ArticlePDF Available

Abstract

Experiments 1 and 2 examined whether winning versus losing led to reckless betting for real prize money. Experiment 2 also assessed whether positive or negative emotions were linked to such reckless betting. College students were randomly assigned to experience primarily either wins or losses during the rigged first round of a computerized card tournament that had 2 independent rounds. For the second round, participants' chip totals were reset and cards were dealt randomly. In Experiment 1 (N=107), participants in the Initial-Winning, as compared with the Initial-Losing, condition bet more recklessly (i.e., bet too many chips when a loss was likely). Experiment 2 (N=72) again showed that Initial-Winning participants bet significantly more recklessly than did Initial-Losing participants. It also revealed that positive affect was significantly positively correlated with such reckless betting. These findings have implications for understanding how college students, those at an age when they are especially vulnerable to problem gambling, can come to lose more money than they can afford. Initially winning and positive affect when gambling could be risk factors.
Winning and Positive Affect Can Lead to Reckless Gambling
Lori F. Cummins, Michael R. Nadorff, and Anita E. Kelly
University of Notre Dame
Experiments 1 and 2 examined whether winning versus losing led to reckless betting for real prize
money. Experiment 2 also assessed whether positive or negative emotions were linked to such reckless
betting. College students were randomly assigned to experience primarily either wins or losses during the
rigged first round of a computerized card tournament that had 2 independent rounds. For the second
round, participants’ chip totals were reset and cards were dealt randomly. In Experiment 1 (N107),
participants in the Initial-Winning, as compared with the Initial-Losing, condition bet more recklessly
(i.e., bet too many chips when a loss was likely). Experiment 2 (N72) again showed that Initial-
Winning participants bet significantly more recklessly than did Initial-Losing participants. It also
revealed that positive affect was significantly positively correlated with such reckless betting. These
findings have implications for understanding how college students, those at an age when they are
especially vulnerable to problem gambling, can come to lose more money than they can afford. Initially
winning and positive affect when gambling could be risk factors.
Keywords: gambling, winning versus losing, risk factors, positive affect
In the past 30 years, more and more people have become willing
to risk their money, possessions, home life, and future in hopes that
the cards will turn out in their favor. According to the most recent
report of the National Gambling Impact Study Commission, be-
tween 1975 and 1999, the proportion of individuals in the United
States who reported having gambled at least once in their lifetime
increased from 68 to 86 percent (University of Chicago, 1999).
The advent of online casinos has made people particularly vulner-
able to the hazards of gambling. In fact, the proximity of a casino
to one’s home strongly predicts the likelihood of developing
problem gambling (University of Chicago, 1999), which is defined
broadly as gambling that causes harm to self or others. A critical
element of problem gambling occurs when one begins to lose and
then continues to make bets to recover previous losses (Campbell-
Meiklejohn, Woolrich, Passingham, & Rogers, 2008).
Young adults aged 18 to 24 years are particularly vulnerable to
problem gambling (Gerstein et al., 1999; Productivity Commis-
sion, 1999; Volberg, Abbott, Ronnberg, & Munck, 2001). A recent
meta-analysis of 19 studies of college students in North America
revealed an alarmingly high lifetime rate of problem gambling of
16.4% among college students (Shaffer & Hall, 2001). As such, it
is important to predict and understand how intelligent, educated
young adults can come to lose more money than they can afford,
so that interventions ultimately can be developed to combat this
key element of problem gambling.
Williams and Connolly (2006) did attempt an intervention for
problem gambling in one recent study. They educated a sample of
college students on the probabilities associated with gambling in
an attempt to reduce the students’ subsequent problem gambling
behavior. The educational intervention did lead students to im-
prove their ability to calculate odds of winning and increased their
resistance to gambling fallacies six months after the intervention.
However, students receiving the intervention, as compared with
those who did not receive it, showed no difference in self-reported
time spent gambling, likelihood of being a problem gambler, or
amount of money spent gambling. The researchers suggested that
mathematically-based interventions are insufficient to produce be-
havioral changes in gambling.
