Content uploaded by Nicholas Epley
Author content
All content in this area was uploaded by Nicholas Epley on Jul 28, 2014
Content may be subject to copyright.
Research Article
Creating Social Connection
Through Inferential
Reproduction
Loneliness and Perceived Agency in Gadgets, Gods, and
Greyhounds
Nicholas Epley,
1
Scott Akalis,
2
Adam Waytz,
1
and John T. Cacioppo
1
1
University of Chicago and
2
Harvard University
ABSTRACT—People are motivated to maintain social con-
nection with others, and those who lack social connection
with other humans may try to compensate by creating a
sense of human connection with nonhuman agents. This
may occur in at least two ways—by anthropomorphizing
nonhuman agents such as nonhuman animals and gadgets
to make them appear more humanlike and by increasing
belief in commonly anthropomorphized religious agents
(such as God). Three studies support these hypotheses both
among individuals who are chronically lonely (Study 1)
and among those who are induced to feel lonely (Studies 2
and 3). Additional findings suggest that such results are
not simply produced by any negative affective state (Study
3). These results have important implications not only for
understanding when people are likely to treat nonhuman
agents as humanlike (anthropomorphism), but also for
understanding when people treat human agents as non-
human (dehumanization).
Few people are strangers to the dream of owning a secluded
island where they can live in supreme isolation, away from the
crush of humanity that surrounds modern urbanites. The expe-
rience of those who approximate this dream, however, suggests
that it may be more of a nightmare. Even moderate levels of
isolation from other people can increase the incidence of clinical
depression and suicide ideation (Heinrich & Gullone, 2006),
elevate blood pressure levels (Hawkley, Masi, Berry, & Cacioppo,
2006), increase levels of stress hormones (Adam, Hawkley,
Kudielka, & Cacioppo, 2006), and compromise one’s immune
system (Cacioppo, Hawkley, & Berntson, 2003). Taken togeth-
er, these health risks make chronic isolation at least as large a
risk factor for morbidity and mortality as cigarette smoking
(House, Landis, & Umberson, 1988). Satisfying social relations
with others is the one demonstrable factor that systematically
differentiates very happy people from unhappy people (Diener
& Seligman, 2002), and most people who find themselves iso-
lated from others fairly quickly begin wanting a bit more urban
crush and a little less desert island.
People engage in a variety of behaviors to alleviate the pain of
social disconnection. For example, they actively seek connec-
tions with other people (Maner, DeWall, Baumeister, & Schaller,
2007), imagine important social relationships (Twenge, Cat-
anese, & Baumeister, 2003), and increase attention to social
cues in the environment (Gardner, Pickett, Jeffries, & Knowles,
2005). Such behaviors involve attempts to establish connections
with other existing humans, but we suggest that disconnected
people may adopt an even more creative approach by inventing
humanlike agents in their environment to serve as potential
sources of connection. People may do so in at least two distinct
ways: by anthropomorphizing nonhuman agents such as me-
chanical devices and nonhuman animals to make them appear
more humanlike or by increasing belief in the existence of
commonly anthropomorphized religious agents (such as God;
Guthrie, 1993). This creation of humanlike agents is likely to
arise through the increased attention to social cues activated by
social disconnection or through the active search for potential
sources of connection.
Address correspondence to Nicholas Epley, University of Chicago
GSB, 5807 South Woodlawn Ave., Chicago, IL 60637, e-mail: epley@
chicagogsb.edu.
PSYCHOLOGICAL SCIENCE
114 Volume 19—Number 2Copyright r2008 Association for Psychological Science
Some existing evidence is consistent with our prediction. Case
studies of people undergoing extreme isolation suggest that they
fairly quickly begin holding conversations with imagined peo-
ple, religious deities, or animals (e.g., God; Kirkpatrick, Shillito,
& Kellas, 1999). Chronic isolation has been a long-standing
explanation for classic examples of anthropomorphism—from
seeing mermaids in the ocean to naming geological features by
their humanlike features (e.g., Loukatos, 1976). Recent survey
evidence suggests that people feel more connected to their god
when praying alone than when praying in a group (Adler, 2005),
that people who are not in a committed relationship are more
likely than those who are to report having a personal relationship
with God (Granqvist & Hagekull, 2000), and that people with
insecure and anxious attachments to others are likely to hold the
strongest religious beliefs (Kirkpatrick & Shaver, 1990). Finally,
the death of a loved one tends to increase religious beliefs (e.g.,
McIntosh, Silver, & Wortman, 1993; Wuthnow, Christiano, &
Kuzlowski, 1980), a coping strategy that may be facilitated by
anthropomorphic conceptions of religious agents (Pargament,
1997).
