Content uploaded by Simona Sacchi
Author content
All content in this area was uploaded by Simona Sacchi on Jan 10, 2018
Content may be subject to copyright.
Journal of Experimental Psychology: Applied
Humanizing Machines: Anthropomorphization of Slot
Machines Increases Gambling
Paolo Riva, Simona Sacchi, and Marco Brambilla
Online First Publication, August 31, 2015. http://dx.doi.org/10.1037/xap0000057
CITATION
Riva, P., Sacchi, S., & Brambilla, M. (2015, August 31). Humanizing Machines:
Anthropomorphization of Slot Machines Increases Gambling. Journal of Experimental
Psychology: Applied. Advance online publication. http://dx.doi.org/10.1037/xap0000057
Humanizing Machines: Anthropomorphization of Slot Machines
Increases Gambling
Paolo Riva, Simona Sacchi, and Marco Brambilla
University of Milano-Bicocca
Do people gamble more on slot machines if they think that they are playing against humanlike minds
rather than mathematical algorithms? Research has shown that people have a strong cognitive tendency
to imbue humanlike mental states to nonhuman entities (i.e., anthropomorphism). The present research
tested whether anthropomorphizing slot machines would increase gambling. Four studies manipulated
slot machine anthropomorphization and found that exposing people to an anthropomorphized description
of a slot machine increased gambling behavior and reduced gambling outcomes. Such findings emerged
using tasks that focused on gambling behavior (Studies 1 to 3) as well as in experimental paradigms that
included gambling outcomes (Studies 2 to 4). We found that gambling outcomes decrease because
participants primed with the anthropomorphic slot machine gambled more (Study 4). Furthermore, we
found that high-arousal positive emotions (e.g., feeling excited) played a role in the effect of anthropo-
morphism on gambling behavior (Studies 3 and 4). Our research indicates that the psychological process
of gambling-machine anthropomorphism can be advantageous for the gaming industry; however, this
may come at great expense for gamblers’ (and their families’) economic resources and psychological
well-being.
Keywords: anthropomorphism, social cognition, mind attribution, slot machines, gambling
They (slot machines) sit there like young courtesans,
promising pleasures undreamed of,
your deepest desires fulfilled, all lusts satiated.
—Frank Scoblete (b. 1947)
As the quote by famous gambler Frank Scoblete suggests,
players might be prone to perceive slot machines as humanlike—
capable of seducing millions of people around the world with their
promise of fulfilling their dreams—rather than as inanimate ob-
jects, programmed with an overall set amount of payouts. Perhaps
it is not coincidental that so many anthropomorphic representa-
tions (including “sexy” women, muscular men, cartoon characters,
and pharaohs) appear on slot machines in casinos and on gambling
websites. It is possible that—similar to how in ancient mythology
challengers perceived the sphinx as possessing a humanlike mind
and intelligence—the gambling industry is selling customers a
challenge against a mind rather than just a machine (or a computer
algorithm).
1
Based on these observations and considering recent
theorizing on anthropomorphism and social perception (Epley,
Waytz, & Cacioppo, 2007), we investigated the hypothesis that
anthropomorphizing slot machines would cause an increase in
gambling.
Slot Machines
It is widely known that excessive gambling can have a detri-
mental impact on a person’s life, including their health, employ-
ment, family, and interpersonal relationships (Griffiths, 2004). In
recent years, we have witnessed a continuous increase in slot
machine (also referred to as “fruit machines” in the United King-
dom, and “poker machines” in Australia and New Zealand; see
Figure 1A for an illustration of the interface of these devices)
gambling. What is more, gambling on slot machines has been
described as one of the most addictive forms of gambling (Grif-
fiths, 2004; Turner & Horbay, 2004).
Despite their harmful effects, the number of slot machines
worldwide is increasing: Slot machines can be found in casinos
(where they are often the most popular gambling game), race-
tracks, local bars, and on websites (e.g., online slot machines).
Managers primarily invest in slot machines because they are
profitable (even considering other gambling games such as
roulette and blackjack). Modern slot machines are controlled
using software that runs an algorithm designed to achieve a
certain overall payback percentage, which is set at the factory
when the software is written (Turner & Horbay, 2004). The
1
This possibility was initially conceived through the observation of a
slot machine in a local bar whose interface was titled “Revenge of
Cheope.”
Paolo Riva, Simona Sacchi, and Marco Brambilla, Department of Psy-
chology, University of Milano-Bicocca.
We thank Marina Azzola, Alessandra Fabris, Elisabetta Moia, Serena
Monaco, Fabrizio Pontiggia, and Michael Schepisi for their help in data
collection. We extend our special thanks to Guglielmo Tobia for helping us
with the data collection of the preliminary study. We also want to thank
Daniel Bennett for proofreading this article.
Correspondence concerning this article should be addressed to Paolo
Riva, University of Milano-Bicocca, Department of Psychology, Piazza
Ateneo Nuovo, 1, 20126 Milano, Italy. E-mail: paolo.riva1@unimib.it
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Journal of Experimental Psychology: Applied © 2015 American Psychological Association
2015, Vol. 21, No. 3, 000 1076-898X/15/$12.00 http://dx.doi.org/10.1037/xap0000057
1
payback percentage can vary (usually somewhere around 80%;
for the United Kingdom, see U.K. Gambling Commission,
2013; for a U.S. state, see Pacode.com., 2015) according to
national jurisdictions, but any number under 100 means that the
casinos, gaming operators, and website owners are the real
winners. It has been estimated that slot machines generate over
70% of the average casino’s income (Chen, Shoemaker, &
Zemke, 2013; Smith & Wynne, 2004).
If the reasons leading operators to invest money in slot machines
are clear, more research is needed to uncover the cognitive and
motivational factors that make slot machines so attractive to gam-
blers (Griffiths, 1994, 2004). The present research addressed this
issue considering recent work on anthropomorphism and mind
attribution.
Anthropomorphism
Anthropomorphism is the tendency to ascribe humanlike fea-
tures, such as mind states, intentions, free will, consciousness, and
emotions (Epley et al., 2007) to nonhuman agents. Targets of
anthropomorphism range from inanimate objects (Chandler &
Schwarz, 2010; Hart, Jones, & Royne, 2013) to religious agents
(Morewedge & Clear, 2008), to all types of machines (Bartneck,
Kuli´
c, Croft, & Zoghbi, 2009), animals (Chartrand, Fitzsimons, &
Fitzsimons, 2008), and nature (Sacchi, Riva, & Brambilla, 2013).
Anthropomorphism permeates our cognitive life, as it satisfies
important cognitive and motivational needs (Epley et al., 2007;
Epley, Waytz, Akalis, & Cacioppo, 2008; Gray, Gray, & Wegner,
2007; Waytz, Morewedge, et al., 2010) and may have implications
for our social life (Sacchi et al., 2013). Recent conceptualizations
of anthropomorphism identified two main determinants: effectance
motivation and sociality motivation (Epley et al., 2007). According
to the former, anthropomorphism makes the agent’s behavior
understandable and predictable, as this behavior is considered to be
driven by intentionality (Epley et al., 2007). Thus, effectance
motivation refers to people’s desire to understand their environ-
ments. Research has shown that individuals who have a stable need
for control seem to be especially likely to endorse anthropomor-
phic beliefs (Epley et al., 2008). According to sociality motivation,
anthropomorphism produces a sense of social connection with
another humanlike mind. Sociality motivation thus refers to the
fundamental need for social connection (Baumeister & Leary,
1995) as well as people’s vigilance for agents (including both
human and nonhuman beings) that can potentially fulfill this
fundamental need. Research has found support for this prediction
by showing that people who feel chronically lonely are more likely
to anthropomorphize nonhuman agents (i.e., pets; see Epley et al.,
2008).
Among different forms of anthropomorphization, the projection
of humanlike features on inanimate objects and machines is a
widespread cognitive tendency (e.g., Bartneck et al., 2009; Eyssel
& Kuchenbrandt, 2012). Whereas the formation of parasocial
relationships between humans and objects was once considered
something peculiar to child development (Winnicott, 1953), recent
evidence suggests that adults report entering into relationships
with objects as well (Gardner & Knowles, 2008). Similarly, people
often reason about anthropomorphized objects as if they were
human beings. Research has shown that people’s behavior toward
technological agents follow the rules they adopt in their social
lives (Nass & Moon, 2000; Waytz, Cacioppo, & Epley, 2010). For
instance, human gender stereotypes can be activated merely by the
presence of a male versus a female computer voice. In one study,
participants rated a computerized male voice as more compelling
than a computerized female voice, whereas a computerized female
was rated as friendlier than a male computerized voice (Nass,
Moon, & Green, 1997). Furthermore, prior findings revealed that
people are prone to apply human social categories such as ethnicity
to computers (Nass & Moon, 2000), and to more strongly anthro-
pomorphize robots that belong to their own ethnic in-group (Eyssel
& Kuchenbrandt, 2012).
