Historically, in the wake of interpersonal violence, the re

Historically, in the wake of interpersonal violence,
the restriction of individual freedoms has often
followed. This pattern can be observed at various
levels of analysis. A recent example at the level of
national policy includes the passage of the contro-
versial USA PATRIOT Act by the US Government
in October, 2001, following the September 11th,
2001 terrorist attacks. Since that time, the Act has
been widely criticized for weakening government
protection of civil liberties. Prior research has
demonstrated that surveillance by an authority
figure(s) in itself tends to be experienced as con-
trolling (Lepper & Greene, 1975). Towns and cities
routinely institute curfews, along with various
other restrictions of freedom, following violent
riots. At the person-level, parents, teachers, and
various other authority figures very typically
respond to violent behavior by exercising more
control and taking away the rights of others to
choose. Certainly, these measures of increased
control and restricted freedom are effective toward
achieving some desired ends, at least temporarily,
Interpersonal control, dehumanization,
and violence: A self-determination
theory perspective
Arlen C. Moller1,2 and Edward L. Deci3
Abstract
Interpersonally controlling approaches are often used to keep

individuals in line, ostensibly in order to
create a safer, more civilized society. Ironically, emerging
research findings indicate that when people feel
controlled, they often respond by behaving in a less civilized,
more antisocial manner (Gagné, 2003; Knee,
Neighbors, & Vietor, 2001; Mask, Blanchard, Amiot, &
Deshaies, 2005; McHoskey, 1999). The present
research investigation explored whether a process of
mechanistic dehumanization might help to explain
the observed relation between interpersonal control and
antisocial behavior, specifically with regard to
tendencies toward violence. The results indicated that a
significant relation between interpersonal control and
tendencies toward interpersonal violence was partially mediated
by perceived mechanistic dehumanization.
Keywords
aggression, autonomy, dehumanization, hostility, interpersonal
control, mechanistic, self-determination
theory, violence
Paper received 27 December 2008; revised version accepted 31
August 2009.
1Northwestern University
2Gettysburg College
3University of Rochester
Corresponding author:
Arlen C. Moller, Department of Preventive Medicine, 680 N
Lake Shore Dr., Suite 1220, Chicago, IL, USA
[email: [email protected]]
Article
Group Processes & Intergroup Relations

13(1) 41–53
© The Author(s) 2009
Reprints and permission: http://www.
sagepub.co.uk/journalsPermission.nav
DOI: 10.1177/1368430209350318
http://gpi.sagepub.com
G
P
I
R
Group Processes &
Intergroup Relations
42 Group Processes & Intergroup Relations 13(1)
and have much intuitive appeal. However, emerg-
ing research on self-determination theory suggests
that these controlling strategies may also have
unintended and ironic consequences. Studies on
interpersonal control indicate that when people
feel that their autonomy has been thwarted, they
often respond by behaving in an even less civilized
and more antisocial manner (Duriez, Vansteenkiste,
Soenens, & De Witte, 2007; Gagné, 2003; Kernis,
1982; Knee et al., 2001; Mask et al., 2005;
McHoskey, 1999). The present research investiga-
tion explored whether the observed relation
between interpersonal control and antisocial ten-
dencies might be at least partially explained by a
process of dehumanization. We review here the

extant evidence for this model, including previous
research indicating that dehumanization is also
predictive of antisocial behavior (Chalk &
Jonassohn, 1990; Goff, Eberhardt, Williams, &
Jackson, 2008; Haslam, 2006; Kelman, 1976;
McAlister, Bandura, & Owen, 2005). We also pres-
ent new data linking the experience of being con-
trolled with mechanistic dehumanization; that is,
feeling less human, and viewing humans as more
machine-like.
Defining autonomy versus
interpersonal control
Self-determination theory (SDT) posits that all
human beings share a basic and universal psycho-
logical need for autonomy (Deci & Ryan, 1985a,
2000, 2002, 2008). In this framework, autonomy
is defined as a subjective experience, character-
ized by feeling free and by endorsing one’s
actions. In particular, the experience of auton-
omy is characterized by feeling free of interper-
sonal coercion. In accord with SDT, when people
feel more autonomous, they experience greater
psychological and physical well-being, they are
happier and healthier. However, to the degree
that satisfaction of the need for autonomy is
thwarted, research findings indicate that people
suffer both psychological and physically.
The psychological need for autonomy is
thwarted, namely, by forces related to interpersonal
control—the attempts by other people to pres-
sure, manipulate, or otherwise influence one’s
will. These forces not only include overt tactics,
such as the use of tangible rewards and punish-

ments used to control people, but also more
subtle forms of control, such as the use of con-
tingent regard. People are contingently regard-
ing in so far as their love and affection are given
or withdrawn contingently on another person’s
behavior. Controlling language can also be used
to subtly pressure people, and includes words
such as, should, must, and have to (i.e., telling
someone, “you should really work harder”).
Finally, yet another form of control identified
by self-determination theory involves pressure
that comes from within a person. That is, when
people pressure themselves in such a way that
they do not feel as though they are freely or
entirely endorsing their actions. From a self-
determination theory perspective, these internal
forms of pressure result from a process of
incompletely internalizing (or introjecting) con-
trolling forces that originate outside a person,
and thus can ultimately be traced back to inter-
personal control.
Interpersonal control, antisocial
behavior, and violence
Research linking the experience of interpersonal
control to antisocial behavior and violence has
begun to accumulate in a variety of forms. We
begin by reviewing this evidence.
An early study, conducted by Kernis (1982),
investigated the influence of three motivational
orientations (autonomy, control, and impersonal)
on the type of anger expression and degree of
subsequent aggressiveness following a self-
esteem threat. The autonomy orientation is char-
acterized by seeing one’s behavior as freely

chosen, whereas the control orientation is charac-
terized by seeing one’s actions as controlled by
external contingencies, such as rewards and pun-
ishments. The impersonal orientation is the
extent to which a person believes that attaining
desired outcomes is beyond his or her control
Moller and Deci 43
and that achievement is largely a matter of luck or
fate. The results indicated that those who scored
higher in autonomy orientation (felt more free)
behaved less aggressively in the lab, while higher
scores on both the control and impersonal orien-
tation scales were related to more self-derogation
(i.e., self-directed aggression). A later study by
Knee and colleagues (2001) demonstrated that
control motivation orientation also predicted
feeling more driving anger as a result of other
drivers’ actions, and was associated with more
aggressive driving and more traffic citations. A
follow-up study by Neighbors, Vietor, and Knee
(2002) monitored 111 participants’ experiences
driving throughout a 10-day period. Again, con-
trol orientation was related to more anger and
aggression while driving.
In a related line of research on prosocial ver-
sus antisocial behaviors, Gagné (2003) found that
an autonomy orientation strongly predicted pro-
social activities both in general, across different
contexts (Study 1), and, specifically, at work
(Study 2). Also, autonomy support from parents
and managers were each marginally significant

predictors of prosocial behavior. Gagné ran sev-
eral meditational models demonstrating that sat-
isfaction of the psychological need for autonomy
partially mediated the relation between autonomy
orientation and prosocial behavior, and fully
mediated the relation between autonomy support
and prosocial behavior, in both studies. Mask and
colleagues (2005) also investigated the relation
between trait-level autonomy and prosocial
behaviors. A measure of trait-level autonomy, or
self-determination, predicted more prosocial
behavior (e.g., helping others) and less moral dis-
engagement, less interpersonal harm (e.g., being
verbally aggressive), and less aggressive driving-
related behaviors (e.g., driving drunk).
The experience of being controlled can also
be operationalized by assessing the nature of
one’s goals or aspirations, as extrinsic goals (e.g.,
financial success) are understood to be more con-
trolled, while intrinsic goals (e.g., building com-
munity) are understood to be more autonomous.
In line with this, McHoskey (1999) found that a
control ori entation and extrinsic (controlled)
goals were significantly related to having greater
Machiavellianism, defined as one’s willingness to
manipulate others, while autonomy orientation
and intrinsic (autonomous) goals were signifi-
cantly related to Machiavellianism in the inverse
direction. Further, McHoskey found that auton-
omy orientation was negatively related to nihil-
ism, while control orientation was positively
related to self-estrangement and antisocial behav-
iors (such as cheating in an exam, plagiarism,
stealing, vandalism, getting drunk several nights a

week, promiscuity, and being arrested for driving
while intoxicated). Recently, Duriez and col-
leagues (2007) followed up on this work, examin-
ing the relation between extrinsic (controlled) and
intrinsic (autonomous) goals, right-wing authori-
tarianism, social dominance, and racial prejudice.
The authors found that an emphasis on extrinsic
goals was positively related to prejudice across
two studies, and that social dominance partially
mediated the relation. The partial mediation
found in these studies, however, leaves open
the potential for future research to identify
other important process variables, such as
dehumanization.
Dehumanization, antisocial behavior,
and violence
Dehumanization is a psychological construct
which has been very broadly defined as the denial
of humanness to others, the negative conse-
quences of which have been well documented
empirically. These consequences include various
forms of antisocial behavior, especially violence
directed toward those dehumanized. Although an
exhaustive review of this literature is beyond the
scope of the present article (for a recent, more
comprehensive review, see Haslam, 2006), we
highlight here some of the most robust and com-
pelling findings.
Goff and colleagues (2008) found evidence
that White participants implicitly associated
Blacks and apes (i.e., animalistic dehumaniza-
tion), and this association in turn was related to
increased endorsement of violence against Black
suspects in a criminal justice context. In a

follow-up, archival study the authors found that
news articles written about Blacks convicted of
capital crimes were more likely to contain ape-
relevant language than news articles written about
White convicts. Further, in the archival data,
those who were portrayed as more apelike in
these articles were more likely to be executed by
the state than those who were not. These findings
reported by Goff and colleagues, taken together
with Duriez and colleagues’ (2007) findings that
having more controlled goals was related to
greater racial prejudice, are consistent with the
assertion that dehumanization may play a role in
linking interpersonal control with racial prejudice
and endorsed violence against Blacks.
Following the September 11th, 2001 terrorist
attacks on the United States, McAlister et al.,
(2005) interviewed 1,499 participants in order to
explore the tendency for people to support the
use of violent military force in retaliation.
Endorsement of military force increased signifi-
cantly following the terrorist attacks, and the
results suggest that one important mediator of
endorsing military force involved dehumanizing
the enemy (e.g., “terrorists do not deserve to be
treated like human beings”, and “enemy rulers
and their followers are no better than animals”),
as dehumanization increased significantly from
pre- to post-September 11th.