None of this research, however, empirically tested the hy-
pothesis that chronic or momentary social disconnection will
lead people to create humanlike agents in their environment. We
conducted such a test in three studies examining people’s
evaluations of mechanical gadgets, pets, and religious agents.
These studies used both experimental manipulations and dis-
positional measures of social connection. One of the defining
features of human agency is the presence of higher-order mental
states, and such mental-state attributions therefore served as our
critical dependent variable in perceptions of gadgets (Study 1)
and pets (Study 3). Because religious agents are notoriously
anthropomorphized (see Guthrie, 1993), we simply measured
belief in the existence of specific supernatural agents or asso-
ciated forces (Studies 2 and 3). We predicted that participants
who were chronically lonely or momentarily induced to think
about loneliness would create agents of social connection by
altering the mental states they attributed to nonhuman agents
(Studies 1 and 3) or by increasing their belief in supernatural
agents (Studies 2 and 3).
STUDY 1: GADGETS
Method
Twenty volunteers (14 female, 6 male) completed an on-line
survey in which they read descriptions of four technological
gadgets: Clocky (a wheeled alarm clock that ‘‘runs away’’ so that
you must get up to turn it off), CleverCharger (a battery charger
designed to prevent overcharging), Pure Air (an air purifier for
people with allergies or respiratory problems), and Pillow Mate
(a torso-shaped pillow that can be programmed to give a ‘‘hug’’).
After reading each description, participants completed five
anthropomorphic mental-state ratings (a5.81): the extent to
which the gadget had ‘‘a mind of its own,’’ had ‘‘intentions,’’ had
‘‘free will,’’ had ‘‘consciousness,’’ and ‘‘experienced emotions.’’
Participants also completed three items unrelated to the cre-
ation of a humanlike agent (a5.71): the extent to which each
device was attractive, efficient, and strong. Finally, participants
completed a three-item loneliness scale (a5.81; e.g., ‘‘How
often do you feel isolated from others?’’ taken from Hughes,
Waite, Hawkley, & Cacioppo, 2004).
Results and Discussion
Gender did not significantly influence any of the analyses and is
therefore not discussed further.
To analyze participants’ responses, we first created three sep-
arate composite measures—one for the anthropomorphic mental-
state ratings, one for the nonanthropomorphic ratings, and one
for the loneliness scale—by averaging ratings for the individual
items within each of these measures.
As predicted, loneliness was significantly correlated with the
anthropomorphic mental-state ratings, r(18) 5.53, p5.02,
p
rep
5.93, but not significantly correlated with the nonan-
thropomorphic ratings, r(18) 5.25, p5.29, p
rep
5.64. The
anthropomorphic mental-state ratings and the nonanthropomor-
phic ratings were themselves significantly correlated, r(18) 5
.58, p5.01, p
rep
5.95, but the correlation between loneliness
and the anthropomorphic mental-state ratings remained signifi-
cant even after we controlled for the nonanthropomorphic rat-
ings, r(17) 5.49, p5.03, p
rep
5.91.
Although the results of this study are consistent with our
prediction, a correlational study using a dispositional measure
of social disconnection cannot demonstrate that social discon-
nection caused the observed results. We therefore adopted an
experimental approach in the next two studies in order to ma-
nipulate social connection directly.
STUDY 2: GODS
Philosophers can vaguely note that religion is an opiate for the
masses (Marx, 1844/1959), but psychologists need to explain
more specifically which pain people might take this drug to al-
leviate. Belief in religious agents can serve to alleviate several
psychological problems (e.g., awareness of one’s own mortality;
Norenzayan & Hansen, 2006), and we predicted that one of
those is the pain of social disconnection (see also Burris, Batson,
Altstaedten, & Stephens, 1994; Kirkpatrick et al., 1999). People
who are induced to feel disconnected from other humans should
therefore report a stronger belief in religious agents than those
induced to feel socially connected.