Indeed, emotions are likely to play a large role in problem
gambling, particularly in losing more money than one can afford
through reckless betting aimed at chasing previous losses. But
which is more likely to lead to reckless betting—initially winning
or initially losing, and feeling good or feeling bad while gambling?
The purpose of the present pair of experiments was to provide
answers to these questions. Currently, there is no clear consensus
in the literature on whether prior wins or prior losses lead to riskier
betting behavior. Some studies suggest that individuals are more
willing to take risks following a period of success (see for example
Thaler & Johnson, 1990), but others have found the opposite to be
true (e.g., Leopard, 1978).
In a study by Thaler and Johnson (1990), participants were
presented with a list of hypothetical statements in the form “you
have won/lost X, now choose between gamble A and sure outcome
B” (Thaler & Johnson, 1990, p. 652). In each statement, gamble A
was a risky option with greater potential pay-off, and sure-outcome
B was a riskless option with a smaller pay-off. The results revealed
Lori F. Cummins, Michael R. Nadorff, and Anita E. Kelly, Department
of Psychology, University of Notre Dame.
Michael Nadorff is now at the Department of Psychology, West Virginia
University, Morgantown, West Virginia.
This research was made possible in part by funding support from the
Undergraduate Research Opportunity Program, Institute for Scholarship in
the Liberal Studies, College of Arts and Letters, and the University of
Notre Dame. The two experiments were conducted toward fulfillment of
the requirements of honors theses under the direction of Anita E. Kelly.
Correspondence should be addressed to Anita E. Kelly, Department of
Psychology, University of Notre Dame, Notre Dame, IN 46556. E-mail:
Kelly.79@nd.edu
Psychology of Addictive Behaviors © 2009 American Psychological Association
2009, Vol. 23, No. 2, 287–294 0893-164X/09/$12.00 DOI: 10.1037/a0014783
287
that following a win, as compared with a loss, participants were
more likely to choose the risky option.
In contrast, Leopard (1978) found that following a loss, indi-
viduals became more risk-taking. In her study, Leopard provided
participants with $10 at the beginning of each of four gambling
tasks. The amount of cash participants had at the end of each
gambling task was combined to determine an overall payout,
which was awarded at the conclusion of the last gambling task.
Results revealed that 67% of the time the participants became
more risk-taking when they were losing rather than when they
were winning.
However, there are some noteworthy limitations of these stud-
ies. Thaler and Johnson (1990) used hypothetical scenarios rather
than an actual gambling task. This is problematic because individ-
uals’ choices in a hypothetical scenario and their actual behaviors
might not converge. Moreover, the results of Leopard’s (1978)
study might be confounded by the payoff scheme she used. Be-
cause the results of all four of her gambling tasks were combined
to determine an overall payoff, it would have been beneficial for
participants who were losing to place riskier bets. By doing so,
these participants would have a better chance of making up for
their losses and increasing their ultimate payoff. Thus, additional
studies are needed to assess risky betting after actual gambling
tasks where prior losses cannot directly influence the participants’
final payoff.
In Experiments 1 and 2, we tested whether college students bet
more recklessly following an experimentally induced winning or
losing streak in a computerized game of cards. Reckless betting
was operationalized as betting too much on hands that were likely
to lose. We studied samples of college students because they are
particularly vulnerable to problem gambling and to losing more
money than they can afford. Moreover, in Experiment 2, we
examined the relationship between affect and reckless betting. In
both experiments, college students played in a two-round card
tournament, where the first round of the tournament was manipu-
lated so that participants experienced either a majority of wins or
a majority of losses. In order to make the experience of winning or
losing salient to participants, they had the opportunity to win one
of three large, actual cash prizes. As such, these studies improved
upon Thaler and Johnson’s study by measuring participants’ risky
betting in an actual gambling task. They also improved upon
Leopard’s study because instead of combining the rounds of the
tournament, we reset participants’ chip totals at the beginning of
the second round and told them that their better score from the two
rounds would be used to determine the winners of the cash prizes.
