed by condition, we ran separate follow-up regressions using
the police power item as a predictor of no-gun errors. Within the uniform con-
dition, perceiving that the police did not abuse their power produced a marginal

Mendoza and Parks-Stamm 2361

increase in shooting unarmed targets, b ¼ �.03, t(85) ¼ 1.83, p ¼ .07, 95% CI
[�.053, .002], suggesting that the uniform’s increase in the propensity to shoot
was found particularly among those with positive attitudes about police use

of power. By comparison, no relationship between power perceptions and

shooting errors emerged within the control condition, b ¼ .004, t(89) ¼ .41,
p ¼ .68, 95% CI [�.016, .024].3 Participants who had more positive attitudes
about police power were more likely to shoot unarmed targets in the uniform

condition only (Figure 2).

Discussion

High-profile shootings of unarmed civilians have impacted public perceptions of

police officers across America and sparked national debates about the factors

that contribute to systematic forms of bias in law enforcement (Nix & Wolfe,

2017; Pew Research Center, 2017). In this study, we moved beyond the well-

established influence of suspect race and investigated how wearing clothing

imbued with power can affect shooting decisions. In line with our theorizing,

we found that participants wearing a police uniform possessed a lower threshold

for shooting, resulting in more unarmed targets being wrongly shot.

Theoretical and practical implications

Our findings have implications for the enclothed cognition literature. All par-

ticipants in our study were playing the role of a police officer in a shooter game,

Figure 2. Predicted values of no-gun error rates (i.e., the shooting of unarmed targets),
illustrating the interaction between police attitudes and uniform condition. Perceptions of
police power were computed at one standard deviation above and below the mean.

2362 Psychological Reports 123(6)

which should have activated similar goals and values across conditions.

Yet, wearing the police uniform uniquely impacted participant behavior in a

manner that also reflected their personal attitudes. Those who were wary of

police abusing their power tended not to show an increase in shooting unarmed

targets as a result of wearing a police uniform, whereas those who were sup-

portive of police power were more likely to shoot unarmed civilians while wear-

ing the uniform. This moderating effect of participants’ attitudes associated with

the article of clothing is consistent with recent research showing that the sym-

bolic meaning of an article of clothing is an independent factor in determining

the effect on one’s cognitions and behavior (L�opez-Pérez et al., 2016).
Our behavioral findings therefore extend our understanding of enclothed

cognition theory by demonstrating that its effects are not limited to general

attentional processes, and that the impact of physically wearing symbolic cloth-

ing can be influenced by more personal associations.
Consistent with prior Shooter Task research, we also found the expected race

by object interaction: participants were more likely to not shoot armed White

versus Black men and were marginally more likely to shoot unarmed Black

versus White men. This pattern emerged across conditions, confirming that

implicit stereotypes do impact shooting decisions, but wearing a police

uniform does not exacerbate this racial bias. Similarly, Civile and Obhi (2017:

Study 1 and Study 2) did not find an effect of wearing a police uniform on

attention toward Black compared to White faces. Although these findings might

seem surprising when considering real-life patterns of police shootings, they are

consistent with prior research showing that power predominantly affects behav-

ior by increasing one’s action orientation (Galinsky et al., 2003), rather than

selectively increasing stereotyping.
One practical implication of this research is that distrust of police power may

actually be beneficial for shooting decision accuracy. In our study, only partic-

ipants who viewed police power as benign showed a higher error rate for shoot-

ing unarmed targets when wearing a police uniform, suggesting that only those

who have positive attitudes about police power feel empowered and emboldened

to action by embodying this role. Those with concerns about …