Not all religious agents, of course, are conceptualized as the
kind one might willingly seek when in need of social connection.
The overwhelming majority of religious systems contain both
loving supernatural agents (e.g., God) who perform desirable
actions (e.g., miracles) and evil supernatural agents (e.g., Devil)
who perform unpleasant actions (e.g., curses). Social discon-
Volume 19—Number 2 115
N. Epley et al.
nection could increase belief in some religious agents or actions
but not in others. Such a result seems unlikely, however, given
the high intercorrelation between belief in one religious agent
within a tradition and belief in other agents within that same
tradition. Belief in God within a Judeo-Christian tradition, for
instance, entails belief in the Devil as well. We measured belief
in a variety of religious agents and associated actions in Study 2,
however, to examine this issue.
Method
A pool of potential participants from the University of Chicago
completed an on-line survey that included the question, ‘‘Do you
believe in God?’’ Ninety-nine of these people (57 female, 42
male)—50 believers who responded ‘‘yes’’ and 49 nonbelievers
who responded ‘‘no’’—were then invited to participate in Study
2 in exchange for $7.
After arriving at the laboratory, participants first completed a
computerized version of the 90-item Eysenck Personality
Questionnaire (Eysenck & Eysenck, 1976) and were then told
that the computer would generate a set of future-life predictions
based on their responses. To bolster participants’ belief in the
accuracy of the predictions, we told them that the personality
questionnaire was a measure of extraversion and provided ac-
curate feedback on the 12-item Extraversion subscale.
We then manipulated social connection using a procedure
developed by Twenge, Baumeister, Tice, and Stucke (2001). All
participants received a paragraph of ‘‘future life predictions’’
ostensibly based on their personality profile. Those in the dis-
connected condition read statements suggesting that they would
be lonely in their lives (e.g., ‘‘You’re the type who will end up
alone later in life’’). Those in the connected condition read
statements suggesting that they would be socially connected in
their lives (e.g., ‘‘You’re the type who has rewarding relation-
ships throughout life’’). Approximately half of the participants
within each experimental condition were religious believers,
and the other half were nonbelievers.
Participants were then asked to ‘‘begin Part 2 of the study,’’ in
which they rated the extent to which they believed in ghosts,
angels, the Devil, miracles, curses, and God, on scales ranging
from 1 (not at all)to10(very much).
Results and Discussion
Responses to the six supernatural-agent items were highly in-
tercorrelated (a5.89). Including the items as a repeated
measures factor did not produce a significant interaction in the
following analysis (F<1) or alter the significance levels of any
overall analyses. We therefore averaged responses to the six
items into a single composite measure to ease presentation.
A 2 (religious believer: yes vs. no) 2 (condition: disconnected
vs. connected) analysis of variance (ANOVA) on participants’
supernatural belief revealed only two significant main effects. Not
surprisingly, participants who reported believing in God before
the experiment reported stronger belief in the supernatural agents
(M55.81) than those who reported not believing in God
(M52.25), F(1, 95) 5126.44, p<.001, p
rep
5.99, Z
2
5.57.
More important, those in the disconnected condition also reported
stronger belief in supernatural agents (M54.35) than those in
the connected condition (M53.71), F(1, 95) 53.98, p<.05,
p
rep
5.88, Z
2
5.04. The interaction was nonsignificant (F5
0.42). Social disconnection does not turn atheists into funda-
mentalists, of course, but it may nudge religious belief in the same
direction for believers and nonbelievers alike.