Having two independent rounds would not put participants in the
position of having to make riskier bets in the second round to
compensate for any earlier losses. Thus, these experiments filled a
gap in the gambling literature by providing a means of testing (a)
more definitively whether winning versus losing causes reckless
betting and (b) the role of affect in reckless betting.
Experiment 1
Based on Thaler and Johnson’s (1990) findings from hypothet-
ical gambling scenarios, we predicted that individuals who expe-
rienced a majority of wins during the first round of the tournament
would perform worse in the second round of the tournament than
would those who experienced a majority of losses during the first
round of the tournament. In particular, we predicted that partici-
pants would bet too much on hands that they were likely to lose
(i.e., would bet recklessly) because of the over-confidence they
would develop from winning so many hands in the first round.
Method
Overview and design. Participants were randomly assigned to
either the Initial-Winning condition or the Initial-Losing condition
for the first round of the tournament. Participants in the Initial-
Winning condition experienced a win on 80% (24 of 30) of hands,
whereas participants in the Initial-Losing condition experienced a
loss on 80% (24 of 30) of hands. Chip totals were reset, and cards
were dealt at random for the second round of the tournament. The
key dependent measures were how skillfully participants bet dur-
ing the second round when they were likely to lose a given hand
(i.e., betting a minimum number of chips) or win a given hand (i.e.,
betting a maximum number of chips).
Participants. Participants were 108 students at a private uni-
versity in the Midwest. They were selected for this study either
because they responded to an advertisement posted online and in
the undergraduate residence halls or because they signed up to
participate through the psychology department. All participants
had the opportunity to win a $250, $100, or $50 cash prize, and
those who signed up through the psychology department also
received extra credit for a psychology class. One participant was
dropped from analyses because he did not complete the materials
properly. Therefore, analyses were conducted using data from 107
participants. Eighty-eight participants identified themselves as
Caucasian (82.2%), 3 as African American (2.8%), 6 as Hispanic
or Latino (5.6%), 7 as Asian (6.5%), and 3 as “other” (2.8%). They
ranged in age from 18 to 32 years, with a mean age of 19.2 (SD
1.72) and the majority of were male (73.8%).
Measures. To get a sense of participants’ experience playing
cards, 2 items were developed. The first item was “How skilled are
you at card games?” It was scored using a 7-point Likert scale
from 1 (not at all skilled)to7(extremely skilled). The second item
was “How much card playing experience do you have?” and was
scored on a 7-point Likert scale from 1 (none)to7(a great deal).
Participants’ betting behavior was assessed using only the bets
they placed during the second round of the tournament. For each
hand, the Acey-Deucey software stored in a data file the number of
chips bet, the percentage of maximum possible bet placed, the
probability of winning the hand, and whether the hand resulted in
a win or a loss of chips for each of the 30 hands. (Note that there
was a computer glitch that did not allow us to use the data from the
last 4 of the 30 hands.) Skill of betting was measured by our 2 key
dependent variables, Expected Chips Won and Expected Chips
Lost, which removed the chance factor behind the randomly dealt
cards. These variables broke down the bets placed on hands where
a win versus a loss would be expected. To explain how these
variables were computed, we first must provide details on how the
card game is played. The game is called Acey-Deucey, and for
each hand, two outer cards are dealt to the player face-up, and the
middle card is displayed face-down. The participant must place a
bet on the likelihood that the middle card will have a value
between the two outer cards. An example of a hand likely to win
is two outer cards of a Deuce (i.e.,a2ofanysuit) and a King. The
bettor should place a maximum bet because there is an 80% chance
288 CUMMINS, NARDORFF, AND KELLY
that the middle card will have a value between a 2 and a King. That
is, only 10 (i.e., 3 Deuces, 3 Kings, and 4 Aces) of the possible 50
cards remaining in a 52-card deck will cause a player to lose on
that hand. Expected Chips Won was how much participants bet
when the probability of winning was greater than 50%. It was
calculated using the formula
h1
26 Bh
PhExpectedChipsWon,
where his the hand number, Bis the actual bet placed, and Pis the
probability of winning. This formula was used only for those hands
where the probability of winning was greater than or equal to 50%.
A higher number is good, meaning that the participant bet more on
hands that were likely to win.