Yet another example of dehumanization being
related to violence concerns the use of dehuman-
ization in connection with genocidal conflicts
(Chalk & Jonassohn, 1990; Haslam, 2006;
Kelman, 1976). Historical accounts reveal that
the perpetrators of genocidal violence have often
espoused ideologies that likened the victims to
vermin and various other “lower” life forms. This
form of dehumanization was documented in
cases that include the Jews during the Holocaust,
Bosnians during the Balkan wars, and Tutsis dur-
ing the genocide in Rwanda. A number of authors
have argued that the process of dehumanization
may make it possible for humans to inflict greater
harm on others by virtue of allowing them to
exclude a group or individual from moral consid-
eration, also known as moral disengagement
(Bandura, 2002; Bandura, Barbaranelli, Caprara,
& Pastorelli, 1996; Bandura, Underwood, &
Fromson, 1975; Kelman, 1976; Opotow, 1990).
Castano and Giner-Sorolla (2006) explored this
relation across three experiments; specifically, the
tendency for people to infrahumanize (an implicit
form of dehumanization) an out-group in response
to interpersonal violence. That is, when participants
were told that their in-group (humans, British,
White Americans) had perpetuated mass killing of
an out-group (aliens, Australian Aborigines, and
Native Americans), they responded by infrahuman-
izing the out-group more, but only if they also per-
ceived a collective responsibility for their in-group’s
actions. The authors theorized that the process of
infrahumanization, in this case, was a strategy for
people to escape collective guilt and reestablish psy-

chological equanimity.
Interpersonal control and
dehumanization
Haslam (2006) recently reviewed the broad litera-
ture on dehumanization, and developed a new
model, differentiating between two forms of
dehumanization: animalistic and mechanistic.
Animalistic dehumanization involves denying
uniquely human attributes to others, representing
them as animal-like, whereas mechanistic dehu-
manization involves denying human nature to
others, representing them as objects or automata.
The present investigation focused on mechanistic
dehumanization, specifically with regard to what
Montague and Matson (1983) referred to as
“technological dehumanization” or “the reduc-
tion of humans to machines” (p. 8). Montague
and Matson posited that this form of dehuman-
ization is a cultural consequence of postmodern
society’s pursuit of industrialization, robotic effi-
ciency, and regularity, and a number of theorists
have since expressed concern over the potential
consequences of technological dehumanization,
per se (Beckers & Schmidt, 2001; Nissenbaum &
Walker, 1998). In designing the present investiga-
tion, we hypothesized that the experience of
being controlled may lead people to feel less
human themselves, and as a consequence see
both themselves and other human beings as
Moller and Deci 45
objects or machines, as opposed to autonomous,

living organisms. To the extent that animals are
understood to be autonomous, we specifically
hypothesized that interpersonal control may
lead to mechanistic, as opposed to animalistic,
dehumanization.
Although, to the best of our knowledge, this
hypothesis had not been previously tested, sev-
eral lines of research and theory offer converging
indirect support. deCharms (1968), for example,
wrote of experiencing oneself as an origin or pawn
in the context of personal causation, and posited
that when one feels controlled, coerced, or
manipulated by another person, they come to feel
more pawn-like, and effectively less human. The
concept of autonomy as understood in self-
determination theory is derived to a large mea-
sure from deCharms’ work on personal causation,
and the pawn metaphor has been used to charac-
terize feeling controlled since (Ryan & Grolnick,
1986). Kelman (1976) posited that dehumaniza-
tion involves denying a person “identity” —the
perception of the person “as an individual, inde-
pendent and distinguishable from others, capable
of making choices” (p. 301, italics added), thus
recognizing a possible conceptual connection
between autonomy and dehumanization.
Empirical research has already linked the
experience of self-determination to the human
capacity for empathy, which is one aspect used to
define humanness. Mask and colleagues (2005)
found that greater self-determination was related
to more empathy. Further, in this study, empathy
mediated the relation between self-determination
and helpfulness, interpersonal harm, and aggres-

sive driving-related behaviors. A second form of
indirect support linking interpersonal control to
dehumanization concerns the experience of vital-
ity, or life force. Subjective vitality is defined as
the state of feeling alive and alert, and is consid-
ered an aspect of eudaimonic well-being (Ryan
& Deci, 2001). The experience of being con-
trolled has been negatively related to vitality at
both the state level and trait level in numerous
studies (Moller, Deci, & Ryan, 2006; Nix, Ryan,
Manly, & Deci, 1999; Ryan & Deci, 2008; Ryan &
Frederick, 1997).
In the present investigation, we principally
sought to directly test the hypothesis that
interpersonal control is positively related to
dehumanization, specifically mechanistic dehu-
manization. We also sought to replicate and
extend previous research findings linking inter-
personal control and dehumanization to a ten-
dency toward violence, and extend that research
by including a range and variety of previously
underexplored indicators to operationalize this
tendency.
Method
Participants
Some 235 (194 female) adults completed the web
questionnaire. Data was collected on-line during a
five-month period using a web survey method
(February 2005 through June 2005). Participants
were recruited from several sites which host links
to social psychology web research, including:
http://www.socialpsychology.org, http://www.
yahoo.com, http://genpsylab-wexlist.unizh.ch,
http://psych.hanover.edu/research/exponnet.

html. All participation was voluntary and no form
of extrinsic compensation was provided. The
mean age for participants was 25.08 years (range:
18–62). The ethnic make-up was as follows:
76.2% White, 4.3% Asian or Pacific Islander, 3%
Black, 10.6% Hispanic, 5.1% Other. The highest
level of education attained by participants in the
sample ranged from some high school education
to the highest levels of graduate education: 1.7%
some high school, 7.2% high school diploma,
44.3% some college, 11.5% associates degree,
14.0% bachelors degree, 6.4% some graduate
school, 7.2% master’s degree, 7.7% held a PhD,
MD, or JD.
Procedure
Each measure was presented on a separate web-
page in the order described below. The order with
regard to scale presentation was not counterbal-
anced. Instructions specific to each scale were
presented at the top of each page.
Measures
General Causality Orientation Scale (GCOS)
This is an individual difference measure of peo-
ple’s relatively enduring motivational orientations
and was developed for use with individuals who
are at least 17 years of age. Subjects answered three
questions for each of 17 vignettes regarding how
likely they were to interpret events in certain ways
(Deci & Ryan, 1985b). The three questions corre-

sponded to autonomy, control, and impersonal
causality orientations. A 5-point scale was used for
each question (1 = very unlikely, 5 = very likely).
The autonomy orientation assesses the extent to
which a person is oriented toward aspects of the
environment that stimulate autonomous motiva-
tion, are optimally challenging, and provide infor-
mational feedback. The control orientation
assesses the extent to which a person is oriented
toward being controlled by rewards, deadlines,
structures, ego-involvements, and the directives of
others. The impersonal orientation assesses the
extent to which a person believes that attaining
desired outcomes is beyond his or her control and
that achievement is largely a matter of luck or fate.
An example of a vignette from the GCOS is: “You
are a plant supervisor and have been charged with
the task of allotting coffee breaks to three workers
who cannot all break at once.” Participants were
asked to rate how likely they would be to respond
in each of the following ways: (a) “Telling the three
workers the situation and having them work with
you on the schedule”; (b) “Find out from someone
in authority what to do or do what was done in the
past”; or (c) “Simply assigning times that each can
break to avoid any problems.” Likelihood ratings
to each hypothetical response correspond to dif-
ferent motivational orientations (in this case, auto-
nomy orientation, control orientation, and
impersonal orientation, respectively). The internal
reliability of each subscale was acceptably high:
Cronbach’s alpha levels of .86 (autonomy orienta-
tion), .77 (control orientation), and .84 (impersonal
orientation).
Self-Determination Scale (SDS) The SDS was

designed to assess individual differences in the
extent to which people tend to function in a self-
determined way (Sheldon, 1995; Sheldon, Ryan, &
Reis, 1996). It is thus considered a relatively endur-
ing aspect of people’s personalities which reflects:
(1) being more aware of their feelings and their
sense of self; and (2) feeling a sense of choice with
respect to their behavior. The SDS is a 10-item scale,
with two 5-item subscales. The first subscale is
awareness of oneself, and the second is perceived
choice in one’s actions. Each items consists of two
statements, and participants are asked to rate the
degree to which one statement is more true of them
on a 7-point scale (1 = only statement A is true of
me; 7 = only statement B is true of me). For exam-
ple, “A. I always feel like I choose the things I do; B.
I sometimes feel that it’s not really me choosing the
things I do” (Perceived Choice); and “A. My emo-
tions sometimes seem alien to me; B. My emotions
always seem to belong to me” (Awareness). The
internal reliability of each subscale was acceptably
high; Cronbach’s alpha levels were .73 (Awareness)
and .77 (Perceived Choice).
Autonomy psychological need satisfaction
The autonomy subscale from the Basic
Psychological Need Satisfaction (BPNS) scale was
used to assess the degree to which participants
experienced satisfaction of the basic psychological
need for autonomy in their lives (Baard, Deci, &
Ryan, 2004; La Guardia, Ryan, Couchman, & Deci,
2000). The self-report measure consists of seven
items (e.g., “I feel free to be who I am”) rated on a
7-point scale (1 = not at all; 7 = very much). The
internal reliability was acceptably high; Cronbach’s

alpha was .73.
Composite interpersonal control A compos-
ite measure of interpersonal control was created
by standardizing the five subscales related to
this core concept (GCOS–Autonomy Orien-
tation; GCOS–Control Orientation; SDS–
Awareness; SDS–Perceived Choice; BPNS–
Autonomy), reverse scoring when appropriate,
and summing the resulting scores. This compos-
ite measure broadly represents the extent to
which one feels they have been controlled by
other people in their life.
Moller and Deci 47
Mechanistic dehumanization Aron and col-
leagues (Aron, Aron, & Smollan, 1992; Aron,
Melinat, Aron, Vallone, & Bator, 1997) developed
the Inclusion of Other in Self (IOS) measure to
assess closeness between two individuals by asking
about the degree to which individuals feel that
another person, initially a romantic partner, is a
part of their conceptualization of self. The mea-
sure has since been widely adapted to measure
individuals’ experiences of closeness with peers,
parents, family members, organizations, and
groups (e.g., Deci, La Guardia, Moller, Scheiner, &
Ryan, 2006). The IOS consists of seven pairs
of circles labeled Self and Other, that overlap to
various degrees, creating a 7-point, interval scale.
Participants select the pair that best describes their
relationship. For our purpose of measuring dehu-
manization, participants selected circles represent-

ing the degree of overlap between “human beings”
and “machines” (collective mechanistic dehuman-
ization; 1 item) and between “me” and “machines”
(individual mechanistic dehumanization; 1 item).
The correlation between responses on these two
items was r = .41, p < .001, a moderate size correla-
tion indicating that although collective and indi-
vidual forms of mechanistic dehumanization share
significant overlap, they are nonetheless conceptu-
ally distinguishable as well. Given that no hypoth-
eses were postulated with regard to distinguishing
these two forms of mechanistic dehumanization,
the two items were also z-scored and combined to
create a composite indicator of overall mechanistic
dehumanization.
Aggression Trait-level aggression was assessed
using Buss and Perry’s (1992) 29-item Aggression
Measure. The measure includes subscales assess-
ing: Physical Aggression (9 items), Verbal
Aggression (5 items), Anger (7 items), and Hostility
(8 items) subscales. Statements (e.g., “Once in a
while I can’t control the urge to strike another per-
son”) are rated on a 7-point scale (1 = extremely
uncharacteristic of me; 7 = extremely characteris-
tic of me). The internal reliability of each subscale
was acceptably high; Cronbach’s alpha levels were
.82 (Physical Aggression), .72 (Verbal Aggression),
.85 (Anger), and .85 (Hostility).
Acceptance of violence Trait-level acceptance
of violence was assessed using the 14-item
Maudsley Violence Questionnaire (MVQ; Walker,
2005). Statements (e.g., “I am totally against vio-
lence”) are rated as either true or false. The internal
reliability of this scale was acceptably high;

Cronbach’s alpha was .79.
Composite violence A composite measure of
aggression was created by standardizing the five
subscales related to this core concept (AM–
Physical Aggression; AM–Verbal Aggression;
AM–Anger; AM–Hostility; and Acceptance of
Violence), reverse scoring when appropriate, and
summing the resulting scores. This composite
measure broadly represents one’s tendency toward
the expression of interpersonal violence.
Results
A linear regression approach was used to
analyze these data. The models run were organized
by testing for evidence of the following relations:
(1) interpersonal control predicting tendencies
toward interpersonal violence; (2) interpersonal
control predicting dehumanization; (3) dehu-
manization predicting tendencies toward interper-
sonal violence; (4) mediation; and (5) moderation.
Interpersonal control ↔ interpersonal violence
Interpersonal control was operationalized in five
different ways (GCOS–Autonomy Orientation;
GCOS–Control Orientation; SDS–Awareness;
SDS–Perceived Choice; BPNS–autonomy), as
well as with a composite measure of all five.
Tendencies toward interpersonal violence were
also operationalized in five different ways (AM–
Physical Aggression; AM–Verbal Aggression;
AM–Anger; AM–Hostility; and Acceptance of
Violence), as well as a composite measure. The
correlations between the measures of inter-
personal control and tendencies toward inter-
personal violence are summarized in Table 1.