Crucial to the purpose of the present research, slot machine
anthropomorphization may have important psychological and be-
havioral consequences. Previous research suggested the possibility
that slot machine anthropomorphism could be linked to gambling
A
B
Figure 1. (A) An example of an interface similar to what we used in
Study 1, Study 2, and Study 3; image by Panzerhorgen (own work; GFDL;
http://www.gnu.org/copyleft/fdl.html) or CC BY-SA 3.0 (http://
creativecommons.org/licenses/by-sa/3.0), via Wikimedia Commons. (B)
The interface created by the authors and used for Study 4. See the online
article for the color version of this figure.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
2RIVA, SACCHI, AND BRAMBILLA
behavior. Walker (1992) analyzed the verbalizations people made
while playing with several electronic devices. Crucially, among
these verbalizations were statements that suggested the “personi-
fication of the machine.” Subsequent research on regular gamblers
has further suggested that one of their most common verbalizations
is the attribution of human features to the gambling device (e.g.,
attributing emotions to the device, insulting the device; Delfabbro
& Winefield, 2000; see also Griffiths, 1994). A more recent study
focusing on slot machine anthropomorphization also reported an
interactive effect among people’s feelings of power, the percentage
of gambling winnings, and anthropomorphization (Kim & McGill,
2011). Those with high social power were more likely to anthro-
pomorphize slot machines after winning than after losing. How-
ever, none of these studies experimentally tested the hypothesis
that anthropomorphizing slot machines would lead people to gam-
ble more or the reasons why such effect might occur, which leads
us to the present research.
The Present Research
Building on prior research into anthropomorphism, the present
research investigated the psychological and behavioral conse-
quences of anthropomorphizing inanimate objects, particularly slot
machines. Based on recent work on anthropomorphism (Aaker,
Fournier, & Brasel, 2004; Epley et al., 2007; Schultz, Kleine, &
Kernan, 1989; Wang, Baker, Wagner, & Wakefield, 2007),
we tested the hypothesis that anthropomorphism and gambling are
linked, and—more specifically—that humanizing a slot machine
can increase gambling behavior. We expect that if a person has the
perception that a gambling device possesses a humanlike mind, it
would lead them to gamble more compared with those who have
the perception that the same gambling device is based on an
algorithm preprogrammed to deliver a certain number of wins and
losses. The latter description is what a slot machine is in actuality:
a computer algorithm that determines where the pictures on the
reels will stop and when to pay out (Turner & Horbay, 2004).
However, the pervasive cognitive tendency to attribute humanlike
features to nonhuman entities (e.g., including machines) might
lead players to perceive a gambling session as a challenge against
a person rather than against a preprogrammed device.
More specifically, building on prior research suggesting that
gamblers experience more excitement during gambling and higher
levels of depressive moods before gambling than nonregular gam-
blers (Griffiths, 1995), together with the literature showing that
people with a strong sense of social connection with a specific
object tend to feel stronger emotions toward it (e.g., Aaker et al.,
2004; Schultz et al., 1989; Wang et al., 2007), we anticipated that
the emotional experience during the game could account for the
relationship between slot machine anthropomorphism and gam-
bling behavior. In particular, we expected that people exposed to
an anthropomorphic description of a slot machine would experi-
ence a higher emotional reaction, which in turn would lead them to
gamble longer.
Overall, the current research tested the hypothesis that anthro-
pomorphizing slot machines would lead people to gamble more.
More specifically, we tested (a) whether anthropomorphism and
gambling behavior are linked, (b) whether anthropomorphizing
slot machines increases gambling behavior, (c) whether anthropo-
morphism affects the gambling outcomes (i.e., the actual win-
nings), and (d) whether the emotional experience during gambling
accounts for the possible link between anthropomorphism and
gambling behavior.
Preliminary Study
We conducted a preliminary study to explore whether anthro-
pomorphism could be linked to gambling behavior. To this effect,
we conducted a quasi-experiment in which we asked a group of
regular players and an equivalent group of nonregular players to
rate slot machines on several human traits.
Method
Participants. Participants were a group of 15 regular slot
machine players (one female; Mage ⫽58.07 years, SD ⫽14.55)
and a group of 15 nonregular slot machine players (one female; M
age ⫽55.47 years, SD ⫽13.88). Both regular and nonregular
players were recruited through snowball sampling among regular
and nonregular slot machine players.
Materials and procedure. Participants in both groups were
asked to complete a questionnaire that included a measure of
participants’ frequency of slot machine play (i.e., “How often do
you play slot machines?”—from less than once a year to every
day) and a measure of participants’ expenditure on slot machines
(i.e., “On average, how much money do you spend weekly playing
slot machines?”). The participants were then asked to express their
agreement on 15 items designed to assess their tendency to an-
thropomorphize slot machines (e.g., “The slot machine acts ac-
cording to its own intentions” and “The slot machine experiences
emotions”; adapted from Waytz, Cacioppo, et al., 2010; see the
Appendix for the complete item descriptions). Responses were
recorded ona1(totally disagree)to5(totally agree) scale and
were averaged to create an overall index of slot machine anthro-
pomorphization (␣⫽.84).
Results and Discussion
All regular players indicated that they usually played at least
once a week. All nonregular players indicated that they usually
played less than once a year. The regular players reported an
average weekly expenditure of €103.67 (SD ⫽124.26), with a
minimum of €5 per week and a maximum of €350 per week. All
nonregular players reported no weekly expenditure (€0) for slot
machine play.
Next, we computed a ttest comparing the two groups. Our
results showed that regular slot machine players ascribed human
traits to slot machines (M⫽2.95, SD ⫽.73) more than nonregular
slot machine players (M⫽1.94, SD ⫽.62), t(28) ⫽4.09, p⬍
.001, d⫽1.49, 95% confidence interval (CI) [0.68, 2.30]. Fur-
thermore, we found that the tendency to attribute a humanlike
mind to gambling devices and the self-reported frequency of play
were positively related, r⫽.63, p⬍.001. Thus, our preliminary
study suggested that, within a community sample, gambling be-
havior and the tendency to attribute human traits to gambling
machines might be linked.
Study 1
Study 1 was designed as the first test of our hypothesis that
anthropomorphizing slot machines increases gambling behavior.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
3
HUMANIZING SLOT MACHINES INCREASES GAMBLING
Indeed, our preliminary study leaves the question of whether slot
machine anthropomorphism per se can increase people’s gambling
behavior completely unanswered. Thus, in Study 1, we manipu-
lated slot machine anthropomorphism and then measured people’s
actual gambling behavior. We predicted that people would gamble
more when they were led to perceive the machine in an anthropo-
morphic way.
Method
Participants and design. Eighty-five participants (36 female;
Mage ⫽26.26 years, SD ⫽8.12) volunteered to participate in this
study. The experiment used a one-way, two-level, between-
subjects design. The independent variable was being primed about
slot machines anthropomorphically (vs. nonanthropomorphically).
The dependent variable was the number of spins that the partici-
pants made while interacting with an authentic online slot ma-
chine.
Materials and procedure. In Study 1, the effect of slot ma-
chine anthropomorphization on gambling behavior was tested
through a cover story. Participants were tested individually and
were told that the study was part of a market research program.
More specifically, participants were informed that the researchers
were interested in the evaluation of the graphical features of slot
machines. For this purpose, participants were told that they could
interact with a slot machine, ostensibly to collect their impressions
concerning the graphical features of the slot machine. They read a
description on the history and the dissemination of the slot ma-
chine on a laptop computer screen. They were then presented with
a series of seven images of slot machine interfaces, including the
one they were connected to later on in the experiment. Then half
of the participants (i.e., those in the anthropomorphism condition)
were presented with the following text on a single slide:
Remember that when you play with a slot machine, you don’t need to
implement any particular strategy. Indeed, the slot machine can decide
whether you will win or lose a series of bets any time she wants.
Sometimes, she may choose to make fun of you, leaving you empty-
handed for several bets; other times, she might want to reward you
with a win. In any case, the slot machine will always choose what will
happen. You just have to start playing and see what happens.
The other half of participants (i.e., those in the nonanthropo-
morphism condition) were presented with the following text on a
single slide:
Remember that when you play with a slot machine, you don’t need to
implement any particular strategy. Indeed, the slot machine is con-
trolled by a mathematical algorithm that is programmed to deliver a
certain overall number of wins and losses. Based on this algorithm,
you can win or lose a series of bets. In any case, the outcome of each
turn of the reels is always run by a computer algorithm. You just have
to start playing and see what happens.
These texts were based on items related to intention and cogni-
tion, which are included in the Mind Attribution Scale (Kozak,
Marsh, & Wegner, 2006; see also Waytz, Cacioppo, et al., 2010)
that is generally adopted to assess anthropomorphization. Partici-
pants were randomly assigned to one of the two experimental
conditions. The remaining part of the instructions was identical in
the two conditions. After reading it, participants were reminded
that their task was to evaluate the graphical features of a slot
machine interface. To complete the task, they were told they could
only use two keys while interacting with a real online slot ma-
chine: “Enter” to spin the reels, or “Esc” to quit the game. Partic-
ipants were told that they could spend as much time as they wanted
interacting with the slot machine and that they could quit the game
whenever they wanted. Participants in both conditions were pre-
sented with an actual online slot machine website
2
(see Figure 1A
for an illustration of an interface similar to that used in Study 1).