Expected Chips Lost was how much the participant bet on
hands in which the probability of winning was less than 50%.
An example of a hand likely to lose is two outer cards ofa6and
an 8 because there is only an 8% chance (i.e., only 4 cards out
of the remaining 50) that the middle card will have a value
between 6 and 8. In this instance, the bettor should place the
minimum bet. Expected Chips Lost was calculated using
the formula
h1
26 Bh
(1Ph)ExpectedChipsLost, where his
the hand number, Bis the actual bet placed, and Pis the
probability of winning. This formula was used only for those
hands where the probability of winning was less than 50%. A
higher number means more reckless, poorer betting.
To illustrate the way the formulas above were employed, con-
sider the following example of calculating Expected Chips Won
(which was computed only on the hands that had greater than or
equal to a 50% chance of winning). A participant plays three hands
of the tournament. On the first hand, the probability of winning is
80% and the participant bets 100 chips. On the second hand, the
probability of winning is 72% and the participant bets 100 chips.
On the third hand, the probability of winning is 88% and the
participant again bets 100 chips. The expected winnings for each
of the three hands are 80, 72, and 88 chips, respectively. Thus, the
value of Expected Chips Won for this participant at the conclusion
of the three hands played equals the sum of the expected winnings
on each hand, or 240 chips. When calculating Expected Chips
Lost, the same procedure was used, except that the computations
were performed only on hands that had less than a 50% chance of
winning and that the number of chips bet on each hand was
multiplied by the probability of losing, or 1 minus the probability
of winning the hand.
Participants also were asked to indicate their age, sex, and race.
Manipulation checks were included at the conclusion of the first
round of card play to ensure that participants perceived themselves
as experiencing a majority of wins or losses, depending on their
assigned condition. These checks included the following 2 items:
(a) “On what percentage of hands played did you experience a
win?” and (b)“On what percentage of hands played did you expe-
rience a loss?” For these items, participants entered a number in
the space provided. A filler stream-of-consciousness writing task
was included to between the two rounds of the tournament to give
participants a mental break between the rounds.
Apparatus. The apparatus for the Acey-Deucey game was an
IBM-compatible computer equipped with a keyboard and mouse.
This computer utilized the Windows XP Home Edition operating
system. The Acey-Deucey software was produced by DeMent
Contract Software, Inc. using Microsoft Visual C⫹⫹ 5.0 with
Microsoft Foundation Class Library. Upon starting the program,
the experimenter is prompted to enter the subject number and
select the condition number. After this, the program proceeds to a
screen of text instructions for the participant, and the participant
presses a button marked “Begin” to start playing the game.
Procedure. Participants came to the laboratory at their as-
signed times and were randomly assigned to either the Initial-
Winning or Initial-Losing condition upon arrival. All participants
signed a consent form and were told that their questionnaires
would be placed in a locked box to which the experimenter would
not have access. To provide a rationale for the study, they were
told that its purpose was to examine the personality characteristics
of different types of card players. Participants responded to the
card playing attitudes and behaviors items, and then the experi-
menter read the instructions for the Acey-Deucey game:
Now you will participate in the actual tournament. For the tourna-
ment, we will be using a computerized version of a card game called
Acey-Deucey. When you press the [Begin] button, the computer will
display three randomly selected cards. The two outer cards will be
displayed face-up, and the middle card will be displayed face-down.
You must place a bet on the likelihood that the middle card will have
a value between the two outer cards. If the middle card has a value
between the two outer cards, then you will win chips on the hand. If
the middle card has the same value as one of the outer cards, this is
counted as a loss. An Ace is the highest card, followed by a King,
Queen, and Jack. An Ace can be either a high or a low card. For
example, if you have the cards Ace, Queen, and Five, then the Ace is
higher than the Queen. However, if you have the cards Ace, Three,
and Six, then the Ace is lower than the Three.