Consistently, trait-level measures of experienced
control were positively correlated with tendencies
toward interpersonal violence, while measures of
experienced autonomy were negatively correlated
with tendencies toward interpersonal violence.
This pattern of correlations remained significant
when controlling for gender.
Interpersonal control ↔ mechanistic
dehumanization
The correlations between measures of interper-
sonal control and dehumanization are summa-
rized in Table 2. Consistently, trait-level measures
of experienced control were positively corre-
lated with the experience of dehumanization,
while measures of experienced autonomy were
negative correlated with the experience of dehu-
manization. This pattern of correlations remained
significant when controlling for gender.
Mechanistic dehumanization ↔ interpersonal
violence
The correlations between measures of mechanis-
tic dehumanization and tendencies toward inter-
personal violence are summarized in Table 3. A
consistent pattern of significant positive correla-
tions was found between dehumanization and
indicators of a tendency toward interpersonal vio-
lence (physical aggression, anger, and hostility),
with the notable exception of verbal aggression.

This pattern of correlations remained significant
Mediation: Interpersonal control →
mechanistic dehumanization →
interpersonal violence
Next, we tested whether mechanistic dehumaniza-
tion would mediate the relation between interper-
sonal control and tendencies toward interpersonal
violence. Baron and Kenny (1986) presented four
steps for establishing mediation. Step 1 involves
showing that the independent variable (i.e., inter-
personal control) is related to the outcome (i.e., a
tendency toward interpersonal violence). This
requirement was met; composite interpersonal
control was significantly related to composite
interpersonal violence, β = .49, p < .001. Step 2
involves showing that the independent variable is
related to the mediator (i.e., mechanistic dehu-
manization). This requirement was met, compos-
ite interpersonal control was significantly related
to composite mechanistic dehumanization, β =
.25, p < .001. Step 3 requires that the mediator
affect the outcome variable, controlling for the
independent variable. This requirement was met;
composite mechanistic dehumanization was
Table 1. Interpersonal control ↔ tendencies toward
Physical
aggress.
Verbal
aggress.

Anger
Hostility
Accept
violence
Composite
aggress.
GCOS
Autonomous orientation -.23** -.01 -.17* -.26** -.26 -.25**
Controlled orientation .34** .33** .21** .34** .30** .41**
Impersonal orientation .18** .05 .23** .48** .10 .28**
SDS
Perceived choice -.19** -.13 -.28** -.44** -.08 -.36**
Awareness -.28** -.14* -.27** -.54** -.10 -.31**
BPNS
Autonomy satisfaction -.25** -.14* .29** -.55** -.15* -.37**
Composite control .41** .24** .33** .53** .30** .49**
Note: Aggress. = Aggression; * p < .05; ** p < .01.
Moller and Deci 49
significantly related to composite interpersonal
violence, controlling for composite interper-
sonal control, β = .17, p < .05. The fourth and

final step for establishing mediation looks at the
relation between the initial predictor variable
and the outcome, controlling for the mediator.
If this effect drops to zero, there is full media-
tion, if it drops significantly (Sobel, 1982), there
is partial mediation. The requirements for par-
tial mediation were met. When composite mech-
anistic dehumanization was controlled for,
the relation between composite interpersonal
control and composite interpersonal violence
dropped (from β = .49 to β = .45); a Sobel test
confirmed that this drop was significant, Sobel’s
z = 2.43, p = .01. Each of the models required
for establishing mediation remained significant
Moderation
Several regression models were run exploring the
potential for direct relations discussed above to
be moderated. In each case, the interaction and
main effects were treated as continuous variables.
First, composite interpersonal control, composite
dehumanization, and the two-way interaction of
these variables were regressed onto composite
interpersonal violence as an outcome; the interac-
tion was nonsignificant, β = -.01, t(229) = -0.16,
p = .87. Next, regression models were run testing
whether age or sex moderated the relations
between either composite interpersonal control
and composite interpersonal violence, or com-
posite dehumanization and composite interper-
sonal violence; none of these interactions
approached significance, all ts < 1.21.
Table 2. Interpersonal control ↔ mechanistic dehumanization

Human–machine Me–machine Composite dehumanization
GCOS
Autonomous orientation -.09 -.05 -.08
Controlled orientation .10 .29** .23**
Impersonal orientation .17** .20** .22**
SDS
Perceived choice -.29** -.22** -.31**
Awareness -.25** -.15* -.23**
BPNS
Autonomy satisfaction -.38** -.24** -.37**
Composite control .18** .25** .25**
Note: * p < .05; ** p < .01.
Table 3. Mechanistic dehumanization ↔ tendencies toward
Physical
aggress.
Verbal
aggress.
Anger
Hostility
Accept

violence
Composite
Aggress.
Human–machine .21** .03 .17* .29** .06 .20**
Me–machine .22** .12 .13* .27** .16* .24**
Composite dehumanization .25** .09 .18* .33** .13* .26**
Note: Aggress. = Aggression; * p < .05; ** p < .01.
General discussion
The results of the present investigation replicated
and extended the evidence for two important
conceptual patterns: (1) the relation between the
experience of interpersonal control and a ten-
dency toward interpersonal violence; and (2) the
relation between experiencing interpersonal con-
trol and the experience of dehumanization.
Further, it introduced evidence for an important
new conceptual connection relating the experi-
ence of interpersonal control to the experience
of dehumanization. The data suggest that trait-
level experiences of interpersonal control were
positively related to both interpersonal violence,
and to the experience of mechanistic dehuman-
ization, whereas feeling more autonomous was
negatively related to these variables. Especially
important were the findings linking the experi-
ence of interpersonal control to greater mecha-
nistic dehumanization (and greater autonomy to

feeling more human, and less machine), the first
data of their kind. Further, the relation between
interpersonal control and interpersonal violence
was shown to be partially explained (or mediated)
by mechanistic dehumanization, although the
amount of variance accounted for was very small.
Limitations and future directions
The correlational nature of the data from this
study represents an important limitation, specifi-
cally with regard to establishing causality. Future
studies will need to investigate the relations
explored here using experimental designs; specifi-
cally, manipulating the experience of control (or
the provision of choice) in order to test whether
being more controlled indeed causes people to feel
less human, and in turn behave more violently.
Another limitation of the present investigation
was a reliance on self-report measures of the cen-
tral concepts. Violent behavioral tendencies can
also be assessed more directly in the lab; for
example, by using the allocation of hot sauce
(Lieberman, Solomon, Greenberg, & McGregor,
1999) or volume of noise-blast administered
(Taylor, 1967). Additionally, given that mechanistic
dehumanization was found to partially (as opposed
to fully) mediate the relation between interpersonal
control and tendencies toward interpersonal vio-
lence, and that only a small degree of variance was
accounted for, this implies that the relation
between interpersonal control and interpersonal
violence is complex and that multiple psycholo-
gical constructs, in addition to mechanistic dehu-
manization, may be required to offer a more
complete account.

The process through which interpersonal con-
trol may lead to dehumanization and interper-
sonal violence could be a developmental process,
not easily observed in a lab setting. For this rea-
son, future studies should employ a longitudinal
design to help establish the direction of the rela-
tions explored here. For example, by investigating
how controlling parenting styles, such as contin-
gent regard toward children, may lead to dehu-
manization and violence among adults. Such
studies could also include indicators of overt vio-
lence, such as convictions for violent crimes
among at-risk populations.
Yet another future direction for this line of
research concerns relating the feeling of being
more controlled by others to different forms of
violence. Specifically, self-determination theory
differentiates between controlled motivation that
is entirely a function of external contingencies
(e.g., tangible rewards or punishments), also
referred to as external regulation, and controlled
motivation that is derived from internal contin-
gencies (e.g., pride or shame), also referred to as
introjected regulation (Ryan & Connell, 1989). As
people can feel pressured or controlled by forces
internal or external to the self, the target of one’s
aggression may also vary. That is, aggression and
violence may be focused inwardly against the self
(e.g., suicide, and various forms of self-mutila-
tion), and violence can, of course, be focused
outwardly against others (e.g., murder, and vari-
ous forms of assault). As such, we offer a specu-
lative hypothesis that introjected forms of
controlled motivation may be more strongly

related to inwardly focused violence, whereas
external forms of controlled motivation may be
more strongly related to outwardly focused
Moller and Deci 51
violence. Future research is needed to test this
hypothesis, as well as the possibility that distinct
forms of dehumanization (e.g., me-machine vs.
humans-machines) may play an important medi-
tational role in understanding this predicted
relation.
Finally, although this research study focused
on mechanistic forms of dehumanization, future
research is needed to investigate whether inter-
personal control may also be related to more
animalistic forms of dehumanization. Haslam
(2006) posits that these two categories of dehu-
manization (animalistic and mechanistic) can be
considered conceptually distinct; however, this
does not rule out the possibility that they might
share common antecedents. Interpersonal con-
trol may also be related to animalistic dehuman-
ization. In particular, the concept of animalistic
dehumanization may be more closely related to
the experience of controlling others, as opposed
to being controlled oneself, as animalistic dehu-
manization is understood to include an implicit
vertical comparison and the belittling or degrada-
tion of the other(s). This follow-up hypothesis is
yet another to be explored in the future.
Conclusion

In sum, the present investigation offers a very
early indication of the potential role that dehu-
manization may play in terms of understanding
the complex relation between the psychological
experience of interpersonal control and the ten-
dency toward violence. The preliminary evidence,
however, indicates that authority figures at vari-
ous levels (e.g., policy makers, managers, teachers,
and parents) would do well to temper their ten-
dency to respond to violence with measures that
thwart people’s psychological need for autonomy,
as the ultimate consequence of these strategies
may ironically be to induce even greater violence,
as a function of dehumanization. We anticipate
that further unraveling the complex dynamics
between interpersonal control, dehumanization,
and violence, at various levels of analysis, and in
a wide range of contexts, will be an important
and generative area for future research.
Acknowledgments
The authors would like to thank Gabriela Llenín for
her help creating the web survey used for this research.
Additionally, we thank Guy Roth for his insightful
feedback and creative suggestions for follow-up
research. Finally, a special debt of gratitude is owed to
Rachel Naylor for her humanizing influence during the
composition of this manuscript.
References
Aron, A., Aron, E. N., & Smollan, D. (1992). Inclusion
of other in the self scale and the structure of inter-
personal closeness. Journal of Personality and Social
Psychology, 63, 596–612.