STUDY 3: GREYHOUNDS
We designed Study 3 to extend this research in three ways. First,
we sought to increase generalizability by investigating percep-
tions of an additional kind of nonhuman agent, namely, pets. We
measured whether the likelihood of attributing humanlike
mental states or traits to pets would be greater among people
induced to feel social disconnection than among people not so
induced. We classified mental-state traits as being humanlike on
the basis of existing research that identifies metacognition as a
critical feature distinguishing traits seen as humanlike from
those shared by other living agents (Cortes, Demoulin, Rodri-
guez, Rodriguez, & Leyens, 2005; Demoulin et al., 2004;
Haslam, Bain, Douge, Lee, & Bastian, 2005). Considering pets
also allowed us to compare the impact of loneliness on evalua-
tions of humanlike traits directly related to providing social
connection or support (e.g., thoughtful, considerate, sympa-
thetic) and evaluations of traits less directly related to social
connection or support (e.g., embarrassable, creative). If social
disconnection leads people to create humanlike agents in their
environment in order to regain social connection or support,
then social disconnection should have a greater influence on
traits closely related to social connection than on mental or
behavioral traits unrelated to social connection. Religious
agents provide a poor test of this specificity hypothesis because
beliefs in different religious agents are highly intercorrelated,
but mental-state attributions should be more independent.
Second, Studies 1 and 2 left open an important ambiguity.
People who are disconnected from others feel lonely, but they
also feel profoundly negative. This negative emotional state
alone—rather than social disconnection per se—may explain
the results observed in those studies. We therefore included a
very different negative mood state in Study 3—fear of another
person—to investigate whether negative affect alone can ex-
plain the results observed thus far. We predicted that it cannot,
and that participants in the disconnected condition would report
greater belief in supernatural agents and evaluate their pet as
being more socially supportive compared with participants in
the fear and control conditions.
Finally, including a fear condition enabled a third test. In
particular, we have suggested that social disconnection leads
116 Volume 19—Number 2
Creating Social Connection
people to create humanlike agents out of existing (or ostensibly
existing) nonhuman agents—a kind of anthropomorphism
(Epley, Waytz, & Cacioppo, 2007). This process is commonly
conflated with the superficially related process of detecting
living agents in one’s environment—a process more closely
related to animism. People detect images of the Virgin Mary in
tree bark, animal forms in clouds, and threatening strangers in
the shadows of darkened alleys (Guthrie, 1993). Detecting an
agent in the immediate environment requires vigilance and
attention, and fear can increase vigilance for specific threat-
ening or fear-inducing agents that may pose a threat (Maner
et al., 2007). Fear caused specifically by the potential threat of
another person should increase the tendency to detect human-
like agents in one’s immediate environment, and we therefore
expected that participants in the fear condition would be more
likely to detect a humanlike agent in an ambiguous display than
would those in the disconnected and control conditions. We
tested this prediction by also asking participants to view 20
ambiguous drawings and report what they saw. Of these 20, 10
were drawn to vaguely resemble a human face (see Fig. 1). We
predicted that participants in the fear condition would report
seeing more faces than those in the disconnected and control
conditions. Such a dissociation between social disconnection
and fear across measures related to anthropomorphism versus
animism would mark an important boundary condition for the
influence of sociality motivation in perceptions of humanlike
agency.
Method
Participants
Fifty-seven Harvard University undergraduates (44 female, 13
male) participated in Study 3 in exchange for course credit or $6.
Procedure
Participants arrived at the laboratory for an experiment inves-
tigating perceptions of ‘‘self, others, and pets.’’ Participants
learned that they would watch a short video clip and were in-
structed to experience the thoughts and emotions felt by the
protagonist in the clip as best they could by empathizing with
him. Participants were then randomly assigned to experimental
condition and watched the 3-min clip associated with that
condition.
We selected clips that, to our knowledge, best exemplified the
psychological states we intended to convey. Participants in the
disconnected condition watched a segment from Cast Away
(Zemeckis & Broyles, 2000) in which the protagonist experi-
ences severe isolation and loneliness while stranded on a de-
serted island. Those in the fear condition watched a segment
from Silence of the Lambs (Demme, Harris, & Tally, 1992) in
which the main protagonist is chasing a serial killer (see Gross &
Levenson, 1995). Those in the control condition watched a clip
from Major League (Ward, 1989) in which baseball players in-
teract in a crowd of people after a victory. This last clip was
a ‘‘control’’ only in that it (a) involved interactions with other
people (unlike the other clips) and (b) did not involve either
loneliness or fear.
Participants then completed three measures ostensibly as part
of an unrelated experiment. First, they rated their belief in the
same supernatural agents and actions used in Study 2. Second,
they were asked to think of a pet that they either owned or knew
well and to pick from a list of 14 traits those 3 that best described
this pet (ownership did not influence any of the reported results).