Following additional instructions about how many chips they
would start with and how much they could bet on each hand, each
participant engaged in the first round of play, where the experi-
mental manipulation occurred. All participants started with 1,500
chips. A maximum bet was set at 20% of the total chip count to
ensure that all participants would have enough chips to play all
hands in the round. (The use of this specific maximum bet amount
ensured that even if a participant bet the maximum number of
chips allowed and lost on every hand, he or she would finish the
round with exactly zero chips.) Participants were required to bet a
minimum of one chip on every hand, and maximum allowed bets
were rounded down to the nearest whole chip.
Participants clicked the “Deal a New Hand” button and the
computer displayed three cards. The two outside cards were dis-
played face-up and the middle card was displayed face-down.
Participants then typed a bet (number of chips) and pressed the
“Enter Bet” button. Once the participant entered his or her bet, the
computer displayed the middle card face-up, and below the cards
the word “Win” or “Lose” was displayed to indicate whether chips
were won or lost on the hand. Participants then pressed the “Deal
a New Hand” button and a new hand was dealt. Each participant
played 30 hands in the first round of cards.
At the conclusion of the first round of play, participants were
prompted by the computer to respond to a few questions about
their first round of tournament play (the manipulation checks).
Participants were then asked to “just take a break and write down
everything that comes to mind for the next five minutes” (the
stream-of-consciousness writing task). Participants typed their re-
sponses into a basic textbox while a timer counted down from five
289
WINNING AND POSITIVE AFFECT
minutes. At the conclusion of this writing task, participants noti-
fied the experimenter that they had finished.
The experimenter then returned to the participant and told
him or her that he or she would have a second chance at winning
the tournament. The experimenter informed the participant that the
rules for the second round of play would be the same as during the
first round. All participants started with a fresh set of 1,500 chips.
The second round of the tournament was exactly the same as the
first round, with the exception that during this round of play the
cards were not rigged. Instead, the computer actually randomly
selected three cards from a standard 52-card deck. Participants
once again played 30 hands of cards during the second round and
notified the experimenter when they had finished.
At the conclusion of card play, all participants were then given
a partial debriefing about the purpose of the study and told that
they would be contacted via campus mail regarding the winners of
the prize money. They were not informed that the winners of the
tournament would be chosen using a random drawing because the
study was contingent upon the participants’ belief that players with
the highest chip totals would win the tournament prizes. At the
conclusion of the study, participants were mailed a full debriefing
form explaining the purpose of the deception (regarding the dis-
tribution of prizes) in the study. Those participants whose names
were drawn as winners were notified and retrieved their cash
prizes from the authors.
Results
Before describing our results, we must mention that we initially
tested to see whether participants’ gender would moderate the
relation between our manipulation and reckless betting. However,
including participants’ gender as a factor in our analyses of vari-
ance (ANOVAs) revealed no interactions between gender and
condition for any of our dependent variables in either Experiment
1 or Experiment 2. Thus, we dropped gender as a factor in our
analyses.
Card playing experience. For the pretest item assessing the
participant’s self-reported skill at playing cards, the mean for this
sample was 4.46 (SD 1.32), with 1 not at all skilled and 7
extremely skilled. For their self-reported experience playing cards,
their mean was 4.44 (SD 1.41), with 1 none and 7 a great
deal. Thus, on average, the participants in this sample perceived
themselves to be moderately skilled and moderately experienced at
card playing.
Manipulation checks. Preliminary analyses were conducted
using data collected from the two manipulation checks concerning
participants’ perceptions of the number of hands that they had just
won or lost in the first round of the tournament. One-way between-
participants ANOVAs comparing the responses of participants in
the Initial-Winning and Initial-Losing condition on each item
revealed an overall significant effect at the p.0001 level for
condition for both items. Individuals in the Initial-Winning con-
dition reported experiencing a win on a greater percentage of the
hands, as compared with those in the Initial-Losing condition, F(1,
106) 1369.70. And they reported experiencing a loss on a lower
percentage of the hands, as compared with those in the Initial-
Losing condition, F(1, 106) 1369.70. See Table 1 for the means
and SDs. In fact, participants guessed very closely the actual
percentage of hands that they won or lost. These results strongly
supported the idea that the manipulation in the first round of the
tournament was successful.