Every time they pressed the “Enter” button, the reels on the screen
spun and the participant could see whether they had won or lost
points on the slot machine interface. A custom-made application
running in the background recorded the number of times each
participant pressed the “Enter” key before quitting the slot ma-
chine.
When participants decided to quit the game (by pressing the
“Esc” key), they were presented with a series of questions. In
actuality, no questions related to the graphical features of the slot
machine were asked. They were asked to complete 15 items
assessing their tendency to ascribe the slot machine with human
mental states such as intention, consciousness, free will, and the
capacity to experience emotions (e.g., “The slot machine acts
according to its own intentions”; ␣⫽.86; see Appendix). Re-
sponses were recorded ona1(totally disagree)to5(totally agree)
scale and were averaged to create an overall index of gambling
machine anthropomorphism. Finally, participants’ demographic
data (i.e., age, sex, and nationality) were collected. At the end of
the experiment, each participant was debriefed and the purpose of
the study was explained in detail.
Results and Discussion
First, in order to check the effectiveness of our manipulation, we
computed an independent samples ttest on anthropomorphism
scores, comparing the two experimental groups. The analysis
revealed that participants anthropomorphized the slot machine
more in the anthropomorphism condition (M⫽2.53, SD ⫽.70)
than in the nonanthropomorphism condition (M⫽2.09, SD ⫽
.68), t(83) ⫽2.91, p⫽.005, d⫽.64, 95% CI [0.20, 1.07].
Then, we computed a second ttest comparing the two experi-
mental conditions on the critical dependent variable, that is, the
number of reel spins made by participants. The analysis showed
that participants in the anthropomorphism condition spun the reels
more times (M⫽31.88, SD ⫽17.64) than the participants in the
nonanthropomorphism condition (M⫽24.19, SD ⫽13.42),
t(83) ⫽2.27, p⫽.026, d⫽.49, 95% CI [0.06, 0.92]. Thus,
presenting a slot machine in an anthropomorphic way increased
gambling behavior. These findings suggest that presenting a gam-
bling game in an anthropomorphic manner might affect people’s
gambling behavior. Indeed, a few lines of text priming participants
to attribute a humanlike mind to a gambling device were enough to
increase the time participants spent with the device.
Study 2
Study 1 suggested that priming people to perceive a gambling
device anthropomorphically caused an increase in gambling be-
2
To access the online slot machine, see http://www.casinoscompared.co
.uk/freeplay.asp?listing⫽57.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
4RIVA, SACCHI, AND BRAMBILLA
havior. However, in Study 1, the only cost for participants’ gam-
bling was their time, that is, the more they chose to gamble the
longer they spent conducting the experiment. Thus, an important
motivational force was missing in our paradigm, namely, the
possibility of winning or losing actual money on the slot machine.
Indeed, Study 1 did not test whether the presence of money would
change the way in which anthropomorphism affects gambling
behavior. For instance, it is possible that the strong motivational
force of money would reduce or even nullify the effect of anthro-
pomorphism on gambling. To address this important limitation, in
Study 2, we gave participants the chance to bet actual money to
further test our prediction that slot machine anthropomorphization
increases gambling behavior. Participants received an initial
amount of money and were then given the chance to win (or lose)
more money on the slot machine. We expected that people would
gamble more when they were led to perceive the machine anthro-
pomorphically, even when money was introduced as a reward.
Method
Participants and design. Fifty participants
3
(34 female; M
age ⫽23.76 years, SD ⫽4.78) agreed to volunteer to participate
in this study. Similar to Study 1, the experiment used a one-way,
two-level, between-subjects design. The independent variable was
being primed about slot machines anthropomorphically (vs. non-
anthropomorphically). The dependent variable was the number of
spins that the participants made while interacting with an authentic
online slot machine.
Materials and procedure. Presented on a computer screen,
the participants read that the researchers were interested in their
experience of playing with a slot machine. The cover story related
to the evaluation of the graphics features of the slot machine was
not used in Study 2 (or in the following studies). Instead, partic-
ipants read a brief description on the history and the dissemination
of the slot machine. Next, they were instructed that they would
start their game session with 750 points and that this initial score
could increase or decrease every time they spun the reels of the slot
machine. Participants were informed that they had to spin the reels
once at the beginning of the session; after this, they could choose
how long to play for and when to stop. The amount of money
participants won at the end of the experiment corresponded to the
score they got on the slot machine at the time they decided to end
their gambling session. The equivalence between the slot machine
score and money was based on a conversion table that was always
visible to participants. The maximum possible payout was €20 (for
scores above 3,500 points), whereas the minimum possible payout
was zero (for scores below 149 points). All participants started
their game session with 750 points that corresponded on the
conversion table to €5.
After being informed of the correspondence between the scores
on the slot machine and money, participants were randomly as-
signed to read the same descriptions (anthropomorphism condition
vs. nonanthropomorphism condition) of the slot machine used in
Study 1. Participants were then automatically connected with a real
slot machine website (see Figure 1A for an example of an interface
similar to the one that was used in this study) and reminded they
could interact with the slot machine as long as they wanted.
However, this time, when participants decided to quit the game,
they had to notify the experimenter, who could read the final score
before pressing the “Esc” key. Finally, participants were asked to
complete the 15 items assessing their tendency to anthropomor-
phize the slot machine (e.g., “The slot machine has conscious-
ness”; ␣⫽.85; see Appendix for the item descriptions) and to
report their demographic data (i.e., age, sex, and nationality). At
the end of the experiment, each participant received the amount of
money they won, was debriefed, and the purpose of the study was
explained in detail.
Results and Discussion
First, we checked the effectiveness of our manipulation by
computing an independent samples ttest on anthropomorphism
scores, comparing the two experimental groups. The analysis
revealed that participants anthropomorphized the slot machine
more in the anthropomorphism condition (M⫽2.50, SD ⫽.67)
than in the nonanthropomorphism condition (M⫽1.98, SD ⫽
.56), t(48) ⫽2.93, p⫽.004, d⫽.84, 95% CI [0.26, 1.41].
Then, we computed a second ttest comparing the two experi-
mental conditions on the number of reel spins made by partici-
pants. The analysis showed that participants in the anthropomor-
phism condition spun the reels more times (M⫽66.69, SD ⫽
33.47) than the participants in the nonanthropomorphism condition
(M⫽46.54, SD ⫽29.51), t(48) ⫽2.25, p⫽.029, d⫽.64, 95%
CI [0.07, 1.21]. Thus, replicating the findings of Study 1, we found
that even when the task involved monetary reinforcement, describ-
ing a slot machine in an anthropomorphic manner increased gam-
bling behavior.
We further investigated in an exploratory way whether our
manipulation affected not only the number of spins participants
made but also the amount of money (corresponding to the slot
machine scores) that they won (or lost). To do so, we computed
another ttest comparing the two experimental conditions on the
amount of money participants received. The analysis showed no
statistically significant differences between the money lost by
participants primed with the anthropomorphic slot machine (M⫽
3.15 euros, SD ⫽3.25) and those lost by participants who were
primed with the nonanthropomorphic slot machine (M⫽4.38
euros, SD ⫽2.80), t(48) ⫽1.42, p⫽.162, d⫽.40, 95% CI
[⫺0.96, 0.16]. Unsurprisingly, we noted that on average partici-
pants of both groups completed their gaming session with less
money than those given at the beginning of the experimental
session (i.e., €5).
In sum, extending our previous findings, Study 2 showed that
presenting a slot machine in an anthropomorphic manner increased
gambling behavior, even when participants were gambling with
real money. The similarity of the findings in Study 1 and Study 2
suggests that the experimental setting adopted in Study 1 (e.g., slot
machine score visible on the screen without any involvement of
monetary reinforcement) can be considered a good proxy to study
3
At the end of the experiment, participants were asked whether they had
ever played on a slot machine at least once in their life. Of 50 participants,
35 (70%) answered “no,” whereas the remaining 15 (30%) answered “yes”
(i.e., that they had played on a slot machine at least once in their life).
Those who answered “yes” were then asked how often they played on slot
machines. Fourteen participants reported that they played less than once
per year, whereas the remaining participant (N⫽1) reported that s/he
played less than once per month. None reported a higher frequency of slot
machine playing.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
5
HUMANIZING SLOT MACHINES INCREASES GAMBLING
the motivational and emotional factors involved when gambling on
a slot machine.
However, we found no evidence for a link between slot machine
anthropomorphization and greater monetary losses. This remains a
plausible hypothesis, considering that slot machines are pro-
grammed to achieve a certain overall payback percentage, which is
always set to make the slot machine operators win over time
(Pacode.com., 2015; U.K. Gambling Commission, 2013). Thus, in
the long run, making people gamble more on a slot machine should
result in greater monetary losses for the gamblers and greater gains
for the slot machine operators.