The list of traits included 3 anthropomorphic traits related to
social connection (thoughtful, considerate, and sympathetic),
4 anthropomorphic traits less related to social connection
(embarrassable, creative, devious, and jealous), and 7 nonan-
thropomorphic traits that are simply behavioral descriptions
(aggressive, agile, active, energetic, fearful, lethargic, and mus-
cular). Finally, participants were shown a series of 20 ambigu-
ous figures (see Fig. 1), each on a separate sheet of paper, and
asked to write down what they saw in each. The number of
faces spontaneously reported served as our key face-detection
measure.
Results and Discussion
We predicted that the three critical measures would show a
dissociation across the experimental conditions. Specifically, we
expected that participants in the disconnected condition would
report greater belief in supernatural agents and have a stronger
tendency to attribute humanlike mental states to nonhuman
Fig. 1. Examples of ambiguous images used in the face-detection mea-
sure of Study 3. The 2 images in the top row are examples of the 10 images
drawn to resemble faces, and the 2 images in the bottom row are examples
of the 10 images drawn to resemble nothing in particular.
Volume 19—Number 2 117
N. Epley et al.
agents (pets) than participants in the control and fear conditions,
but that participants in the fear condition would be more likely
to spontaneously report seeing faces in the ambiguous figures
than would those in the control and disconnected conditions.
Table 1 presents all relevant means and significance levels for
simple-effects tests. We present analyses of the individual
measures and then the overall analysis across all measures to
ease presentation.
Belief in Supernatural Agents
As in Study 2, beliefs in the different supernatural agents and
actions were highly interrelated (a5.89) and therefore col-
lapsed into a composite measure. As predicted, orthogonal
planned contrasts showed that participants in the disconnected
condition reported stronger belief in supernatural agents than
participants in the other two conditions combined, t(54) 52.36,
p5.02, p
rep
5.92, d50.57, whereas beliefs in the control and
fear conditions did not differ, p5.77.
Pet Ratings
Participants selected three traits that best described their pet.
Our critical measure was the proportion of traits selected from
each trait type out of the total number possible for that type. As
predicted, orthogonal planned contrasts revealed that partici-
pants in the disconnected condition were more likely to select
one of the three social-connection traits than were participants
in the other two conditions combined, t(54) 53.01, p<.01,
p
rep
5.97, d50.82, whereas the likelihood of selecting the
social-connection traits did not differ between the fear and
control conditions, p5.40. There was no difference across
conditions in the likelihood of selecting either the nonconnec-
tion traits or the behavioral traits, both ps>.15, and none of the
individual nonconnection or behavioral traits was chosen sig-
nificantly more often in one condition than in the others (all
ps>.2). A 3 (condition: disconnection, fear, control) 3 (traits:
connection, nonconnection, behavioral) mixed-model ANOVA
yielded only the predicted interaction, F(4, 108) 53.69, p5
.01, Z
2
5.12. These results suggest that individuals lacking
social connection may create agents that are useful conduits for
social connection.
Face Detection
Our prediction for face detection was the complement of our
prediction for the preceding two measures—that participants in
the fear condition would report detecting more faces in the
ambiguous figures than participants in the disconnected and
control conditions. Indeed, orthogonal planned contrasts con-
firmed this predicted pattern, t(54) 52.32, p5.02, p
rep
5.92,
d50.63, and the number of faces reported did not differ be-
tween the disconnected and control conditions, p5.97.
Overall Analysis
To test our overall prediction of a dissociation across measures,
we standardized participants’ responses across our three critical
dependent measures: belief in supernatural agents, percentage
of social-connection traits selected, and number of faces report-
ed. A 3 (condition: disconnected, fear, control) 3(measure:
supernatural agents, social-connection traits, faces identified)
mixed-model ANOVA using these standardized measures yielded
only the predicted interaction presented in the results already
discussed, F(4, 108) 53.67, p5.01, Z
2
5.12.
Discussion
Study 3 again confirmed our main prediction that social dis-
connection increases the tendency to create humanlike agents
out of nonhuman agents in one’s environment, and demonstrates
that this result is not simply produced by any negative emotional
state. This experiment also draws a tentative distinction between
creating humanlike agents out of nonhumans (anthropomor-
phism) and seeing humanlike agents in one’s immediate envi-
ronment (something akin to animism). At the very least, this
dissociation demonstrates that the fear manipulation in this
study had some unique effect on judgment and was not simply a
second control condition.