Aron, A., Melinat, E., Aron, E. N., Vallone, R., &
Bator, R. (1997). The experimental generation of
interpersonal closeness: A procedure and some
preliminary findings. Personality and Social Psychology
Bulletin, 23, 363–377.
Baard, P. P., Deci, E. L., & Ryan, R. M. (2004). Intrinsic
need satisfaction: A motivational basis of perform-
ance and well-being in two work settings. Journal of
Applied Social Psychology, 34, 2045–2068.
Bandura, A. (2002). Selective moral disengagement
in the exercise of moral agency. Journal of Moral
Education, 31, 101–119.
Bandura, A., Barbaranelli, C., Caprara, G. V., &
Pastorelli, C. (1996). Mechanisms of moral disen-
gagement in the exercise of moral agency. Journal
of Personality and Social Psychology, 71, 364–374.
Bandura, A., Underwood, B., & Fromson, M. E.
(1975). Disinhibition of aggression through dif-
fusion of responsibility and dehumanization
of victims. Journal of Research in Personality, 9,
253–269.
Baron, R. M., & Kenny, D. A. (1986). The moderator–
mediator variable distinction in social psychological
research: Conceptual, strategic, and statistical con-
siderations. Journal of Personality and Social Psychology,
51, 1173–1182.
Beckers, J. J., & Schmidt, H. G. (2001). The structure of
computer anxiety: A six-factor model. Computers in
Human Behavior, 17, 35–49.

Buss, A. H., & Perry, M. (1992). The Aggression
Questionnaire. Journal of Personality and Social
Castano, E., & Giner-Sorolla, R. (2006). Not quite
human: Infra-humanization as a response to col-
lective responsibility for intergroup killing. Journal
of Personality and Social Psychology, 90, 804–818.
Chalk, F., & Jonassohn, K. (1990). The history and sociol-
ogy of genocide: Analyses and case studies. New Haven,
CT: Yale University Press.
deCharms, R. (1968). Personal causation. New York:
Academic Press.
Deci, E. L., La Guardia, J. G., Moller, A. C., Scheiner,
M. J., & Ryan, R. M. (2006). On the benefits of
giving as well as receiving autonomy support:
Mutuality in close friendship. Personality and Social
Psychology Bulletin, 32, 313–326.
Deci, E. L., & Ryan, R. M. (1985a). Intrinsic motivation
and self-determination in human behavior. New York:
Plenum Publishing Co.
Deci, E. L., & Ryan, R. M. (1985b). The general causal-
ity orientations scale: Self-determination in person-
ality. Journal of Research in Personality, 19, 109–134.
Deci, E. L. & Ryan, R. M. (2000). The “what” and

“why” of goal pursuits: Human needs and the self-
determination of behavior. Psychological Inquiry, 11,
227–268.
Deci, E. L., & Ryan, R. M. (Eds.). (2002). Handbook of
self-determination research. Rochester, NY: University
of Rochester Press.
Deci, E. L., & Ryan, R. M. (2008). Facilitating optimal
motivation and psychological well-being across
life’s domains. Canadian Psychology, 49, 14–23.
Duriez, B., Vansteenkiste, M., Soenens, & De Witte,
H. (2007). The social costs of extrinsic relative to
intrinsic goal pursuits: Their relation with social
dominance and racial and ethnic prejudice. Journal
of Personality, 75, 757–782.
Gagné, M. (2003). The role of autonomy support
and autonomy orientation in prosocial behavioral
engagement. Motivation and Emotion, 27, 199–223.
Goff, P. A., Eberhardt, J. L., Williams, M. J., & Jackson,
M. C. (2008). Not yet human: Implicit knowledge,
historical dehumanization, and contemporary con-
sequences. Journal of Personality and Social Psychology,
94, 292–306.
Haslam, N. (2006). Dehumanization: An integrative review.
Personality and Social Psychology Review, 10, 252–264.
Kelman, H. C. (1976). Violence without restraint:
Reflections on the dehumanization of victims and
victimizers. In G. M. Kren & L. H. Rappoport
(Eds.), Varieties of psychohistory (pp. 282–314). New

York: Springer.
Kernis, M. H. (1982). Motivational orientations, anger,
and aggression in males. Doctoral dissertation,
University of Rochester.
Knee, C. R., Neighbors, C., & Vietor, N. (2001).
Self-determination theory as a framework for
understanding road rage. Journal of Applied Social
La Guardia, J. G., Ryan, R. M., Couchman, C. E., &
Deci, E. L. (2000). Within-person variation in secu-
rity of attachment: A self-determination theory
perspective on attachment, need fulfillment, and
well-being. Journal of Personality and Social Psychology,
79, 367–384.
Lepper, M., & Greene, D. (1975). Turning play into
work: Effects of adult surveillance and extrinisic
motivation. Journal of Personality and Social Psychology,
31, 479–486.
Lieberman, J. D., Solomon, S., Greenberg, J., &
McGregor, H. A. (1999). A hot new way to meas-
ure aggression: Hot sauce allocation. Aggressive
Behavior, 25, 331–348.
Mask, L., Blanchard, C. M., Amiot, C. E., & Deshaies,
J. (2005). Can self-determination benefit more
than the self ? A pathway to prosocial behaviours.
Poster presented at the Society for Personality and
Social Psychology’s Annual Meeting, New Orleans,
Louisiana, January.
McAlister, A. L., Bandura, A., & Owen, S. V. (2005).

Mechanisms of moral disengagement in support
of military force: The impact of Sept. 11. Journal of
Social and Clinical Psychology, 25, 141–165,
McHoskey, J. W. (1999). Machiavellianism, intrinsic
versus extrinsic goals, and social interest: A self-
determination theory analysis. Motivation and
Emotion, 23, 267–283.
Moller, A. C., Deci, E. L., & Ryan, R. M. (2006).
Choice and ego-depletion: The moderating role of
autonomy. Personality and Social Psychology Bulletin, 32,
1024–1036.
Montague, A., & Matson, F. (1983). The dehumanization
of man. New York: McGraw-Hill.
Neighbors, C., Vietor, N. A., & Knee, C. R. (2002). A
motivational model of driving anger and aggression.
Personality and Social Psychology Bulletin, 28, 324–335.
Moller and Deci 53
Nissenbaum, H., & Walker, D. (1998). Will computers
dehumanize education? A grounded approach to
values at risk. Technology in Society, 20, 237–273.
Nix, G. A., Ryan, R. M., Manly, J. B., & Deci, E. L.
(1999). Revitalization through self-regulation: The
effects of autonomous and controlled motivation
on happiness and vitality. Journal of Experimental
Social Psychology, 35, 266–284.
Opotow, S. (1990). Moral exclusion and injustice: An

introduction. Journal of Social Issues, 46, 1–20.
Ryan, R. M., & Connell, J. P. (1989). Perceived locus of
causality and internalization: Examining reasons
for acting in two domains. Journal of Personality and
Social Psychology, 57, 749–761.
Ryan, R. M., & Deci, E. L. (2001). To be happy or to be
self-fulfilled: A review of research on hedonic and
eudaimonic well-being. In S. Fiske (Ed.), Annual
review of psychology (Vol. 52, pp. 141–166). Palo Alto,
CA: Annual Reviews, Inc.
Ryan, R. M., & Deci, E. L. (2008). From ego-depletion
to vitality: Theory and findings concerning the
facilitation of energy available to the self. Social and
Personality Psychology Compass, 2, 702–717.
Ryan, R. M., & Frederick, C. M. (1997). On energy, per-
sonality and health: Subjective vitality as a dynamic
reflection of well-being. Journal of Personality, 65,
529–565.
Ryan, R. M., & Grolnick, W. S. (1986). Origins and pawns
in the classroom: Self-report and projective assess-
ments of individual differences in children’s per-
ceptions. Journal of Personality and Social Psychology, 50,
550–558.
Sheldon, K. M. (1995). Creativity and self-determination
in personality. Creativity Research Journal, 8, 61–72.
Sheldon, K. M., Ryan, R. M., & Reis, H. R. (1996).
What makes for a good day? Competence and
autonomy in the day and in the person. Personality
and Social Psychology Bulletin, 22, 1270–1279.

Sobel, M. E. (1982). Asymptotic confidence intervals
for indirect effects in structural equation mod-
els. In S. Leinhardt (Ed.), Sociological methodology
1982 (pp. 290–312). Washington, DC: American
Sociological Association.
Taylor, S. P. (1967). Aggressive behavior and physiologi -
cal arousal as a function of provocation and the ten-
dency to inhibit aggression. Journal of Personality, 35,
297–310.
Walker, J. S. (2005). The Maudsley Violence Questionnaire:
Initial validation and reliability. Personality and Individual
Differences, 38, 187–207.
Biographical notes
arlen c. moller is a research assistant professor in
the Department of Preventive Medicine at North-
western University. He received his BA in Psy-
chology from Cornell University, and his Ph.D.
in Social and Personality Psychology from
the University of Rochester. His research focuses
on issues related to human motivation and
well-being.
edward l. deci is a professor of psychology and
the Gowen Professor in the Social Sciences in
the Department of Clinical and Social Psy-
chology at the University of Rochester. He
received his Ph.D. in Psychology from Carnegie
Mellon University. His research examines a vari-
ety of issues in human motivation, isolating
basic processes and testing their application to
education, health care, parenting, mental health,
and work organizations in the US and across

cultures. Using both laboratory and field
methods, his work focuses primarily on the
nature and development of self-determination.
For more detail, please see: http://www.psych.
rochester.edu/SDT/
Parental Desensitization to Violence and Sex in Movies
WHAT’S KNOWN ON THIS SUBJECT: Movie ratings designed
to
warn parents about violence and sexual content have permitted
increasing amounts of each in popular films. One potential
explanation for this “ratings creep” is parental desensitization
to
this content as it becomes more prevalent in movies.
WHAT THIS STUDY ADDS: This study adds experimental
evidence
that parents become desensitized to movie violence and sex and
are more willing to allow children to view such content.
abstract
OBJECTIVES: To assess desensitization in parents’ repeated
exposure
to violence and sex in movies.
METHODS: A national US sample of 1000 parents living with
at least 1
target child in 1 of 3 age groups (6 to 17 years old) viewed a
random
sequence of 3 pairs of short scenes with either violent or sexual
content