Study 3
Study 3 aimed to further replicate the effect of slot machine
anthropomorphization on gambling behavior and explore the me-
diating mechanism driving this effect. Thus, in Study 3, we again
manipulated slot machine anthropomorphism and then measured
people’s gambling behavior. Furthermore, in Study 3 we tested
whether the emotional experience during the game could account
for the relationship between slot machine anthropomorphism and
gambling behavior. More specifically, we expected that people
exposed to an anthropomorphic description of a slot machine
would experience a higher emotional reaction, which in turn would
lead them to gamble longer.
Method
Participants and design. Eighty participants
4
(44 female; M
age ⫽25.65 years, SD ⫽6.24) agreed to volunteer to participate
in this study. The experiment used a one-way, two-level, between-
subjects design. The independent variable was being primed about
slot machines anthropomorphically (vs. nonanthropomorphically).
The dependent variables were the number of spins that the partic-
ipants made while interacting with an online slot machine and the
participants’ self-reported emotional experiences during the game.
Materials and procedure. Participants were randomly as-
signed to read the same descriptions (anthropomorphism condition
vs. non anthropomorphism condition) of the slot machine used in
Studies 1 and 2. Participants were then automatically connected
with a real slot machine website (see Figure 1A for an example of
an interface similar to the one that was used in this study) and were
informed they could interact with the slot machine for as long as
they wanted. When participants decided to quit the game (by
pressing the “Esc” key), they were presented with a series of
questions. First, they were asked to complete 15 items assessing
their tendency to ascribe the slot machine human mental states
such as intention, consciousness, free will, and the capacity to
experience emotions (e.g., “The slot machine has free will”; ␣⫽
.96). Next, participants rated their emotional experience during the
game using three items of positive emotions (i.e., “The gaming
experience was fun”; “The gaming experience was exciting”; “The
gaming experience was stimulating”; ␣⫽.95) and three items of
negative emotions (i.e., “The gaming experience made me angry”;
“The gaming experience made me furious”; “The gaming experi-
ence was very irritating”; ␣⫽.80). Responses were recorded on a
1(totally disagree)to5(totally agree) scale and were averaged to
create an overall index of gambling-machine anthropomorphism,
an overall index of positive emotions, and an overall index of
negative emotions. Finally, participants’ demographic data (i.e.,
age, sex, and nationality) were collected. At the end of the exper-
iment, each participant was debriefed and the purpose of the study
was explained in detail.
Results and Discussion
In order to check the manipulation effectiveness, we computed
an independent samples ttest on anthropomorphism scores, com-
paring the two experimental groups. The analysis revealed that
participants anthropomorphized the slot machine more in the an-
thropomorphism condition (M⫽2.94, SD ⫽.69) than in the
nonanthropomorphism condition (M⫽1.60, SD ⫽.46), t(78) ⫽
10.32, p⬍.001, d⫽2.29, 95% CI [1.72, 2.85].
Then, we computed a second ttest comparing the two experi-
mental conditions on the number of reel spins made by partici-
pants. The analysis showed that participants in the anthropomor-
phism condition spun the reels more times (M⫽28.23, SD ⫽
11.65) than the participants in the nonanthropomorphism condition
(M⫽15.63, SD ⫽8.35), t(78) ⫽5.56, p⬍.001, d⫽1.24, 95%
CI [0.76, 1.72]. Replicating the findings of Studies 1 and 2, we
again found that presenting a slot machine in an anthropomorphic
manner increased gambling behavior.
We then computed a third ttest comparing the two experimental
conditions on the self-reported emotional reaction, considering
both positive and negative emotions. The analysis showed that
participants in the anthropomorphism condition reported higher
levels of both positive (M⫽2.91, SD ⫽.94) and negative
emotions (M⫽1.78, SD ⫽.73) compared with the participants in
the nonanthropomorphism condition (for positive emotions, M⫽
1.76, SD ⫽.91; for negative emotions, M⫽1.23, SD ⫽.32), (for
positive emotions, t[78] ⫽5.57, p⬍.001, d⫽1.24, 95% CI [0.76,
1.72]; for negative emotions, t[78] ⫽4.38, p⬍.001, d⫽.98, 95%
CI [0.51, 1.44]. Thus, imbuing slot machines with human charac-
teristics and mental states led participants of Study 3 to report a
stronger emotional reaction (including both positive and negative
emotions) during the gambling period.
Finally, we conducted mediational analyses to test for a possible
mechanism that might account for the effects of anthropomor-
phism (coded as low anthropomorphism ⫽⫺1, high anthropomor-
phism ⫽1) on gambling behavior. The manipulation of slot
machine anthropomorphism predicted gambling behavior (B⫽
6.30, SE ⫽1.13, p⬍.001). Furthermore, the manipulation of slot
machine anthropomorphism positively predicted both positive
(B⫽0.58, SE ⫽.10, p⬍.001) and negative (B⫽.28, SE ⫽.06,
p⬍.001) emotions. Finally, when both positive and negative
emotions were included in the regression equation, both positive
(B⫽6.93, SE ⫽.99, p⬍.001) and negative (B⫽3.74, SE ⫽
1.62, p⫽.024) emotions predicted the gambling behavior,
whereas the direct effect of the manipulation of slot machine
anthropomorphism on gambling behavior was no longer signifi-
cant (B⫽1.29, SE ⫽1.18, p⫽.28).
4
Participants were asked whether they had ever played on a slot ma-
chine at least once in their life. Out of 80 participants, 69 (86%) reported
“no,” whereas the remaining 11 (14%) reported they had played on a slot
machine at least once in their life. Of the latter group of participants, eight
reported that they played less than once per year, whereas the remaining
participants (n⫽3) reported that they played less than once per month.
None reported a higher frequency of slot machine playing.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
6RIVA, SACCHI, AND BRAMBILLA
To test for mediation, we employed the bootstrapping method
developed by Preacher and Hayes (2008). The analysis provided
support for the mediating role of positive emotions (B⫽3.98,
SE ⫽1.02; 95% CI [2.40, 6.57]) but not of negative emotions (B⫽
1.03, SE ⫽0.75; 95% CI [–.36, 2.54]; 1,000 bootstrap resamples).
This finding suggested that although slot machine anthropomor-
phization increased the experience of both positive and negative
emotions, only the experience of positive emotions (e.g., “The
gaming experience was exciting”) mediated the direct effect of
anthropomorphism on gambling behavior.
In sum, Study 3 further corroborates our main prediction, show-
ing that leading people to perceive a gambling device in an
anthropomorphic manner generally increases gambling behavior.
Furthermore, Study 3 explores a mechanism that could account for
the effect, providing initial evidence of a link between the anthro-
pomorphism of slot machines with participants’ emotional expe-
rience and their actual gambling behavior. Specifically, we found
that slot machine anthropomorphism increases the experience of
positive emotions (e.g., excitement, fun), which in turn leads
people to gamble more.
Study 4
Across three studies, we tested the possibility that mind attri-
bution can affect gambling behavior. We found that when people
are confronted with a slot machine described in humanizing terms
(e.g., “The slot machine can decide whether you will win or lose”),
gambling behavior increases. We also found that this effect held
even when the task involved monetary reinforcement. However,
the studies conducted so far did not provide conclusive evidence
related to the link between slot machine anthropomorphization and
gambling outcomes (e.g., people’s actual winnings). Furthermore,
we found some evidence for a possible link between the self-
reported emotional experience during the game and gambling
behavior; however, this evidence was obtained through an ad hoc
created measure of emotions. Furthermore, this scale included only
high-arousal emotions (e.g., “The gaming experience was excit-
ing”; “The game experience was very irritating”), whereas low-
arousal emotions were not considered. Finally, another limitation
of the studies conducted so far is the low statistical power. To
address these issues, we conducted Study 4.
The aim of Study 4 was to test the predictions of the current
research and its derived assumptions in a properly powered sample
of participants. To do so, we first conducted an a priori power
analysis for sample size estimation (using GPower 3.1; Faul,
Erdfelder, Lang, & Buchner, 2007), with ␣⫽.01 and power ⫽
.80, which suggested that the projected sample size needed to
detect a medium effect size (d⫽.50) was approximately N⫽192
for a between-groups comparison.
Thus, Study 4 tested again the main prediction of the current
study, namely, that anthropomorphizing slot machines would in-
crease gambling behavior. Furthermore, Study 4 tested the as-
sumption that anthropomorphizing slot machines would not only
affect gambling behavior but also gambling outcomes (i.e., the
actual winnings). Specifically, we expected that the anthropomor-
phization of slot machines would make participants gamble more,
which in turn would decrease participants’ winnings at the slot
machine. Finally, using a validated measure of emotions, Study 4
tested whether subcategories of emotions (i.e., positive and nega-
tive emotions, with high or low arousal) could account for the link
between slot machine anthropomorphization and gambling behav-
ior. Again, we expected that the self-reported emotional experi-
ence during the game would account for the effect of anthropo-
morphism on gambling behavior.