Betting behavior. We conducted one-way between-
participants ANOVAs on Expected Chips Won and Expected
Chips Lost. The analysis revealed no significant difference in
Expected Chips Won for individuals in the Initial-Winning and
Initial-Losing conditions ( p.10). However, the analysis of
Expected Chips Lost did reveal a significant difference for indi-
viduals in the Initial-Winning and Initial-Losing conditions F(1,
106) 4.18, p.04. See Table 2 for the means and SDs.
Participants in the Initial-Winning, as compared with the Initial-
Losing, condition bet more chips on hands where the probability of
winning was less than 50%.
Discussion
Consistent with our hypothesis, participants in the Initial-
Winning condition bet more recklessly in the second round of the
tournament than did those in the Initial-Losing condition. Specif-
ically, they placed inappropriately high bets on hands where their
probability of winning was below 50%. At the same time, there
was no significant difference between the Initial-Winning and
Initial-Losing groups in their betting when a win was likely.
These results demonstrate that winning actually can cause an
individual to play worse on subsequent gambles. This finding is
consistent with the findings of Thaler and Johnson (1990) and
supports the notion that their results can apply in an actual gam-
bling scenario. As predicted, the results of this study contradict
those of Leopard (1978), where most participants had become
more risk-taking in response to a losing streak.
One possible explanation for these findings is the same reason
that individual investors in the stock market tend to invest poorly,
buying stocks when they have recently risen in value and selling
them when they have recently lost value (Dalbar Inc., 2003):
People expect trends to continue when they are gambling. If they
Table 1
Experiments 1 and 2: Mean Scores (and SDs) on Manipulation
Checks
Experiment Item
Condition
Initial-Winning Initial-Losing
Experiment 1 On what percentage of
hands played did
you experience a
win?
80.84 (7.96)
a
18.71 (9.38)
b
On what percentage of
hands played did
you experience a
loss?
19.16 (7.96)
a
81.29 (9.38)
b
Experiment 2 On what percentage of
hands played did
you experience a
win?
81.00 (11.56)
a
19.41 (10.73)
b
On what percentage of
hands played did
you experience a
loss?
19.00 (11.56)
a
80.59 (10.73)
b
Note. Means in the same row that do not share subscripts differ at p
.0001.
290 CUMMINS, NARDORFF, AND KELLY
have defied odds by winning more than they than they should in
previous gambles, then they expect to continue to win even in the
face of slim odds and thus end up betting recklessly. In a similar
phenomenon known as “playing the rush” in casinos across the
world, when card players are winning they often decide to play
hands that they would normally fold because they expect to con-
tinue to win despite the odds (Miller, Sklansky, & Malmuth,
2004).
Although playing the rush typically is a conscious process,
another possible explanation for why participants in the Initial-
Winning condition bet recklessly in the second round of the
tournament does not require conscious thought. It is simply that
because they had been reinforced for betting recklessly in the
first round, they continued in the second round to bet on hands
that were likely to lose. After all, many studies have shown that
reinforcement of a given behavior leads to an increase in that
behavior (see Skinner, 1963).
Yet another possible explanation for the present findings is that
participants in the Initial-Winning condition experienced an ele-
vation in positive mood after the first tournament relative to those
in the Initial-Losing condition. Several studies suggest that people
experiencing positive affect tend to overestimate the likelihood of
experiencing positive events (Johnson & Tversky, 1983; Nygren,
Isen, Taylor, & Dulin, 1996; Wright & Bower, 1992). Thus, in the
present study, individuals in the Initial-Winning, as compared with
those in the Initial-Losing, condition may have felt more positive
at the end of the first round of the tournament and subsequently
overestimated their likelihood of winning on hands where the odds
were not in their favor. This notion is consistent with the research
of Anderson and Galinsky (2006), who have shown that having a
more optimistic perception of risk leads individuals to make riskier
choices on decision-making tasks. Perhaps the experience of win-
ning leads to more positive affect, which in turn results in more
optimistic perceptions of risk. However, Experiment 1 did not
assess participants’ affect after the first tournament. Moreover,
there are reasons to believe that not just positive affect, but also
negative affect could lead to reckless betting based on observations
that feeling aggravated after losing is linked to impulsive acts (e.g.,
Parke & Griffiths, 2004, 2005a, 2005b). Thus, a follow-up exper-
iment was necessary to test the role of affect in reckless betting.