from popular movies that were unrestricted to youth audiences
(rated
PG-13 or unrated) or restricted to those underage 17 years
without adult
supervision (rated R). Parents indicated the minimum age they
would
consider appropriate to view each film. Predictors included
order of
presentation, parent and child characteristics, and parent movie
viewing history.
RESULTS: As exposure to successive clips progressed, parents
sup-
ported younger ages of appropriate exposure, starting at age
16.9
years (95% confidence interval [CI], 16.8 to 17.0) for violence
and
age 17.2 years (95% CI, 17.0 to 17.4) for sex, and declining to
age
13.9 years (95% CI, 13.7 to 14.1) for violence and 14.0 years
(95%
CI, 13.7 to 14.3) for sex. Parents also reported increasing
willingness
to allow their target child to view the movies as exposures
progressed.
Desensitization was observed across parent and child
characteristics,
violence toward both human and non-human victims, and movie
rating.
Those who frequently watched movies were more readily
desensitized
to violence.
CONCLUSIONS: Parents become desensitized to both violence
and sex
in movies, which may contribute to the increasing acceptance of

both
types of content by both parents and the raters employed by the
film
industry. Pediatrics 2014;134:877–884
AUTHORS: Daniel Romer, PhD,a Patrick E. Jamieson, PhD,a
Brad J. Bushman, PhD,b,c Amy Bleakley, PhD, MPH,a Anli
Wang, PhD,a Daniel Langleben, MD,a,d and Kathleen Hall
Jamieson, PhDa
aAnnenberg Public Policy Center, and dDepartment of
Psychiatry,
Perelman School of Medicine, University of Pennsylvania,
Philadelphia, Pennsylvania; bSchool of Communication and
Department of Psychology, The Ohio State University,
Columbus,
Ohio; and cDepartment of Communication Science, VU
University
Amsterdam, The Netherlands
KEY WORDS
movie violence, movie sex, movie ratings, desensitization,
parents, adolescents
ABBREVIATIONS
CARA—Classification and Rating Administration
CI—confidence interval
MPAA—Motion Picture Association of America
Dr Romer conceptualized and designed the study, conducted the
statistical analyses, and drafted the initial manuscript; Dr
Jamieson conceptualized and designed the study and reviewed
and revised the manuscript; Drs Bushman and Bleakley
contributed to the design of the study and reviewed and revised
the manuscript; Drs Wang and Langleben contributed to the

analysis of the data and reviewed and revised the manuscri pt;
Dr Hall Jamieson conceptualized and designed the study and
reviewed and revised the manuscript; and all authors approved
the final manuscript as submitted.
www.pediatrics.org/cgi/doi/10.1542/peds.2014-1167
doi:10.1542/peds.2014-1167
Accepted for publication Aug 27, 2014
Address correspondence to Daniel Romer, PhD, Annenberg
Public
Policy Center, University of Pennsylvania, 202 S 36th St,
Philadelphia PA, 19104-3806. E-mail: [email protected]
PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-
4275).
Copyright © 2014 by the American Academy of Pediatrics
FINANCIAL DISCLOSURE: The authors have indicated they
have
no financial relationships relevant to this article to disclose.
FUNDING: All phases of this research were supported by
internal funds at the Annenberg Public Policy Center.
POTENTIAL CONFLICT OF INTEREST: The authors have
indicated
they have no potential conflicts of interest to disclose.
COMPANION PAPER: A companion to this article can be found
on
page 1024, and online at www.pediatrics.org/cgi/doi/10.1542/
peds.2014-2803.

PEDIATRICS Volume 134, Number 5, November 2014 877
ARTICLE
by guest on March 29,
2021www.aappublications.org/newsDownloaded from
mailto:[email protected]
http://www.pediatrics.org/cgi/doi/10.1542/peds.2014-2803
http://www.pediatrics.org/cgi/doi/10.1542/peds.2014-2803
In recognition that parents might not
want their children exposed toviolent or
sexuallyexplicitfilms,theMotionPicture
AssociationofAmerica(MPAA),themovie
industry’s trade association, assigns
ratings that warn parents of that con-
tent.1 In recent years, however, the
MPAA’s rating board, the Classification
and Rating Administration (CARA), has
allowed more violent content into the
films it rates as acceptable for adoles-
centsunder theageof 17years(itsPG-13
category for viewers ages 13 years and
older).2–5 For example, the amount of
gun violence in top-grossing PG-13 films
has more than tripled since the PG-13
rating was introduced in 1985.2 Indeed,
in 2012, popular PG-13 films contained
significantly more gun violence than R-
rated films (restricted tochildren under
age 17 years unless accompanied by an
adult), which are more likely to contain
explicit sex.2,4 Although sexual content
has been classified more consistently in

the R category,4 there is also evidence
that sex has become more prevalent in
that category.5,6
CARA acknowledges that its standards
have changed over time and attributes
this to the changing standards of par-
ents.1 An alternative and not necessarily
incompatible explanation for this “rat-
ings creep” is that parents on CARA’s
rating board and in the general pop-
ulation have become desensitized by
repeatedexposuretoincreasingamounts
of violence and sex in films.5 Desensi-
tization occurs when repeated expo-
sure to a disturbing stimulus reduces
the emotional response to that con-
tent,7 a phenomenon that has been
observed in response to violent8-13 and
sexual media.14 Consistent with this
interpretation, a Kaiser Family Foun-
dation survey of over 1000 parents
found that parents were less con-
cerned about their children’s exposure
to sex (51% vs 67%) and violence (46%
vs 62%) in 2006 than in 1998.15
Totestthedesensitizationhypothesis,we
asked an online panel of parents to view
a succession of brief movie scenes
containing either violent or sexual con-
tent.Aftereachclip,parentsratedtheage
at which they considered it appropriate
forachildtoviewthefilmfromwhichthe
scene was taken. If desensitization af-

fects the reactions of parents, exposure
to successive scenes should reduce the
age at which they would consider it ap-
propriateforachildtoviewsuchcontent.
In addition, desensitization should be
more apparent among parents who con-
sume more movies.
To test the limits of desensitization, we
also examined the effect of increasing
exposure to violence in films not only
directed toward humans but also to-
wardhuman-likecharacters.Inaddition,
we examined whether desensitization
can transfer between violence and sex.
The desensitization hypothesis predicts
that repeated exposure to either dis-
turbing stimulus (graphic violence or
overtsex)candesensitize parents tothe
other type of content.
METHODS
Participants
Participants were 1000 parents who
were members of the online research
panel sponsored by Research Now.15
The panel is representative of online
households, and the firm uses quality
controls that restrict excessive partic-
ipation in surveys. Participants for this
study were adult parents or guardians
of a child residing in the home in 1 of
the following 3 target age groups: 6 to

9 years(n=301),10to12years(n=301),
and 13 to 17 years (n = 398). By MPAA
standards, all of the movies we tested
would be inappropriate for the 2
younger groups. A parent with a child
in more than 1 age group was ran-
domly assigned to 1 group.
Movie Content
We identified brief scenes of violence in
6movies,5ofwhichwereratedaseither
PG-13 or R (see Table 1). An additional
clip was taken from a PG-13 film that
was unrated when released on DVD.
In each case, a character was shown
harming another character, most often
by using a firearm. Two of the clips from
the Terminator series involved victims
of violence that were robots in human
form that attacked humans but were
repulsed with violent action. We also
identified 2 films with brief scenes in-
volving sexual behavior, 1 of which was
rated PG-13 and the other R. The R-rated
clip showed a young male character
TABLE 1 Characteristics of Films Shown in Pairs According to
Content: 1 Set With Sex, Two Sets With Only Human Violence
(A, B), and 1 Set With
Nonhuman Victims of Violence, Resulting in 6 Different Orders
of Film Content
Film Paired Clips Content Length(s) Rating Year of Release
Percent Previously Seen

8 Mile 1 Sex 30 R 2002 45
Casino Royale 1 Sex 42 PG-13 2006 48
Collateral 2A Violence 30 R 2004 45
Taken 2 2A Violence 18 PG-13 2012 42
Die Hard 2B Violence 15 R 1998 50
Live Free or Die Harder 2B Violence 18 Unrated 2007 46
Terminator 3 Nonhuman victim 59 R 1984 49
Terminator Salvation 3 Nonhuman victim 37 PG-13 2009 46
878 ROMER et al
initiating sex with a female, whereas the
PG-13 film showed a female character
initiating less explicit sex with a male.
All of the films were among the top-25
grossing movies in the year of their
release (see http://www.annenbergpu-
blicpolicycenter.org/test-videos-for-parent-
study/), and between 42% and 50% of the
parents in the sample reported previous
exposure to the films.
Experimental Procedure
Parents were shown 3 pairs of clips in
a random order. The order within each
pair was also random, containing a clip
either from a film rated PG-13 or
R/Unrated.Halfoftheparentssaw1pair
of the 4 human violence clips, and the
other half saw the otherpair. All parents
saw the 2 sex and the 2 violent clips with

non-human victims. This design enabled
us to determine order effects for both
sex and violence, with more power to
study effects of violence than of sex.
Wetoldparents,“Wewillshowyoushort
scenes from movies and ask you to tell
us how you feel about children and
adolescents viewing those films.” After
viewing each clip, they were asked
a series of questions about the film
from which the clip was taken. Here we
focus on the question: At what age do
you think it would be ok for a child to
see the movie from which this clip was
taken? Choices ranged from age 6 to 19
years. A second question regarding the
likelihood that parents would allow
their target child to view the film in the
future was also asked. Parents also
reported whether they had previously
seen the movie from which the clip was
taken. After rating all the clips, parents
reported how many movies they watched
on television or a computer in the past
week, including broadcast, cable, and
video streaming. They also completed
a 3-item parental monitoring scale
(Cronbach a = 0.85) that queried their
practices regarding the target child’s
whereabouts and behavior outside the
home.16 Thesurveytookabout20minutes
and was deemed exempt from review
by the University of Pennsylvania In-
stitutional Review Board.

Analysis
Mixed-effects regression using Stata 13
(Stata Corp, College Station, TX) was
used to analyze trends in age appro-
priateness across the 6 movie-clip
presentations, with parent and child
characteristics as predictors. This anal-
ysis allowed us to test the effects of
consecutive presentations of either vi-
olent or sexual content as well as dif-
ferences in parent and target child
characteristics.
RESULTS
Demographic characteristics of the
sample (Table 2) indicate that most of
the participants were married biologi-
cal parents and 56% were mothers.
Violence
Table 3 contains the mean ages reported
by parents after viewing violent movies
by presentation order, and various child
and parent characteristics. Preliminary
analyses revealed that the violence clips
were rated equivalently for age appro-
priateness, including the ones involving
non-human characters. Hence, we col-
lapsed the ratings for the 6 violent films
at each presentation order.