Method
Participants and design. Two hundred participants
5
(83 fe-
male; Mage ⫽35.41 years, SD ⫽15.33) participated on a
voluntary basis. Similar to our previous studies, the experiment
used a one-way, two-level, between-subjects design. The indepen-
dent variable was being primed about slot machines anthropomor-
phically (vs. nonanthropomorphically); the dependent variables
were (a) the number of spins that the participants made while
interacting with a slot machine, (b) the points scored by each
participants at the end of the game, (c) the actual winnings of each
participant (based on the scores the obtained at the end of the
game), and (d) the self-reported emotions participants experienced
during the game.
Materials and procedure. Participants were randomly as-
signed to read the same descriptions (anthropomorphism condition
vs. non anthropomorphism condition) of the slot machine used in
our previous studies. Because the online interface of the slot
machine we used in Studies 1 to 3 was no longer available at the
time we conducted Study 4, we decided to program an ad hoc slot
machine interface. The interface represented a slot machine with a
simple graphic, including three reels, a “play” button (that partic-
ipants could use to spin the reels), and a quit icon (see Figure 1B).
Furthermore, by clicking a specific button, participants could
access a card depicting the winning combinations and the points
that each of these combinations was worth. While gambling,
participants saw the points they had scored next to the three reels.
In this study, we gave participants an equivalence of candies rather
than money. Similar to Study 2, the winnings were based on the
actual score participants got at the time they decided to quit
gambling. The equivalence between the slot machine score and the
number of candies was based on a conversion table that was
always kept visible to participants. The maximum payout was 10
candies (for scores above 100 points), whereas the minimum
payout was zero (for scores below 9 points). All participants
started their game session with 50 points that corresponded to five
candies on the conversion table. Participants were told that they
could interact with the slot machine for as long as they wanted and
could quit the game at any time.
After being informed of the correspondence between scores and
the winnings, participants were randomly assigned to read the
same descriptions (anthropomorphism condition vs. nonanthropo-
morphism condition) of the slot machine used in Studies 1 to 3.
The slot machine interface was loaded on a computer screen and
participants could interact with it. When participants pressed the
5
Participants were asked whether they had ever played on a slot ma-
chine at least once in their life. Of 200 participants, 138 (69%) chose the
“no” option, whereas 62 (31%) reported they had played on a slot machine
at least once in their life. Of the latter group of participants, 53 reported that
they played less than once per year; six participants reported that they
played less than once per month; two participants reported they play about
once a week; and one participant reported playing several times a week.
None reported a higher frequency (e.g., every day) of slot machine playing.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
7
HUMANIZING SLOT MACHINES INCREASES GAMBLING
quit button, the ad hoc created slot machine interface generated a
log file including the number of times each participant spun the
reels and the total score earned.
Next, similar to previous studies, participants were asked to
complete the 15 items assessing their tendency to anthropomor-
phize the gambling machine (e.g., “The slot machine decides
everything in the game”; ␣⫽.81; see Appendix for the item
descriptions). Then, participants were asked to report their emo-
tional experience during the game through one of the most
frequently used emotion scales in the psychological literature,
the Positive and Negative Affect Schedule: Expanded Form
(PANAS-X; Watson, & Clark, 1994). Responses were recorded
ona1(not at all)to5(very much) scale. Based on a recent
study showing that the PANAS-X items may be classified along
the dimensions of both valence (positive vs. negative) and
arousal (high vs. low; Wyczesany & Ligeza, 2015), we aver-
aged responses to this scale to create four overall indexes of
high-arousal negative emotions (e.g., angry), low-arousal neg-
ative emotions (e.g., tired), high-arousal positive emotions
(e.g., excited), and low-arousal positive emotions (e.g., calm).
Finally, participants reported their demographic data (i.e., age,
sex, and nationality). At the end of the experiment, each participant
received the number of candies they won and was debriefed.
Results and Discussion
Manipulation check. An independent samples ttest was first
computed on anthropomorphism scores, comparing the two exper-
imental groups. The analysis revealed that participants anthropo-
morphized the slot machine more in the anthropomorphism con-
dition (M⫽2.53, SD ⫽.65) than in the nonanthropomorphism
condition (M⫽2.23, SD ⫽.63), t(198) ⫽3.28, p⬍.001, d⫽.46,
95% CI [0.19, 0.75].
6
Outcome variables. Next, we compared the two experimental
conditions on the number of reel spins participants made. Further
replicating our previous findings, the analysis showed that partic-
ipants in the anthropomorphism condition spun the reels more
times (M⫽42.25, SD ⫽29.09) than the participants in the
nonanthropomorphism condition (M⫽30.08, SD ⫽20.45),
t(198) ⫽3.42, p⬍.001, d⫽.48, 95% CI [0.20, 0.77].
The analysis on the final score participants obtained on the slot
machine showed that participants presented with the anthropomor-
phic slot machine lost more points (M⫽32.55, SD ⫽13.42) than
the participants presented with the nonanthropomorphic slot ma-
chine (M⫽37.40, SD ⫽11.48), t(198) ⫽2.74, p⫽.007,
d⫽⫺0.39, 95% CI [⫺0.67, ⫺0.11]. In line with a typical
algorithm governing a slot machine (Turner & Horbay, 2004), the
average scores of both experimental conditions were lower than
the starting point (i.e., 50 points); however, our findings show that
those primed with the anthropomorphic slot machine went on
playing for longer and ended up losing more slot machine points.
The two experimental conditions were also compared on the
number of candies participants received based on their winnings.
Again, we found that participants presented with the anthropomor-
phic slot machine lost more candies (M⫽2.91, SD ⫽1.28) than
the participants presented with the nonanthropomorphic slot ma-
chine (M⫽3.33, SD ⫽1.05), t(198) ⫽1.42, p⫽.012, d⫽⫺0.36,
95% CI [⫺0.63, ⫺0.08]. On average, participants in both condi-
tions obtained fewer candies than what they started with at the
beginning of the game (i.e., five candies), but this reduction was
stronger for those primed with the anthropomorphic slot machine.
We then computed a series of ttests comparing the two exper-
imental conditions on the four indices of self-reported emotional
reaction considering both valence (ranging from negative to pos-
itive) and arousal (low to high). The analysis showed no significant
differences for the overall indexes of high-arousal negative emo-
tions, high-arousal positive emotions, and low-arousal positive
emotions (all ps⬎.064). By contrast, the analysis showed that
participants presented with the anthropomorphic slot machine re-
ported higher levels of low-arousal negative emotions (e.g., fa-
tigue; M⫽1.74, SD ⫽.59) than the participants presented with the
nonanthropomorphic slot machine (M⫽1.56, SD ⫽.51),
t(198) ⫽⫺2.25, p⫽.025, d⫽0.32, 95% CI [0.05, 0.61].
Mediation analyses. Next, we examined whether different
types of emotions (varying along the dimensions of valence and
arousal; Posner, Russell, & Peterson, 2005) mediated the relationship
between the manipulation of slot machine anthropomorphization
(high vs. low) and gambling behavior (see Table 1 for correlations
between variables).
We used a bootstrapping procedure (Hayes, 2013) estimating
direct and indirect effects with multiple potential mediators. First,
we found that the slot machine manipulation (coded as low an-
thropomorphism ⫽⫺1, high anthropomorphism ⫽1) predicted
the dependent variable (i.e., the number of reel spins made by
participants), B⫽6.04, SE ⫽1.79, p⬍.001. However, slot
machine manipulation did not predict high-arousal negative emo-
tions (B⫽.03, SE ⫽.04, p⫽.422), high-arousal positive
emotions (B⫽.07, SE ⫽.04, p⫽.105), or low-arousal positive
emotions (B⫽.13, SE ⫽.07, p⫽.064). By contrast, we found a
significant association between our manipulation and low-arousal
negative emotions (B⫽.09, SE ⫽.04, p⫽.025).
According to Hayes (2013, Chapter 6; see also Hayes, 2009),
whether an indirect effect is significant is not pertinent to the
pattern of significance or nonsignificance for individual paths in a
mediation model. Thus, we estimated indirect effects for all four
potential mediators, even though some of the paths between the
manipulation and the mediators were not statistically significant.
When the four emotional clusters were included in the regression
equation, high-arousal positive emotions predicted gambling be-
havior (B⫽9.42, SE ⫽3.05, p⫽.002), whereas the direct effect
of the manipulation of slot machine anthropomorphization on
gambling outcomes decreased (B⫽5.46, SE ⫽1.75, p⫽.002).