GENERAL DISCUSSION
And God stepped out on space,
And he looked around and said:
I’m lonely—
I’ll make me a world . . . (Johnson, 1927/1990, p. 17)
Physicists have the scientific tools to suggest that Johnson
may have gotten his poem profoundly wrong, but psychologists
have the scientific tools to suggest that Johnson may have gotten
his poem profoundly backward. In three studies, people who
were chronically disconnected from others (Study 1) or mo-
mentarily led to think about disconnection (Studies 2 and 3)
TABLE 1
Primary Dependent Measures From Study 3
Measure
Experimental condition
Disconnected Control Fear
Belief in supernatural agents 4.77
a
3.42
ab
3.20
b
Pet ratings
Social-connection traits .44
a
.23
b
.14
b
Nonconnection traits .08
a
.11
a
.13
a
Behavioral traits .20
a
.27
a
.25
a
Face detection 6.60
a
6.56
a
8.16
b
Note. Pet ratings are reported as the average proportion of traits selected out
of the total number possible for the indicated category (three social-connec-
tion traits, four nonconnection traits, and seven behavioral traits). Face de-
tection was measured as the number of images in which participants reported
seeing a face (of the 20 images, 10 were drawn to bear at least some resem-
blance to faces). Within each row, means that do not share a superscript differ
at p<.05, two-tailed. For belief in supernatural agents, the post hoc com-
parison between the disconnected and control conditions was marginally sig-
nificant, t(54) 51.88, p5.06, p
rep
5.85, d50.51.
118 Volume 19—Number 2
Creating Social Connection
appeared to create humanlike agents in their environment—
from gadgets to pets to supernatural agents such as God. These
studies go beyond simply demonstrating that social disconnec-
tion leads people to seek companionship from nonhuman agents,
showing that social disconnection can alter the way these agents
are conceptualized or represented. Lonely people cannot
make themselves a world, of course, but they can make them-
selves a mindful gadget, a thoughtful pet, or a god to populate
that world.
We believe the results of these studies inspire four especially
intriguing questions. First, does social disconnection lead peo-
ple to seek only a sense of connection with others, or does it
lead to a more specific search for positive social support? Study
3 provides some tentative evidence for a selective creation of a
socially supportive agent in participants’ pet ratings, but people
also prefer to be with socially unsupportive others than to be
isolated (Van Beest & Williams, 2006). Our research does not
unambiguously identify the exact kind of humanlike agent a
socially disconnected person is likely to create.
Second, these experiments suggest that social disconnection
leads people to create humanlike agents in their environment,
but is this inferential reproduction effective? The positive cor-
relation between pet ownership and well-being is well docu-
mented (Serpell, 1991), as is the correlation between religious
belief and well-being (Cacioppo & Brandon, 2002), but surely
some agents (e.g., God) satisfy the need for social connection
better than others (e.g., gadgets). Does variance in the degree to
which these agents satisfy social connection come from the ease
with which they can be anthropomorphized? If so, encouraging
anthropomorphism among socially disconnected individuals
may have some surprising therapeutic benefits.
Third, these experiments provided participants with the op-
portunity to alleviate a sense of social disconnection by hu-
manizing nonhuman agents, but might some people actually
prefer this method for alleviating social disconnection over
seeking connection through other humans? Individuals who are
rejected or ostracized by one group of people tend to avoid at-
tempting to reconnect with that group and instead seek con-
nection with other groups (Maner et al., 2007). It is at least
possible that people who are chronically rejected or ostracized
from other humans may eventually come to avoid seeking con-
nection with other humans altogether and instead satisfy their
sociality motivation by deliberately seeking connection with
nonhuman agents.