Preliminary analyses on the sample’s
mean age ratings by presentation order
indicated that a quadratic function
provided extremely good fits (R2 values
$0.97), the results of which we show in
the Figure 1. Analyses shown in Table 4
revealed that presentation order (linear
TABLE 2 Percentages of Sample According to Parental and
Child Characteristics by Age of the
Target Child
Parent Characteristic Age of Target Child, y Total
6 to 9 10 to 12 13 to 17
Marital status
Married 70.8 75.1 71.6 72.4
Single 12.6 8.3 8.8 9.8
Living with partner 9.3 4.7 7.8 7.3
Divorced/other 7.3 11.9 11.9 10.5
Relation to child
Mother 61.1 53.8 54.5 56.3
Father 28.9 34.6 33.2 32.2
Grandparent 4.7 4.3 3.8 4.2
Other 5.3 7.3 8.5 7.3
Age, y
18 to 34 43.5 20.3 12.3 24.1
35 to 44 37.2 46.2 31.7 37.7
45+ 19.3 33.6 56.0 38.2
Child gender
Male 49.1 51.9 54.0 51.9
Female 50.9 48.1 46.0 48.1

Education
,High school 1.6 1.6 2.5 2.0
High school 15.6 16.3 18.8 17.1
Some college 35.9 38.2 43.0 39.4
Bachelor degree 33.6 29.9 20.6 27.3
Post-graduate degree 13.3 14.0 15.1 14.2
Race/ethnicitya
White 83.7 87.0 87.4 86.2
Black 9.0 7.6 6.8 7.7
Hispanic 11.0 13.0 10.3 11.3
Asian 4.0 2.7 5.3 4.1
Other 3.3 4.3 1.3 2.8
a Racial/ethnic categories can overlap.
ARTICLE
http://www.annenbergpublicpolicycenter.org/test-videos-for-
parent-study/
parent-study/
parent-study/
and quadratic terms), parental monitor-
ing, and previous exposure to the films
wererelatedtoageappropriateness.Age
judgments declined as a function of

presentation order with the quadratic
term reducing the effect of successive
exposures (Fig 1). Parents who had pre-
viously seen more of the clips were less
restrictive overall. They were also more
likelytohavewatchedmoremoviesinthe
pastweek(r=0.28).Theconstanttermin
the model when adjusted for other pre-
dictors represents the age that parents
assigned to the first clip (16.9 years; 95%
CI, 16.8 to 17.0). With each succeeding
clip, however, parents reduced their age
judgment by a little over 1 year, until the
quadratic term became large enough to
reduce the order effect.
Theinteractionbetweenorderandparent
age indicated that older parents were
less subject to desensitization (Fig 1A). In
addition, the more movies parents had
watched in the past week, the less re-
strictivetheywereastheysawadditional
clips (Fig 1B). In particular, those viewing
11 or more movies rated the first clip at
age 16.6 years (95% CI, 15.6 to 17.6) but
the last 1 at age 12.0 years (95% CI, 11.0
to 13.0). For parents who viewed fewer
movies, the effect of order was weaker.
Parents who viewed no movies rated the
last clip at age 14.8 years (95% CI, 14.2 to
15.4), whereas those who viewed be-
tween 3 and 4 movies rated the final clip
at age 13.9 years (95% CI, 13.5 to 14.3).
Target child’s gender did not influence
parents’ age judgments, nor did parent

education and income. Analysis of the
random effects (not shown) indicated
that parents who assigned older ages
to the first clip displayed greater de-
sensitization as they watched sub-
sequent clips.
Sex
Table 3 contains the mean ages repor-
ted by parents for the average of the
2 movies containing sex. Like violence,
order of presentation (linear and qua-
dratic terms) was related to judgments
ofageappropriateness(Table4).Parents
rated the first sexual clip at age 17.2
years (95% CI, 17.0 to 17.4) and the last
at age 14.0 years (95% CI, 13.7 to 14.3).
Previous exposure to the films was also
inversely related to age judgments. Un-
like violence, the target child’s gender
and age mattered for sex. Parents were
more restrictive if their target child was
older (Fig 1C) or female (Fig 1D). Parents
who monitored their children more
closely were also more restrictive.
However, age of child or gender did not
interact with order of presentation.
Like violence, parent age interacted with
presentationordersuchthatolderparents
were less subject to desensitization.
Unlike violence, movie-watching experi-
ence did not interact with order of pre-
sentation. Nevertheless, it is clear that
desensitizationtosexoccurrednotonly

when the 2 sex clips were shown in the
first 2 positions, but also when they
followed clips with violent content.
Desensitizationtoviolenceoccurredeven
when sexual clips preceded violent ones.
Agejudgmentsinresponsetoviolencein
thethirdandfourthpositionswererated
equivalently regardless of whether the
first 2 clips contained sex or violence
(data not shown).
Unliketherandomcomponentresultsfor
violence, initial age judgments were
positivelyrelatedtoorder,indicatingthat
parentswhostartedoutmorerestrictive
exhibited less desensitization for sexual
content as they watched more clips.
TABLE 3 Mean Ages Assigned to Movies With Either Violence
or Sex by Age of Child, Order of
Presentation, Parental Monitoring, Number of Movies Seen
Before, Child Gender,
Number of Movies Watched Past Week, and Parent Age
Predictor Age of Child, y
Total6 to 9 10 to 12 13 to 17
Violence Sex Violence Sex Violence Sex Violence Sex
Order
1 16.7 17.0 16.8 17.2 17.1 17.4 16.9 17.2
2 15.5 15.8 16.0 15.9 16.0 16.5 15.8 16.1
3 14.6 14.3 14.9 15.0 15.3 15.3 15.0 14.9
4 14.5 14.2 14.5 14.7 14.9 15.1 14.7 14.6

5 13.7 13.3 13.6 13.8 14.3 14.7 13.9 14.0
6 13.4 13.5 13.7 13.6 14.4 14.5 13.9 14.0
Total by age 14.7 14.8 14.9 15.1 15.3 15.6 15.0 15.2
Parental monitoringa
Low 14.7 14.3 14.4 14.6 15.3 15.7 14.9 15.1
Medium 14.7 14.9 15.1 15.3 15.3 15.4 15.1 15.3
High 14.8 14.9 15.0 15.1 15.5 15.7 15.0 15.2
Movies seen before
None 16.1 16.0 16.8 16.0 16.6 16.6 16.5 16.2
1 to 2 15.1 15.2 15.0 15.3 15.5 15.9 15.3 15.5
3 to 5 14.4 14.4 14.7 14.9 15.0 15.2 14.7 14.9
6 13.7 13.8 12.9 13.9 14.5 15.0 13.7 14.2
Child gender
Male 14.6 14.4 14.9 14.9 15.2 15.4 15.0 15.0
Female 14.8 15.1 14.9 15.2 15.5 15.8 15.1 15.4
Recent movies watched
None 15.4 15.5 15.6 15.9 15.8 15.9 15.7 15.8
1 to 2 14.7 14.9 15.1 15.6 15.4 15.8 15.1 15.4
3 to 4 15.1 14.9 14.7 14.7 15.2 15.5 15.0 15.1
5 to 10 14.2 14.0 14.8 14.8 15.1 15.1 14.7 14.7
11+ 14.3 14.8 13.4 13.2 15.1 15.3 14.1 14.2
Parent age, y
18 to 34 14.2 14.5 14.1 14.2 15.1 15.4 14.4 14.6
35 to 44 15.1 14.8 15.1 15.2 14.9 14.9 15.0 15.0
45+ 15.1 15.3 14.9 15.4 15.6 16.0 15.4 15.7
a Trichotomized into approximately equal proportions of the
sample.
880 ROMER et al

Target Child Viewing
Likelihoodofallowingthetargetchildto
view the films was inversely correlated
with age judgments (r’s = 20.45 and
20.47 for violence and sex, respectively).
In addition, linear and quadratic order
effects were evident with greater le-
niency as exposures progressed (see
Tables 5 and 6). A major difference was
that the age of the child played a
stronger role than the age of the par-
ent for both violence and sex, with
parents being less restrictive for older
children.
DISCUSSION
When parents viewed film clips con-
tainingeitherviolenceorsexthatmight
FIGURE 1
Best fitting curves (R2 $ 0.97) for age appropriateness by order
of presentation for violence according to A, parents of different
ages and B, parents with
different amounts of movie watching, and for sex according to
C, children of different ages and D, gender.
TABLE 4 Fixed Effects in Regression Analysis of Age

Appropriateness for Violent and Sexual Movies
Predictor Violence Sex
Coefficient 95% CI Probability Coefficient 95% CI Probability
Linear order 21.29 21.48 to 21.11 ,0.001 21.80 22.10 to 21.50
,0.001
Quadratic order 0.09 0.07 to 0.11 ,0.001 0.15 0.12 to 0.18
,0.001
Child age 0.16 20.00 to 0.32 0.055 0.35 0.18 to 0.53 ,0.001
Child gendera 0.00 20.24 to 0.25 0.985 0.37 0.10 to 0.63 0.006
Parental monitoring 0.16 0.00 to 0.31 0.051 0.21 0.04 to 0.38
0.015
Movies seen before 20.57 20.72 to 20.41 ,0.001 20.55 20.72 to
20.37 ,0.001
Parent age 0.15 20.06 to 0.35 0.162 0.00 20.31 to 0.32 0.976
Recent movies 20.03 20.17 to 0.12 0.737 20.12 20.34 to 0.10
0.294
Watched 0.06 0.01 to 0.11 0.024 0.08 0.00 to 0.17 0.041
Order 3 parent age
Order 3 movies watched 20.04 20.08 to 20.01 0.020 20.02 20.08
to 0.05 0.624
Constant 18.25 17.5 to 19.0 ,0.001 18.69 17.7 to 19.7 ,0.001
a Coded as male = 1 and female = 2.
ARTICLE
be inappropriate for children, desensi-

tization was clearly evident even within
the short duration of this study. As suc-
cessive clips were seen, movies regarded
as only appropriate for older adoles-
cents soon became more acceptable for
younger ages. The order effect occurred
regardless of whether the violence was
directed at a human or a human-like
character and regardless of the film’s
rating. Desensitization also occurred for
sexual content, not only when the 2
sexual clips were shown first, but also
later in the viewing sequence, indicating
transferfromviolence tosex. Inaddition,
there was no difference in response to
violent clips whether they were pre-
ceded by sexual content or not. These
patterns suggest that desensitization is
possible not only within but also across
violent and sexual content.
The results support the operation of
a very basic reduction in emotional
response produced by repeated expo-
sures to sex and violence. Desensiti-
zation has also been found to transfer
from fictional media to real-life vio-
lence.8,11 Although the arousal pro-
duced by both sexual and violent
content probably played a role, it is
likely that the disturbing nature of the
content contributed to the transfer
between violence and sex. One study
that compared repeated exposure to

arousing comedic versus violent film
clips found a much weaker and differ-
ent pattern of desensitization in re-
sponse to comedic clips, suggesting
that arousal is not sufficient to produce
desensitization.13 Other research reached
the same conclusion.18
TABLE 5 Mean Ratings of Likelihood of Allowing Target Child
to View Movies With Either Violence
or Sex by Age of Child, Order of Presentation, Parental
Monitoring, Number of Movies
Seen Before, Child Gender, Number of Movies Watched Past
Week, and Parent Age
Predictor Age of Child, y
6 to 9 10 to 12 13 to 17 Total
Violence Sex Violence Sex Violence Sex Violence Sex
Order
1 1.96 1.65 2.08 1.81 3.20 2.73 2.49 2.10
2 2.37 2.16 2.77 2.56 4.18 3.87 3.22 2.92
3 2.72 2.63 3.20 3.13 4.51 4.23 3.57 3.41
4 2.92 2.68 3.46 3.39 4.82 4.50 3.84 3.61
5 3.18 3.16 3.98 3.90 5.13 5.16 4.17 4.22
6 3.11 3.02 3.85 3.97 4.93 5.12 4.02 4.18
Total by age 2.72 2.52 3.23 3.12 4.45 4.30 3.56 3.40
Parental monitoring
Low 3.25 2.92 3.60 3.41 4.46 4.38 3.93 3.77
Medium 2.78 2.39 3.11 2.94 4.49 4.33 3.69 3.48
High 2.48 2.41 3.10 3.09 4.35 4.13 3.15 3.06
Movies seen before