The analysis provided support for the idea that high-arousal pos-
itive emotions (B⫽.70, SE ⫽.52; 95% CI [.0072, 2.11]) mediated
the relationship between slot manipulation and gambling behav-
6
To test the possibility that our manipulation influenced only the two
components of anthropomorphism (i.e., agency and experience; Gray et al.,
2007; Waytz & Young, 2014), we first created an overall index of agency
(12 items, ␣⫽.801) and an overall index of experience (three items, ␣⫽
.652). Then, we meta-analytically combined the results from the effect
sizes of Studies 1 to 4 (N⫽415) considering the effects of our manipu-
lation of slot machine anthropomorphization on the two components of
anthropomorphism (i.e., agency and experience). The meta-analysis
showed a significant effect on both the weighted combined zscore for
agency ratings (z⫽7.12, p⬍.001), and experience ratings (z⫽3.81, p⫽
.002). Therefore, our manipulation affected both the attribution of agency
and the attribution of experience to the slot machine.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
8RIVA, SACCHI, AND BRAMBILLA
ior.
7
By contrast, high-arousal negative emotions did not predict
gambling behavior (B⫽.79, SE ⫽5.29, p⫽.880), nor did they
play any mediation role (B⫽.02, SE ⫽.27; 95% CI [⫺.36, 1.01]).
Similarly, low-arousal positive emotions did not predict gambling
behavior (B⫽⫺1.10, SE ⫽1.87, p⫽.556), nor did they play any
mediation role (B⫽.15, SE ⫽.32; 95% CI [⫺1.18, .30]). Finally,
low-arousal negative emotions did not predict gambling behavior
(B⫽⫺5.96, SE ⫽4.82, p⫽.22), nor did they play any mediation
role (B⫽.53, SE ⫽.53; 95% CI [⫺2.02, .13]).
Finally, we conducted another mediation analysis to test the
assumption that the gambling behavior (i.e., the number of times
participants spun the slot machine reels) could account for the link
between slot machine anthropomorphization and the gambling
outcomes (i.e., scores earned and number of candies; see Table 1
for correlations between variables). The manipulation of slot ma-
chine anthropomorphism negatively predicted both slot machines
points (B⫽⫺2.42, SE ⫽0.88, p⫽.007) and the number of
candies earned (B⫽⫺0.21, SE ⫽0.08, p⫽.012). Furthermore,
the manipulation of slot machine anthropomorphism predicted
gambling behavior (B⫽6.08, SE ⫽1.78, p⫽.001). Finally, when
gambling behavior was included in the regression equations, gam-
bling behavior predicted gambling scores (B⫽⫺.42, SE ⫽.01,
p⬍.001) and the number of candies earned (B⫽⫺
.04, SE ⫽
.002, p⬍.001), whereas the direct effects of the manipulation of
slot machine anthropomorphism on slot machine scores (B⫽.15,
SE ⫽.47, p⫽.75) and number of candies earned was no longer
significant (B⫽.03, SE ⫽.05, p⫽.57).
The analysis provided support for the mediating role of gam-
bling behavior on both slot machine scores (B⫽⫺2.58, SE ⫽
0.72; 95% CI [⫺4.17, ⫺1.32]) and the number of candies earned
(B⫽⫺.24, SE ⫽.07; 95% CI [⫺.38, ⫺.12]). These findings
suggest that participants primed with the anthropomorphic slot
machine finished with fewer points and lost more candies because
they gambled more.
Summary. In sum, in Study 4 we replicated the main finding
of the current studies in a large sample of participants, namely,
gambling increases when a slot machine is displayed in an anthro-
pomorphic manner. We also showed that the effects of slot ma-
chine anthropomorphization extend to gambling outcomes; partic-
ipants primed with different descriptions of a slot machine
reported different scores when gambling.
Then, unlike Study 3, Study 4 did not provide support for the
hypothesis that priming participants with an anthropomorphic slot
machine per se results in a stronger emotional reaction (except for
low-arousal negative emotions, such as fatigue). However, we
found an indirect effect of high-arousal positive emotions, such
that feeling alert, confident, and excited indirectly mediated the
link between slot machine anthropomorphization and gambling
behavior. Finally, Study 4 showed that gambling behavior ac-
counted for the link between slot machine anthropomorphization
and gambling outcomes, that is, the more participants played, the
less they won.
General Discussion
The present research tested the hypothesis that anthropomor-
phizing gambling devices increases gambling. Our hypothesis was
based on the rationale that people might gamble more when they
think they are engaged by a mind rather than just a mathematical
algorithm. First, we conducted a preliminary study investigating
whether gambling behavior and slot machine anthropomorphiza-
tion might be linked. Accordingly, we found that regular players
anthropomorphized slot machines more than nonregular players.
Specifically, we found that the more people attribute a humanlike
mind to gambling devices, the more frequent they report their
gambling to be. Then, in Study 1, we manipulated slot machine
anthropomorphization and found an increase in participants’ gam-
bling behavior. In Study 2, we replicated the findings of Study 1
and showed that the effect held when the task involved monetary
reinforcement. In Study 3, we explored the possible role of self-
reported emotions on the link between anthropomorphism and
gambling behavior, showing that positive emotions mediated the
relationship between slot machine anthropomorphization and gam-
bling behavior. Finally, in Study 4, we extended the effects of
anthropomorphism from gambling behavior to gambling out-
comes, showing that participants primed with a humanized device
lose more because they play longer. Furthermore, in Study 4, we
found further evidence for a possible mediational role of emotions,
showing that high-arousal positive emotions played a role linking
anthropomorphism and gambling.
The present research substantially extends existing work on the
consequences of humanizing technical devices, providing useful
implications for the construct of anthropomorphism. Few works
have looked at the potential consequences of anthropomorphism,
7
A similar indirect mediational pattern of high-arousal positive emo-
tions emerged on the number of candies participants received based on
their winnings (B⫽⫺0.026, SE ⫽0.02; 95% CI [⫺0.08, ⫺0.0004]), but
not on the final score participants obtained on the slot machine
(B⫽⫺0.29, SE ⫽0.23; 95% CI [⫺.91, 0.01]).
Table 1
Correlations Between Variables (Study 4)
MSD12345678
1. Slot machine manipulation .50 .50 1 .236
ⴱⴱ
⫺.191
ⴱⴱ
⫺.178
ⴱⴱ
.057 .158
ⴱ
.115 .131
2. Gambling behavior 36.16 25.82 1 ⫺.859
ⴱⴱ
⫺.842
ⴱⴱ
⫺.021 ⫺.031 .236
ⴱⴱ
.057
3. Slot machine score 34.97 12.69 1 ⫺.946
ⴱⴱ
.008 .093 ⫺.213
ⴱⴱ
⫺.097
4. Number of candies 3.12 1.18 1 ⫺.009 .108 ⫺.211
ⴱⴱ
⫺.092
5. High-arousal negative emotions 1.47 .54 1 .747
ⴱⴱ
.315
ⴱⴱ
⫺.267
ⴱⴱ
6. Low-arousal negative emotions 1.65 .55 1 .175
ⴱ
⫺.156
ⴱ
7. High-arousal positive emotions 1.90 .65 1 .234
ⴱⴱ
8. Low-arousal positive emotions 3.12 1.04 1
ⴱ
p⬍.05.
ⴱⴱ
p⬍.01.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
9
HUMANIZING SLOT MACHINES INCREASES GAMBLING
and, when they do, they are mostly limited to assessment of
attitudes (e.g., Butterfield, Hill, & Lord, 2012; Chandler, Griffin,
& Sorensen, 2008; Chandler & Schwarz, 2010; Sacchi et al., 2013;
Tam, Lee, & Chao, 2013; Zagefka, Noor, Brown, de Moura, &
Hopthrow, 2011). Furthermore, the majority of studies have con-
sidered the attribution of humanlike emotions the target of anthro-
pomorphism (e.g., Norenzayan, Hansen, & Cady, 2008; Sacchi et
al., 2013), whereas the current research explored the agent’s emo-
tional response as a function of machine anthropomorphization
and its role as a mediating mechanism.
Many scholars have argued that people have a strong tendency
to attribute mental states to nonhuman agents. Some theorists have
argued that theory of mind, that is, the attribution of mental states
to oneself and others, is such a pervasive force such that people
cannot help but perceive mental states in objects and supernatural
religious agents (Bering, 2002). In our studies, we consistently
found that priming participants with an anthropomorphic slot
machine led them to perceive it as more humanlike (compared
with participants primed with a nonanthropomorphic machine).
However, the manipulation check means provided by the nonan-
thropomorphic slot machine group were consistently higher than
zero, thus suggesting that the anthropomorphization of an object
occurs spontaneously, but it can be increased by environmental
cues. Overall, our findings indicate that even trivial objects such as
slot machines can—in the eyes of the perceiver— become moral
agents who are conscious, act according to their own intentions,
and experience emotions. Furthermore, they suggest that gambling
behavior can be manipulated through targeted communication to
decrease the attribution of human mind features to gambling
devices. In our studies, the description of slot machines in the
anthropomorphism conditions was based on some key aspects that
the existing literature on anthropomorphism has considered (e.g.,
the anthropomorphized agent has consciousness, free will, and it is
considered responsible for its actions). By contrast, the specific
description of the slot machine shown in the nonanthropomor-
phism condition was based on descriptions of slot machines and
how they function (Turner & Horbay, 2004).