Finally, these experiments investigated perceptions of non-
human agents, but we believe this research has implications for
perceptions of humans as well. If social disconnection increases
the tendency to seek humanlike agents in one’s environment,
then a strong sense of social connection should decrease this
tendency to seek humanlike agents. A lack of motivation to
connect with other humans should decrease the tendency to
perceive humanlike traits in these other humans as well. This
reasoning suggests that when evaluating other individuals,
people who are especially socially connected might also be more
likely to dehumanize those to whom they are not socially con-
nected (see also Harris & Fiske, 2006; Leyens et al., 2003). We
found exactly this pattern in one recent experiment in which
participants induced to feel strongly connected to another per-
son were less likely to attribute humanlike mental states to
members of an out-group than were those not induced to feel
connected to another person (Waytz, Epley, & Cacioppo, 2007).
Social disconnection induces the motivation to create human-
like agents of social support, and social connection reduces that
motivation. Although being socially connected has many de-
sirable consequences for one’s mental and physical health, it
may have some undesirable consequences as well.
Acknowledgments—We thank Jasmine Kwong, Carlos Lozano,
and Nick Josefowitz for assisting with data collection, and the
National Science Foundation (Grant SES0241544), the Uni-
versity of Chicago Graduate School of Business, and the Temple-
ton Foundation for financial support.
REFERENCES
Adam, E.K., Hawkley, L.C., Kudielka, B.M., & Cacioppo, J.T. (2006).
Day-to-day dynamics of experience-cortisol associations in a
population-based sample of older adults. Proceedings of the Na-
tional Academy of Sciences, USA,103, 17058–17063.
Adler, J. (2005, August 29). In search of the spiritual. Newsweek,146,
46–64.
Burris, C.T., Batson, C.D., Altstaedten, M., & Stephens, S.K. (1994).
What a friend: Loneliness as a motivator of intrinsic religion.
Journal for the Scientific Study of Religion,33, 326–334.
Cacioppo, J.T., & Brandon, M.E. (2002). Religious involvement and
health: Complex determinism. Psychological Inquiry,13, 204–
206.
Cacioppo, J.T., Hawkley, L.C., & Berntson, G.G. (2003). The anatomy
of loneliness. Current Directions in Psychological Science,12, 71–
74.
Cortes, B.P., Demoulin, S., Rodriguez, R.T., Rodriguez, A.P., & Leyens,
J.P. (2005). Infrahumanization or familiarity?: Attribution of
uniquely human emotions to the self, the ingroup, and the out-
group. Personality and Social Psychology Bulletin,31, 243–253.
Demme, J. (Director), Harris, T. (Writer), & Tally, T. (Writer). (1992).
Silence of the lambs [Motion picture]. United States: Orion Pic-
tures Corp.
Demoulin, S., Leyens, J.P., Paladino, M.P., Rodriguez, R.T., Rodriguez,
A.P., & Dovidio, J.F. (2004). Dimensions of ‘‘uniquely’’ and ‘‘non-
uniquely’’ human emotions. Cognition & Emotion,18, 71–96.
Diener, E., & Seligman, M.E.P. (2002). Very happy people. Psycho-
logical Science,13, 80–83.
Epley, N., Waytz, A., & Cacioppo, J.T. (2007). On seeing human: A
three-factor theory of anthropomorphism. Psychological Review,
114, 864–886.
Eysenck, H.J., & Eysenck, S.B.G. (1976). Psychoticism as a dimension
of personality. London: Hodder & Stoughton.
Gardner, W.L., Pickett, C.L., Jefferies, V., & Knowles, M. (2005). On
the outside looking in: Loneliness and social monitoring. Per-
sonality and Social Psychology Bulletin,31, 1549–1560.
Volume 19—Number 2 119
N. Epley et al.
Granqvist, P., & Hagekull, B. (2000). Religiosity, adult attachment,
and why ‘‘singles’’ are more religious. International Journal for
the Psychology of Religion,10, 111–123.
Gross, J.J., & Levenson, R.W. (1995). Emotion elicitation using films.
Cognition & Emotion,9, 87–108.
Guthrie, S.E. (1993). Faces in the clouds: A new theory of religion. New
York: Oxford University Press.
Harris, L.T., & Fiske, S.T. (2006). Dehumanizing the lowest of the low:
Neuroimaging responses to extreme out-groups. Psychological
Science,17, 847–853.
Haslam, N., Bain, P., Douge, L., Lee, M., & Bastian, B. (2005). More
human than you: Attributing humanness to self and others. Jour-
nal of Personality and Social Psychology,89, 937–950.