None 1.89 1.76 2.00 2.38 3.40 3.22 2.57 2.57
1 to 2 2.20 2.11 3.07 2.77 3.99 3.98 3.18 3.06
3 to 5 2.96 2.74 3.46 3.42 4.92 4.68 3.88 3.71
6 4.25 3.67 4.17 3.70 5.42 5.31 4.63 4.24
Child gender
Male 2.94 2.67 3.27 3.31 4.63 4.56 3.74 3.64
Female 2.54 2.39 3.21 2.94 4.21 3.98 3.38 3.16
Recent movies watched
None 1.97 2.26 2.81 2.32 3.85 3.90 3.11 3.08
1 to 2 2.72 2.35 3.01 2.75 4.31 4.15 3.39 3.13
3 to 4 2.66 2.48 3.42 3.38 4.49 4.16 3.66 3.46
5 to 10 3.00 2.89 3.23 3.45 4.99 5.04 3.81 3.87
11+ 3.58 2.77 4.21 4.03 4.98 4.15 4.20 3.67
Parent age, y
18 to 34 3.00 2.62 3.61 3.71 4.42 4.38 3.44 3.26
35 to 44 2.56 2.51 3.13 3.08 4.72 4.62 3.49 3.42
45+ 2.42 2.29 3.12 2.82 4.30 4.11 3.70 3.49
Parents were asked, “How likely is it that you would allow
(target child) to see this movie in the future?” Response options
ranged from extremely unlikely (1) to neither unlikely or likely
(4) to extremely likely (7).
TABLE 6 Fixed Effects in Regression Analysis of Likelihood of
Allowing Child to View Either Violent or Sexual Movies in the
Future
Predictor Violence Sex
Coefficient 95% CI Probability Coefficient 95% CI Probability
Linear order 0.80 0.68 to 0.93 ,0.001 1.05 0.88 to 1.24 ,0.001
Quadratic order 20.08 20.09 to 20.06 ,0.001 20.10 20.12 to

20.08 ,0.001
Child age 0.73 0.56 to 0.90 ,0.001 0.49 0.26 to 0.73 ,0.001
Child gendera 20.17 20.35 to 0.01 0.069 20.32 20.52 to 20.11
0.002
Parental monitoring 20.19 20.31 to 20.08 0.001 20.21 20.33 to
20.08 0.002
Movies seen before 0.61 0.48 to 0.73 ,0.001 0.54 0.41 to 0.68
,0.001
Parent age
Recent movies 20.10 20.22 to, 0.30 0.133 20.11 20.25 to 0.03
0.141
Watched 0.04 20.08 to 0.17 0.488 0.11 20.07 to 0.28 0.238
Order 3 child age 0.05 0.01 to 0.09 0.007 0.11 0.05 to 0.17
,0.001
Order 3 movies watched 20.04 0.01 to 0.06 0.021 0.01 20.04 to
0.06 0.724
Constant 1.49 0.92 to 2.07 ,0.001 1.20 0.53 to 1.88 ,0.001
a Coded as male = 1 and female = 2.
882 ROMER et al
We also found that the more movies
parents had watched in the past week,
the less restrictive they became as
exposure to violent clips progressed. In
addition, previous exposure to the
movies shown, which was related to
frequency of movie viewing, was a pre-
dictor of greater leniency for both vio-
lence and sex. These findings support
the view that not only does desensi-

tization occur in the study’s short time
frame, but that it can also accrue over
time, so that parents become less dis-
turbed by violent and sexual content in
movies the more they view it.
The finding that desensitization to vio-
lence can also extend to sex suggests
thattherecentriseinviolentfilmsinthe
PG-13 category2 may also have influ-
enced the way parents evaluate films
with sex. This conclusion is consistent
with declining reported parental con-
cern about both violence and sex in
films.15 In short, recent increases in
film violence may have made parents
more tolerant of both violence and sex
in movies.
The wide variation in response to these
films underscores the challenges that
any global rating system faces in pro-
vidingparentsusefulinformationabout
film content.19 Older parents were less
sensitive to presentation order for
both violence and sex, perhaps be-
cause they have stronger emotional
reactions to the explicit content in
these films. In addition, parents of all
ages were more likely to consider their
own child’s age and gender in response
to sex, recommending greater restric-
tion when they had older or female
children. Nevertheless, when it came to
deciding what their own children
should watch in the future, parents of

older children appeared to recognize
that their children would soon be ma-
ture enough to view the films.
The finding that violence directed
against human-like characters elicited
the same response asviolence directed
toward humans suggests that it is the
violent act itself that parents found
disturbing. Indeed, responses to the
films did not differ despite different
ratings for either sex or violence. These
broad results suggest that desensi-
tization may have far-reaching effects.
As parents become inured to violence
and sex in films, they will be less likely
to shield their children from such
content. Children may then also be-
come desensitized to violence, which
could reduce their empathy for the
suffering of others11,12,20 and encourage
aggressive responses to conflict.21,22
Exposure to sex may also lead to early
sexual initiation23 as well as increased
risk for teen pregnancy.24
Althoughour findingsmay notextendto
CARA’s behavior, CARA board members
are also parents selected to represent
families across the country.1 Given that
they preview and rate hundreds of
films a year,25 they could also be
desensitized to disturbing content in
films and thus more likely to be lenient

when it comes to evaluating the appro-
priateness of such content for chil-
dren.5 This could help to explain the
ratings creep that has occurred in
films containing violence over the past
20 years. As the industry has pushed
for the PG-13 rating, especially for vio-
lent films that draw large audiences in
the United States and abroad, parents
as well as CARA may have become more
accepting of violent content. As a re-
sult, our entire culture may be un-
dergoing desensitization to violent
movies with consequences that remain
unknown. One possible outcome of this
desensitization is the greater accep-
tance of the use of guns, which are
heavily featured in violent PG-13 mov-
ies.2 Indeed, use of guns in violent acts
has increased in US youth over the past
decade.26 Considering that movies may
enhance associations between guns
and violence,2,27 the effects on chil-
dren’s exposure to violent use of weap-
ons in movies remains surprisingly
unstudied.28 The present findings sug-
gest that this should be a high priority
for future research.
An alternative explanation for our
findings is that viewing any series of
movie clips would result in more tol-
erance of the portrayed behavior.
However, this seems unlikely, because
neutral clips would presumably be ac-

ceptableforchildrenatamuchyounger
age than 17 years, the approximate age
suggested for all of the studied clips
when they were first viewed. Thus, even
if there were some effects on age
judgmentsowingtorepeatedviewing,it
is unlikely that they would produce the
steepagedeclineobservedinthisstudy.
Despite the clear findings, we also
recognize some limitations that invite
further research. Because of concerns
regarding respondent burden, we did
not measure participants’ emotional
reactions to each film clip. Thus, we do
not have direct evidence that emotional
reactions declined over successive
viewings. However, earlier research
has established that emotional reac-
tions are less intense with repeated
viewings of violence.8,18,29 With the ex-
ception of pornography,14 less is known
about repeated viewings of sex and its
effects on subsequent viewing of vio-
lence. In addition, we did not expose
parents to as many clips containing sex
as violence. Thus, we have less confi-
dence in the generality of the findings
regarding sexual content.
Our sample was drawn from a national
panel that is not necessarily repre-
sentative of parents in the United
States. However, characteristics such
as education and income were not
related to desensitization. Furthermore,

ourstudywasnotdesignedtodetermine
national rates of tolerance to movie vi-
olence and sex, but rather to determine
whether parents across a wide range
of characteristics would exhibit a sys-
tematic pattern of desensitization to
ARTICLE
violent and sexual content over repeated
viewings.
CONCLUSIONS
Thefindingssuggestthatdesensitization
to films with violent or sexual content
poses major challenges for the assign-
ment of film ratings. CARA raters who
presumably see many films in a week
may be subject to desensitization that
affects their ratings. At the same time,
parents may also be more accepting of
lenient ratings as they are repeatedly
exposed to films with violent or sexual
content. and may be more willing to al-
low their children to view it.
ACKNOWLEDGMENTS

WethanktheteamatResearchNowand
Jeremy Quattlebaum of APPC for pro-
gramming the online study materials.
REFERENCES
1. Classification and Rating Administration.
History of the film rating system. Available
at: filmratings.com. Accessed August 1, 2014
2. Bushman BJ, Jamieson PE, Weitz I, Romer
D. Gun violence trends in movies. Pediat-
rics. 2013;132(6):1014–1018 doi:10.1542/
peds.2013-1600
3. Gentile DA. The rating systems for media
products. In: Calvert S, Wilson BJ, eds. The
Handbook of Children, Media, and De-
velopment. Malden, MA: Blackwell; 2008:
527–551
4. Nalkur PG, Jamieson PE, Romer D. The ef-
fectiveness of the Motion Picture Associa-
tion of America’s rating system in
screening explicit violence and sex in top-
grossing movies from 1950 to 2006. J
Adolesc Health. 47(5):440–447
5. Potts R, Belden A. Parental guidance:
a content analysis of MPAA motion picture
rating justifications 1993–2005. Curr Psy-
chol. 2009;28(4):266–283
6. Leone R, Osborn L. Hollywood’s triumph and
parents’ loss: an examination of the PG-13
rating. Pop Commun. 2004;2(2):85–101

7. Wolpe J. The Practice of Behavior Therapy,
3rd ed. New York: Pergamon Press; 1982
8. Carnagey NL, Anderson CA, Bushman BJ.
The effect of video game violence on
physiological desensitization to real-life vi-
olence. J Exp Soc Psychol. 2007;43:489–496
doi:10.1016/j.jesp.2006.05.003
9. Engelhardt CR, Bartholow BD, Kerr GT,
Bushman BJ. This is your brain on violent
video games: Neural desensitization to vi-
olence predicts increased aggression fol-
lowing violent video game exposure. J Exp
Soc Psych. 2011;47:1033–1036
10. Anderson CA, Shibuya A, Ihori N, et al. Vio-
lent video game effects on aggression,
empathy, and prosocial behavior in eastern
and western countries: a meta-analytic
review. Psychol Bull. 2010;136(2):151–173
doi:10.1037/a0018251
11. Thomas MH, Horton RW, Lippincott EC,
Drabman RS. Desensitization to portrayals
of real-life aggression as a function of ex-
posure to television violence. J Pers Soc
Psychol. 1977;35(6):450–458
12. Cline VB, Croft RG, Courrier S. Desensitization
of children to television violence. J Pers Soc
Psychol. 1973;27(3):360–365
13. Fanti KA, Vanman E, Henrich CC, Avraamides