Our research has shown that the consequences of slot machine
anthropomorphization extend from gambling behavior to gambling
outcomes. Indeed, we found that people exposed to a humanized
slot machine lost more points compared with people exposed to a
nonhumanized slot machine. Furthermore, anthropomorphization
resulted in a greater loss of slot machine payouts (Study 4). In
actuality, we found that both groups ended up with less than what
they had started with, although anthropomorphism resulted in a
greater loss of scores and relative payouts. As we noted, this is not
surprising, considering that the payback percentage with which
slot machines are set at varies usually between 70% and 90%, and
thus it varies the degree to which people lose money by playing
more at the slots. In any case, from a logical standpoint, on
average, any payback percentage under 100% implies that people
who play more at the slot machine will lose more money. This is
exactly what we found when we tested this hypothesis with ade-
quate statistical power (see Study 4): People primed with a hu-
manized slot machine lost more slot machine payouts because they
went on playing for longer.
How does anthropomorphism increase gambling? Our research
was able to identify a possible mechanism that can account for the
anthropomorphism/gambling-behavior link. Indeed, emotional
arousal is known to increase attention (Mather, 2007). In both
Study 3 and Study 4, we found that high-arousal positive emotions
(e.g., feeling excited, determined, and confident) played a role in
the effect of anthropomorphism on gambling behavior. Thus, in
our experiments, when they believed they were playing against a
mind (cf. a mathematical algorithm), people might have gambled
more because of the more intense emotional experiences they
underwent while gambling. In other words, our participants may
have felt more engaged during the game and thus gambled longer.
This finding could be in line with previous studies suggesting that
the emotional quality of a person’s experience with a nonhuman
entity may approximate that of their experience with other human
beings (Aaker et al., 2004; Schultz et al., 1989; Wang et al., 2007).
However, whereas in Study 3 we found a direct effect of the
anthropomorphism manipulation on both positive and negative
emotions, in Study 4, we did not replicate this finding (except for
the low-arousal negative emotion cluster; e.g., fatigue). It is pos-
sible that the different measures we used in these studies (also
considering the different number of items that characterized the
different scales) account for these inconsistencies. Therefore, fu-
ture studies should further explore the role of emotions in account-
ing for the link between anthropomorphism and gambling behav-
ior.
From an applied standpoint, our research suggests some avenues
for communications and interventions. We showed that it is pos-
sible to affect gambling behavior (and the related gambling out-
comes) by simply framing gambling devices with different terms.
Thus, when the aim is reducing gambling, we recommend framing
slot machines in a mechanistic way. This implies conveying the
message of what slot machines are: physical objects controlled by
mathematical algorithms, which are programmed to deliver a cer-
tain amount of wins and losses. By contrast, any form of human-
ization (either in the form of words or images) of the slot machine
should be avoided. Accordingly, our studies showed that human-
izing slot machines led people to gamble and lose more, perhaps
because of the psychological state (i.e., high-arousal positive emo-
tions) that accounted for the effect of anthropomorphization on
gambling. Again, when the aim is to reduce gambling, anthropo-
morphization of the gambling device should be prevented in order
to restrain gambling behavior and the economic losses that derive
from it.
There are some limitations to the present research. First, in our
studies, we used a verbal description of slot machine to manipulate
slot machine anthropomorphization. Future studies should test
other ways to modulated people’s perception of these gambling
devices, perhaps through the use of anthropomorphic cues (e.g.,
eyes) displayed on the interface. Moreover, future studies should
adopt an implicit measure of emotions or physiological recording
during the gambling experience to further test for the role of
emotions in accounting for the link between humanizing machines
and gambling behavior.
Second, future research should also consider other potential
mechanisms that are also likely to play a role in the
anthropomorphism– behavior link. As we have already noted, re-
cent work showed that nonhuman entities are more likable when
they resemble concepts with which people have familiarity (i.e.,
humans; see Epley et al., 2007, 2008; Gray et al., 2007; Waytz,
Morewedge, et al., 2010). This reason could account for the effect
of anthropomorphism on gambling behavior. Furthermore, previ-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
10 RIVA, SACCHI, AND BRAMBILLA
ous research has suggested that anthropomorphism satisfies effec-
tance motivation; that is, anthropomorphizing a target makes it
appear more predictable and understandable (Epley et al., 2007;
see also Waytz, Morewedge, et al., 2010). Accordingly, the per-
ceived increased sense of mastery over the machine’s behavior,
that is, the illusionary control related to the attribution of a hu-
manlike mind to it might lead players to gamble more frequently.
In addition to increasing a sense of efficacy, anthropomorphism
can also increase a sense of connection with nonhuman agents.
Magical thinking, expectations of winning, feelings of competi-
tiveness, and perceptions of randomness might also contribute to
the effect. Thus, future research should further investigate the role
of these potential cognitive and affective mechanisms in account-
ing for the link between slot machine anthropomorphization and
gambling behavior.
Finally, future research should also consider the possible individual
differences that can play a role in the anthropomorphization–
gambling link. For instance, it could be tested whether people with an
external locus of control tend to respond more strongly to a manipu-
lation of slot machine anthropomorphization. In this sense, the ques-
tions raised here are the beginning of a program of research investi-
gating machine anthropomorphization and its impact on human
behavior.
Conclusions
Slot machines are designed to induce people to play more and to
spend as much money as possible (Griffiths, 1994, 2004). Our
investigation demonstrates one of these—that anthropomorphism
represents a psychological means to increase a gambler’s time on
the machine or, adopting an industry jargon, to encourage “playing
to extinction”—that is, playing until a gambler’s money is gone. In
the eyes of the gaming industry, the psychological process of
gambling-machine anthropomorphism is advantageous; however,
this may come at great expense for gamblers’ (and their families’)
economic resources and psychological well-being. Therefore, the
present findings may also have important implications for social
life. As our data suggest, one of the simplest ways to reduce
gambling behavior might be to convey the message of what a slot
machine actually is: not a person, not a mind, just a machine.
References
Aaker, J., Fournier, S., & Brasel, S. A. (2004). When good brands do bad.
Journal of Consumer Research, 31, 1–16. http://dx.doi.org/10.1086/
383419
Bartneck, C., Kuli´
c, D., Croft, E., & Zoghbi, S. (2009). Measurement
instruments for the anthropomorphism, animacy, likeability, perceived
intelligence, and perceived safety of robots. International Journal of
Social Robotics, 1, 71– 81. http://dx.doi.org/10.1007/s12369-008-0001-3
Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for
interpersonal attachments as a fundamental human motivation. Psycho-
logical Bulletin, 117, 497–529. http://dx.doi.org/10.1037/0033-2909.117
.3.497
Bering, J. M. (2002). The existential theory of mind. Review of General
Psychology, 6, 3–24. http://dx.doi.org/10.1037/1089-2680.6.1.3
Butterfield, M. E., Hill, S. E., & Lord, C. G. (2012). Mangy mutt or furry
friend? Anthropomorphism promotes animal welfare. Journal of Exper-
imental Social Psychology, 48, 957–960. http://dx.doi.org/10.1016/j.jesp
.2012.02.010
Chandler, J., Griffin, T. M., & Sorensen, N. (2008). In the “I” of the storm:
Shared initials increase disaster donations. Judgment and Decision Mak-
ing, 3, 404 – 410.
Chandler, J., & Schwarz, N. (2010). Use does not wear ragged the fabric
of friendship: Thinking of objects as alive makes people less willing to
replace them. Journal of Consumer Psychology, 20, 138 –145. http://dx
.doi.org/10.1016/j.jcps.2009.12.008
Chartrand, T. L., Fitzsimons, G. M., & Fitzsimons, G. J. (2008). Automatic
effects of anthropomorphized objects on behavior. Social Cognition, 26,
198 –209. http://dx.doi.org/10.1521/soco.2008.26.2.198
Chen, S. C., Shoemaker, S., & Zemke, D. M. V. (2013). Segmenting slot
machine players: A factor-cluster analysis. International Journal of
Contemporary Hospitality Management, 25, 23– 48. http://dx.doi.org/10
.1108/09596111311290200
Delfabbro, P. H., & Winefield, A. H. (2000). Predictors of irrational
thinking in regular slot machine gamblers. The Journal of Psychology,
134, 117–128. http://dx.doi.org/10.1080/00223980009600854
Epley, N., Waytz, A., Akalis, S., & Cacioppo, J. T. (2008). When we need
a human: Motivational determinants of anthropomorphism. Social Cog-
nition, 26, 143–155. http://dx.doi.org/10.1521/soco.2008.26.2.143
Epley, N., Waytz, A., & Cacioppo, J. T. (2007). On seeing human: A
three-factor theory of anthropomorphism. Psychological Review, 114,
864 – 886. http://dx.doi.org/10.1037/0033-295X.114.4.864
Eyssel, F., & Kuchenbrandt, D. (2012). Social categorization of social
robots: Anthropomorphism as a function of robot group membership.