Experiment 2
The first purpose of Experiment 2 was to try to replicate the
findings from Experiment 1, which showed that winning in one
tournament led to betting too much on hands that were likely to
lose (i.e., betting recklessly) in the next. The second purpose was
to investigate the role of affect, both positive and negative, in such
reckless betting. We predicted that participants in the Initial-
Winning condition would experience more positive affect than
would those in the Initial-Losing condition, and that positive affect
would be linked to betting too much on hands that were likely to
lose. At the same time, we predicted that negative affect also
would be associated with more reckless betting. We based this
latter prediction on the growing number of findings that have
linked gambling to negative affect (see Daughters, Lejuez, Lesieur,
Strong, & Zvolensky, 2003, for a review) and to aggressive be-
haviors, such as impulsively verbally lashing out at a bypassing
casino employee (Griffiths, Parke, & Parke, 2003; Mulenman,
DenOtter, Wadman, Tran, & Anderson, 2002; Parke & Griffiths,
2004, 2005a, 2005b). We reasoned that these findings might be
accounted for by the negative emotions that likely occur when
gambling and underlie such impulsive acts.
Method
Participants. Participants were 73 students at a private univer-
sity in the Midwest. They were selected for this study because they
signed up to participate through the psychology department. All
participants had the opportunity to win a $250, $100, or $50 cash
prize, and they all received extra credit in a psychology course in
exchange for their participation. One male participant was dropped
because he was approximately 8 SDs from the mean on the
primary dependent variable, Expected Chips Lost. This left 72
participants who ranged in age from 18 to 22 years, with a mean
age of 19.2 years (SD 1.18). Of the participants, 58 identified
themselves as Caucasian (81.7%), 2 as African American (2.8%),
6 as Hispanic or Latino (8.5%), 3 as Asian (4.2%), 2 as “other”
(2.8%). Approximately 61% of the participants were male (N
43). One participant did not respond to the demographic questions.
The same participant also did not complete the affect measure and
therefore was excluded from analysis of the affect data. This
participant was included in the analyses of betting behavior.
Measures. The Positive and Negative Affect Schedules
(PANAS; Watson, Clark, & Tellegen, 1988) were included to
assess affect following the experience of induced winning or
induced losing in the first round of the tournament. The PANAS is
a 20-item measure. Participants are provided with a list of adjec-
tives (e.g., for positive affect: inspired, attentive, and excited; for
negative affect: hostile, scared, and upset) and are asked to rate the
extent to which each adjective describes their feelings at the
moment on a scale from 1 (very slightly or not at all)to5
(extremely). Supporting its validity are correlations between the
PANAS and related constructs of distress, depression, and state
anxiety (see Watson et al., 1988). Cronbach’s alphas for this
experiment were .90 for positive affect (PA) and .82 for negative
affect (NA).
Once again, participants’ betting behavior was assessed using
the bets they placed during the second round of the tournament.
For each hand, the Acey-Deucey software stored the number of
chips bet, the percentage of maximum possible bet placed, the
probability of winning the hand, and whether the hand resulted in
a win or a loss of chips in a data file.
Table 2
Experiments 1 and 2: Mean Scores (and SDs) on Expected
Chips Won and Lost
Experiment Dependent variable
Condition
Initial-Winning Initial-Losing
Experiment 1 Expected chips won 1227.70 (875.88)
a
1061.44 (715.78)
a
Expected chips lost 791.38 (535.79)
a
577.65 (545.19)
b
Experiment 2 Expected chips won 1176.42 (752.67)
a
1028.35 (719.54)
a
Expected chips lost 756.34 (468.96)
a
321.59 (229.84)
b
Note. Means in the same row that do not share subscripts differ signifi-
cantly at p.05.