MN. Desensitization to media violence over
a short period of time. Aggress Behav.
2009;35(2):179–187 doi:10.1002/ab.20295
14. Zillmann D, Bryant J. Pornography, sexual
callousness, and the trivialization of rape. J
Commun. 1982;32(4):10–21
15. Rideout V. Parents, Children & Media: A
Kaiser Family Foundation Survey. Menlo
Park, CA: The Henry J. Kaiser Foundation;
2007:4. Available at: www.kff.org
16. Research Now. Research Now Survey Sampling
and Quality Control. Available at: www.
researchnow.com/en-US/Services/Sample.aspx
17. Sieverding JA, Adler N, Witt S, Ellen J. The
influence of parental monitoring on ado-
lescent sexual initiation. Arch Pediatr Ado-
lesc Med. 2005;159(8):724–729
18. Linz D, Donnerstein E, Adams SM. Physio-
logical desensitization and judgments
about female victims of violence. Hum
Commun Res. 1989;15(4):509–522
19. Gentile DA, Maier JA, Hasson MR, Lopez de
Bonetti B. Parents’ evaluation of media
ratings a decade after the television rat-
ings were introduced. Pediatrics. 2011;128
(1):36–44 doi:10.1542/peds.2010-3026
20. Funk JB, Baldacci HB, Pasold T, Baumgardner
J. Violence exposure in real-life, video games,
television, movies, and the internet: is there

desensitization? J Adolesc. 2004;27(1):23–39
doi:10.1016/j.adolescence.2003.10.005
21. Anderson CA, Bushman BJ. Psychology. The
effects of media violence on society. Sci-
ence. 2002;295(5564):2377–2379
22. Browne KD, Hamilton-Giachritsis C. The in-
fluence of violent media on children and
adolescents:a public-health approach. Lan-
cet. 2005;365(9460):702–710 doi:10.1016/
S0140-6736(05)17952-5
23. Brown JD, L’Engle KL, Pardun CJ, Guo G,
Kenneavy K, Jackson C. Sexy media matter:
exposure to sexual content in music, mov-
ies, television, and magazines predicts
black and white adolescents’ sexual be-
havior. Pediatrics. 2006;117(4):1018–1027
doi:10.1542/peds.2005-1406
24. Chandra A, Martino SC, Collins RL, et al. Does
watching sex on television predict teen
pregnancy? Findings from a national longi-
tudinal survey of youth. Pediatrics. 2008;122
(5):1047–1054
25. Waxman S. Rated S, for secret: a former
film rater breaks the code of silence
and tells how 13 people judge 760
movies a year. Washington Post. April 8,
2001:1
26. Children’s Defense Fund. Protect Children
Not Guns. Washington, DC; 2013. Available
at: www.childrensdefense.org/child-research-

data-publications/data/protect-children-
not-guns-2013.pdf2013. Accessed August
1, 2014
27. Bushman BJ. The weapons effect. JAMA
Pediatr. 2013;167(12):1094–1095 doi:10.1001/
jamapediatrics.2013.3824
28. National Research Council. Priorities for
Research to Reduce the Threat of Firearm-
Related Violence. Washington, DC: The Na-
tional Academies Press; 2013
29. Guo X, Zheng L, Wang H, et al. Exposure to
violence reduces empathetic responses to
other’s pain. Brain Cogn. 2013;82(2):187–
191 doi:10.1016/j.bandc.2013.04.005
884 ROMER et al
http://www.kff.org
http://www.researchnow.com/en-US/Services/Sample.aspx
http://www.researchnow.com/en-US/Services/Sample.aspx
http://www.childrensdefense.org/child-research-data-
publications/data/protect-children-not-guns-2013.pdf2013
DOI: 10.1542/peds.2014-1167 originally published online
October 20, 2014;
2014;134;877Pediatrics

Daniel Langleben and Kathleen Hall Jamieson
Daniel Romer, Patrick E. Jamieson, Brad J. Bushman, Amy
Bleakley, Anli Wang,
Services
Updated Information &
http://pediatrics.aappublications.org/content/134/5/877
including high resolution figures, can be found at:
References
http://pediatrics.aappublications.org/content/134/5/877#BIBL
This article cites 20 articles, 5 of which you can access for free
at:
Subspecialty Collections
http://www.aappublications.org/cgi/collection/screen_time_sub
Screen Time
http://www.aappublications.org/cgi/collection/media_sub
Media
http://www.aappublications.org/cgi/collection/firearms_sub
Firearms
son_prevention_sub
http://www.aappublications.org/cgi/collection/injury_violence_-
_poi
Injury, Violence & Poison Prevention
following collection(s):
This article, along with others on similar topics, appears in the
Permissions & Licensing
http://www.aappublications.org/site/misc/Permissions.xhtml

in its entirety can be found online at:
Information about reproducing this article in parts (figures,
tables) or
Reprints
http://www.aappublications.org/site/misc/reprints.xhtml
Information about ordering reprints can be found online:
http://http://pediatrics.aappublications.org/content/134/5/877
http://pediatrics.aappublications.org/content/134/5/877#BIBL
_poison_prevention_sub
_poison_prevention_sub
http://www.aappublications.org/cgi/collection/firearms_sub
http://www.aappublications.org/cgi/collection/media_sub
http://www.aappublications.org/cgi/collection/screen_time_sub
http://www.aappublications.org/site/misc/Permissions.xhtml
http://www.aappublications.org/site/misc/reprints.xhtml
DOI: 10.1542/peds.2014-1167 originally published online
October 20, 2014;
2014;134;877Pediatrics
Daniel Langleben and Kathleen Hall Jamieson
Daniel Romer, Patrick E. Jamieson, Brad J. Bushman, Amy
Bleakley, Anli Wang,
located on the World Wide Web at:

The online version of this article, along with updated
information and services, is
by the American Academy of Pediatrics. All rights reserved.
Print ISSN: 1073-0397.
the American Academy of Pediatrics, 345 Park Avenue, Itasca,
Illinois, 60143. Copyright © 2014
has been published continuously since 1948. Pediatrics is
owned, published, and trademarked by
Pediatrics is the official journal of the American Academy of
Pediatrics. A monthly publication, it
lable at ScienceDirect
Computers in Human Behavior 69 (2017) 268e274
Contents lists avai
Computers in Human Behavior
journal homepage: www.elsevier.com/locate/comphumbeh
Review
Cyberbullying in higher education: A literature review
Lynette K. Watts a, *, Jessyca Wagner b, Benito Velasquez c,
Phyllis I. Behrens d
a Radiologic Sciences Department, Midwestern State
University, 3410 Taft Blvd., Bridwell Hall Office 201E, Wichita
Falls, TX 76308, USA

b Radiologic Sciences Department, Midwestern State
University, 3410 Taft Blvd., Bridwell Hall, Office 216, Wichita
Falls, TX 76308, USA
c Athletic Training Department, School of Allied Health,
Lincoln Memorial University, 6965 Cumberland Gap Parkway,
Harrogate, TN 37752, USA
d University of Missouri Extension, 3950 Newman Road,
Plaster Hall 107A, Joplin, MO 64801-1512, USA
a r t i c l e i n f o
Article history:
Received 29 January 2016
Received in revised form
10 November 2016
Accepted 15 December 2016
Available online 18 December 2016
Keywords:
Higher education cyberbullying
Cyberbullying prevalence
Cyberbullying awareness
Social media cyberbullying
* Corresponding author.
E-mail address: [email protected] (L.K. Wa
http://dx.doi.org/10.1016/j.chb.2016.12.038
0747-5632/© 2016 Elsevier Ltd. All rights reserved.
a b s t r a c t
With the more pervasive use of technology, the prevalence of
cyberbullying has increased. Universities and
colleges have implemented more ways to reach students over
the Internet and by other electronic means;
consequently, itisimportant to investigate the prevalence of
cyberbullyinginhighereducation.Thisliterature
review was created to raise awareness of this continuing trend

of cyberbullying among college students.
Cyberbullying is defined as intent to consistently harass another
individual to cause harm via any electronic
method, including social media, and includes defamation, public
disclosure of private facts, and intentional
emotional distress. An exhaustive search of current literature
was conducted using a variety of databases
including Academic Search Complete, Education Source, E-
Journals, ERIC, PsycArticles, Psychology and
Behavioral Sciences Collection, and PsycINFO. Google, Google
Scholar, and reference lists from key studies
were also used to gather relevant studies for inclusion in this
review. Articles were used to define cyberbul-
lying, build a historical base of cyberbullying among
adolescents, examine factors involved in cyberbullying,
describeeffects ofcyberbullying,and examine thistrend among
college students. Strategies suchasreporting,
monitoring online interactions, and legal actions to address
cyberbullying in all areas of education were
included. Future research should investigate if educational
programs lead to decreased cyberbullying,
increased rates of reporting, and how cyberbullying may change
social media etiquette.
© 2016 Elsevier Ltd. All rights reserved.
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
2. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
3. Bullying and cyberbullying defined . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269

3.1. Types of cyberbullying . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
4. Historical perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
5. Factors contributing to cyberbullying . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
5.1. Anonymity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
5.2. Psychological needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
5.3. Social dominance theory . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
5.4. Poor interpersonal child/adult relationships . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . 271
5.5. Other social cues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
6. Effects of cyberbullying behavior . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
7. Undergraduates and cyberbullying . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
8. Strategies to resolve cyberbullying . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272

9. Conclusions and recommendations . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
tts).
mailto:[email protected]
http://crossmark.crossref.org/dialog/?doi=10.1016/j.chb.2016.12
.038&domain=pdf
www.sciencedirect.com/science/journal/07475632
www.elsevier.com/locate/comphumbeh
L.K. Watts et al. / Computers in Human Behavior 69 (2017)
268e274 269
1. Introduction
The Internet can have an impact on users socially and psycho-
logically (Bessi�ere, Pressman, Kiesler, & Kraut, 2010; Kraut
et al.,
1998, 2002). More individuals are using the Internet for social
communication with e-mail, instant messaging, chat rooms, and
gaming along with general information retrieval, health
resources,
and entertainment (Bessi�ere et al., 2010; Kraut et al., 2002).
While
the use of the Internet has its benefits, there has been much

speculation as to whether the Internet may also be harmful to
one's
social and psychological well-being. The argument has been
pre-
sented the Internet could cause users to become socially inept
and
isolated, whereas others believe the increased interaction could
improve social relationships (Kraut et al., 1998). Unfortunately,
the
use of the Internet to perpetrate cyberbullying may present a
stronger argument for the detrimental effects of the technology.
Bullying is a major problem in today's society and occurs at
many different ages and in many different forms. With the
increase
in the use of technology, a disturbing trend worldwide is cyber -
bullying, where individuals can harass others online through
emails, text messaging, and social media websites 24 h a day,
seven
days a week. The anonymity cyberbullying provides gives
bullies a
sense of power and control that otherwise might not be present
if
they were face-to-face with their victims. Also, the prevalence
of
this topic in current news and media can be empowering to an
individual because of gained publicity, even if he or she is the
only
one aware he or she is the bully in question.
The purpose of this literature review was to enhance awareness
of the prevalence of cyberbullying in higher education among
higher education students, administrators, and faculty. While
several studies regarding cyberbullying in higher education
exist,
further investigation is needed to examine characteristics of

Historically, in the wake of interpersonal violence, the re

Historically, in the wake of interpersonal violence, the re

Recommended

Recommended

More Related Content

Similar to Historically, in the wake of interpersonal violence, the re

Similar to Historically, in the wake of interpersonal violence, the re (20)

More from SusanaFurman449

More from SusanaFurman449 (20)

Recently uploaded

Recently uploaded (20)

Historically, in the wake of interpersonal violence, the re