British Journal of Social Psychology, 51, 724 –731. http://dx.doi.org/10
.1111/j.2044-8309.2011.02082.x
Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G
ⴱ
Power 3: A
flexible statistical power analysis program for the social, behavioral, and
biomedical sciences. Behavior Research Methods, 39, 175–191.
Gardner, W. L., & Knowles, M. L. (2008). Love makes you real: Favorite
television characters are perceived as “real” in a social facilitation
paradigm. Social Cognition, 26, 156 –168. http://dx.doi.org/10.1521/
soco.2008.26.2.156
Gray, H. M., Gray, K., & Wegner, D. M. (2007). Dimensions of mind
perception. Science, 315, 619. http://dx.doi.org/10.1126/science
.1134475
Griffiths, M. (1995). The role of subjective mood states in the maintenance
of fruit machine gambling behaviour. Journal of Gambling Studies, 11,
123–135. http://dx.doi.org/10.1007/BF02107111
Griffiths, M. (2004). Betting your life on it. British Medical Journal, 329,
1055–1056. http://dx.doi.org/10.1136/bmj.329.7474.1055
Griffiths, M. D. (1994). The role of cognitive bias and skill in fruit machine
gambling. British Journal of Psychology, 85, 351–369. http://dx.doi.org/
10.1111/j.2044-8295.1994.tb02529.x
Hart, P. M., Jones, S. R., & Royne, M. B. (2013). The human lens: How
anthropomorphic reasoning varies by product complexity and enhances
personal value. Journal of Marketing Management, 29, 105–121. http://
dx.doi.org/10.1080/0267257X.2012.759993
Hayes, A. F. (2009). Beyond Baron and Kenny: Statistical mediation
analysis in the new millennium. Communication Monographs, 76, 408 –
420. http://dx.doi.org/10.1080/03637750903310360
Hayes, A. F. (2013). Introduction to mediation, moderation, and condi-
tional process analysis: A regression-based approach. New York, NY:
Guilford Press.
Kim, S., & McGill, A. L. (2011). Gaming with Mr. Slot or gaming the slot
machine? Power, anthropomorphism, and risk perception. Journal of
Consumer Research, 38, 94 –107. http://dx.doi.org/10.1086/658148
Kozak, M. N., Marsh, A. A., & Wegner, D. M. (2006). What do I think
you’re doing? Action identification and mind attribution. Journal of
Personality and Social Psychology, 90, 543–555. http://dx.doi.org/10
.1037/0022-3514.90.4.543
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
11
HUMANIZING SLOT MACHINES INCREASES GAMBLING
Mather, M. (2007). Emotional arousal and memory binding: An object-
based framework. Perspectives on Psychological Science, 2, 33–52.
http://dx.doi.org/10.1111/j.1745-6916.2007.00028.x
Morewedge, C. K., & Clear, M. E. (2008). Anthropomorphic God concepts
engender moral judgment. Social Cognition, 26, 182–189. http://dx.doi
.org/10.1521/soco.2008.26.2.182
Nass, C., & Moon, Y. (2000). Machines and mindlessness: Social re-
sponses to computers. Journal of Social Issues, 56, 81–103. http://dx
.doi.org/10.1111/0022-4537.00153
Nass, C., Moon, Y., & Green, N. (1997). Are machines gender neutral?
Gender-stereotypic responses to computers with voices. Journal of Ap-
plied Social Psychology, 27, 864 – 876. http://dx.doi.org/10.1111/j.1559-
1816.1997.tb00275.x
Norenzayan, A., Hansen, I. G., & Cady, J. (2008). An angry volcano?
Reminders of death and anthropomorphizing nature. Social Cognition,
26, 190 –197. http://dx.doi.org/10.1521/soco.2008.26.2.190
Pacode.com. (2015). 058 Pa. Code § 461a.7. Slot machine minimum
design standards. Retrieved August 16, 2015, from http://www.pacode
.com/secure/data/058/chapter461a/s461a.7.html
Posner, J., Russell, J. A., & Peterson, B. S. (2005). The circumplex model
of affect: An integrative approach to affective neuroscience, cognitive
development, and psychopathology. Development and Psychopathology,
17, 715–734. http://dx.doi.org/10.1017/S0954579405050340
Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling
strategies for assessing and comparing indirect effects in multiple me-
diator models. Behavior Research Methods, 40, 879 – 891. http://dx.doi
.org/10.3758/BRM.40.3.879
Sacchi, S., Riva, P., & Brambilla, M. (2013). When Mother Earth rises up:
Anthropomorphizing nature reduces support for natural disaster victims.
Social Psychology, 44, 271–277. http://dx.doi.org/10.1027/1864-9335/
a000112
Schultz, S. E., Kleine, R. E., & Kernan, J. B. (1989). “These are a few of
my favorite things”: Toward an explication of attachment as a consumer
behavior construct. Advances in Consumer Research, 16, 359 –366.
Smith, G. J., & Wynne, H. J. (2004). VLT gambling in Alberta: A
preliminary analysis. Alberta Gambling Research Institute. Available
at http://www.abgaminginstitute.ualberta.ca/documents/research/
VLT_Gambling_Alberta.pdf
Tam, K. P., Lee, S. L., & Chao, M. M. (2013). Saving Mr. Nature:
Anthropomorphism enhances connectedness to and protectiveness to-
ward nature. Journal of Experimental Social Psychology, 49, 514 –521.
http://dx.doi.org/10.1016/j.jesp.2013.02.001
Turner, N., & Horbay, R. (2004). How do slot machines and other elec-
tronic gambling machines actually work? Journal of Gambling Issues,
11, 1– 41.
U.K. Gambling Commission. (2013, May 3). Gaming machine (fruit
machine, slot machine) minimum percentage payout. Retrieved from
http://www.gamblingcommission.gov.uk/gf-faqs/gaming_machines/
how_much_should_a_gaming_machi.aspx
Walker, M. B. (1992). Irrational thinking among slot machine players.
Journal of Gambling Studies, 8, 245–261. http://dx.doi.org/10.1007/
BF01014652
Wang, L. C., Baker, J., Wagner, J. A., & Wakefield, K. (2007). Can a retail
website be social? Journal of Marketing, 71, 143–157. http://dx.doi.org/
10.1509/jmkg.71.3.143
Watson, D., & Clark, L. A. (1994). The PANAS-X: Manual for the positive
and negative affect schedule-expanded form. Unpublished manuscript,
University of Iowa.
Waytz, A., Cacioppo, J., & Epley, N. (2010). Who sees human? The
stability and importance of individual differences in anthropomorphism.
Perspectives on Psychological Science, 5, 219 –232. http://dx.doi.org/10
.1177/1745691610369336
Waytz, A., Morewedge, C. K., Epley, N., Monteleone, G., Gao, J. H., &
Cacioppo, J. T. (2010). Making sense by making sentient: Effectance
motivation increases anthropomorphism. Journal of Personality and
Social Psychology, 99, 410 – 435. http://dx.doi.org/10.1037/a0020240
Waytz, A., & Young, L. (2014). Two motivations for two dimensions of
mind. Journal of Experimental Social Psychology, 55, 278 –283. http://
dx.doi.org/10.1016/j.jesp.2014.08.001
Winnicott, D. W. (1953). Transitional objects and transitional phenomena;
a study of the first not-me possession. The International Journal of
Psychoanalysis, 34, 89 –97.
Wyczesany, M., & Ligeza, T. S. (2015). Towards a constructionist ap-
proach to emotions: Verification of the three-dimensional model of
affect with EEG-independent component analysis. Experimental Brain
Research, 233, 723–733. http://dx.doi.org/10.1007/s00221-014-4149-9
Zagefka, H., Noor, M., Brown, R., de Moura, G. R., & Hopthrow, T.
(2011). Donating to disaster victims: Responses to natural and humanly
caused events. European Journal of Social Psychology, 41, 353–363.
http://dx.doi.org/10.1002/ejsp.781
(Appendix follows)
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
12 RIVA, SACCHI, AND BRAMBILLA
Appendix
Items on the Anthropomorphization Scale
1. The slot machine acts according to its own intentions.
2. The slot machine has consciousness.
3. The slot machine has free will.
4. The slot machine perceives stimuli.
5. The slot machine experiences emotions.
6. The slot machine decides everything in the game.
7. The slot machine punishes me whenever she wants.
8. The slot machine rewards me whenever she wants.
9. Whenever I play, the slot machine chooses what hap-
pens.
10. The slot machine can be kind to me.
11. The slot machine can be mean to me.
12. The slot machine can act maliciously against me.
13. The slot machine can have mercy on me.
14. The slot machine rewards whoever she wants.
15. Sometimes, the slot machine persecutes certain people.
Received July 15, 2014
Revision received July 21, 2015
Accepted July 21, 2015 䡲
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
13
HUMANIZING SLOT MACHINES INCREASES GAMBLING