Hawkley, L.C., Masi, C.M., Berry, J.D., & Cacioppo, J.T. (2006).
Loneliness is a unique predictor of age-related differences in
systolic blood pressure. Psychology and Aging,21, 152–164.
Heinrich, L.A., & Gullone, E. (2006). The clinical significance of
loneliness: A literature review. Clinical Psychology Review,26,
695–718.
House, J.S., Landis, K.R., & Umberson, D. (1988). Social relationships
and health. Science,241, 540–545.
Hughes, M.E., Waite, L.J., Hawkley, L.C., & Cacioppo, J.T. (2004). A
short scale for measuring loneliness in large surveys: Results
from two population-based studies. Research on Aging,26, 655–
672.
Johnson, J.W. (1990). God’s trombones: Seven Negro sermons in verse.
New York: Viking-Penguin. (Original work published 1927)
Kirkpatrick, L.A., & Shaver, P.R. (1990). Attachment theory and re-
ligion, childhood attachments, religious beliefs, and conversion.
Journal for the Scientific Study of Religion,29, 315–334.
Kirkpatrick, L.A., Shillito, D.J., & Kellas, S.L. (1999). Loneliness,
social support, and perceived relationships with God. Journal of
Social and Personal Relationships,16, 513–522.
Leyens, J.-P., Cortes, B.P., Demoulin, S., Dovidio, J., Fiske, S.T.,
Gaunt, R., et al. (2003). Emotional prejudice, essentialism, and
nationalism. European Journal of Social Psychology,33, 703–
717.
Loukatos, D. (1976). Personification of capes and rocks in the Hellenic
seas. In A. Bharati (Ed.), The realm of the extra-human (pp. 467–
474). Paris: Mouton.
Maner, J.K., DeWall, C.N., Baumeister, R.F., & Schaller, M. (2007).
Does social exclusion motivate interpersonal reconnection? Re-
solving the ‘‘porcupine problem.’’ Journal of Personality and
Social Psychology,92, 42–55.
Marx, K. (1959). Economic and philosophic manuscripts of 1844.
Moscow: Progress Publishers. (Original work published 1844)
McIntosh, D.N., Silver, R.C., & Wortman, C.B. (1993). Religion’s role in
adjustment to a negativelife event: Coping with the loss of a child.
Journal of Personality and Social Psychology,65, 812–821.
Norenzayan, A., & Hansen, I.G. (2006). Belief in supernatural agents
in the face of death. Personality and Social Psychology Bulletin,
32, 174–187.
Pargament, K. (1997). The psychology of religious coping. New York:
Guilford.
Serpell, J. (1991). Beneficial effects of pet ownership on some aspects
of human health and behavior. Journal of the Royal Society of
Medicine,84, 717–720.
Twenge, J.M., Baumeister, R.F., Tice, D.M., & Stucke, T.S. (2001). If
you can’t join them, beat them: Effects of social exclusion on
aggressive behavior. Journal of Personality and Social Psychol-
ogy,81, 1058–1069.
Twenge, J.M., Catanese, K.R., & Baumeister, R.F. (2003). Social ex-
clusion and the deconstructed state: Time perception, meaning-
lessness, lethargy, lack of emotion, and self-awareness. Journal of
Personality and Social Psychology,85, 409–423.
Van Beest, I., & Williams, K.D. (2006). When inclusion costs and
ostracism pays, ostracism still hurts. Journal of Personality and
Social Psychology,91, 918–928.
Ward, D.S. (Writer/Director). (1989). Major league [Motion picture].
United States: Mirage.
Waytz, A., Epley, N., & Cacioppo, J.T. (2007). [Social connection in-
creases dehumanization]. Unpublished raw data.
Wuthnow, R., Christiano, K., & Kuzlowski, J. (1980). Religion and
bereavement: A conceptual framework. Journal for the Scientific
Study of Religion,19, 408–422.
Zemeckis, R. (Director), & Broyles, W., Jr. (Writer). (2000). Cast away
[Motion picture]. United States: DreamWorks SKG.
(RECEIVED 4/9/07; REVISION ACCEPTED 6/29/07)
120 Volume 19—Number 2
Creating Social Connection