291
WINNING AND POSITIVE AFFECT
This experiment also included demographic questions regarding
the participants’ age, sex, and race. The same manipulation-check
items used in Experiment 1 were included at the conclusion of the
first round of card play to ensure that participants perceived
themselves as experiencing a majority of wins or losses, depending
on their assigned condition.
Procedure. Participants came to the laboratory at their as-
signed times and were randomly assigned to either the Initial-
Winning or the Initial-Losing condition upon arrival. All partici-
pants signed a consent form and were given instructions by the
experimenter.
The procedure for Experiment 2 was largely the same as Ex-
periment 1, except that the participants did not complete a writing
task after answering the manipulation check questions. Instead,
they completed the PANAS. They also completed the Iowa Gam-
bling Task (Bechara, Damasio, Tranel, & Anderson, 1994) as a
filler task to provide a break between the Acey-Deucey rounds.
The experimenter then read them instructions, and they played the
second round of Acey-Deucey.
Results
Manipulation checks. Means and SDs for each manipulation-
check item in Experiment 2 are reported in Table 1. On average,
individuals in the Initial-Winning condition, as compared with
those in the Initial-Losing condition, reported experiencing a win
on a greater percentage of the hands, and experiencing a loss on a
lower percentage of the hands. One-way between-participants
ANOVAs conducted on the responses to each item revealed sig-
nificant differences between the two conditions, F(1, 71)
545.01, p.0001, for both items. We concluded that the manip-
ulation in the first round of the tournament was successful once
again.
Primary analyses. A one-way between-participants ANOVA
revealed no significant difference in Expected Chips Won between
participants in the Initial-Winning and Initial-Losing conditions
(p.10). However, a one-way between-participants analysis of
Expected Chips Lost did reveal that participants in the Initial-
Winning condition bet more chips on hands that were likely to lose
than did those in the Initial-Losing condition, F(1, 71) 24.03,
p.0001. See Table 2 for the means and SDs. This result
replicates the findings of Experiment 1.
Also, consistent with the suggestion that participants in the
Initial-Winning condition experienced an elevation in positive
mood following their winning streak relative to participants in the
Initial-Losing condition, a one-way between-participants ANOVA
on affect scores revealed that participants in the Initial-Winning
condition (M31.78, SD 6.49), compared with those in the
Initial-Losing condition, had significantly higher levels of positive
affect (M28.29, SD 8.45), F(1, 70) 3.85, p.05. (Note
that this result does not indicate whether participants in the Initial-
Winning condition had an increase in affect relative to a baseline
of affect, given that pretest affect was not measured.) There was no
significant difference in negative affect between the two groups
(p.10).
We followed up these between-participants ANOVAs with cor-
relational analyses, collapsing across conditions. The scores on
positive affect and negative affect ranged from 10 to 49 and from
10 to 25, respectively. As predicted, we found a significant posi-
tive correlation between positive affect and Expected Chips Lost,
r(71) .29, p.01. At the same time, we found a significant
negative correlation between positive affect and Expected Chips
Won, r(71) ⫽⫺.24, p.04. Thus, positive affect was linked to
poorer betting. However, contrary to our predictions, we found no
significant correlation between negative affect and Expected Chips
Lost or Expected Chips Won, ps.10.
We next conducted a multiple-regression analysis that entered
condition (coded as 1for Initial-Winning condition and 2for
Initial-Losing condition) and positive affect scores simultaneously
as predictor variables of Expected Chips Lost (i.e., reckless bet-
ting) to see if the effect of our manipulation on reckless betting
was mediated by positive affect. Together, these 2 predictors
accounted for 30% of the variance in Expected Chips Lost. A
significant unique predictive effect was obtained for condition,
standardized ␤⫽⫺.47, t(70) ⫽⫺4.51, p.0001; and a mar-
ginally significant effect was obtained for positive affect scores,
standardized ␤⫽.19, t(70) 1.78, p.08. Because this analysis
still showed a strong significant link between winning in the first
tournament and reckless betting in the second tournament even
when statistically controlling for positive affect scores, it suggests
that this link was not mediated by positive affect.
General Discussion
Experiments 1 and 2 tested whether winning versus losing in an
initial tournament would lead college students to bet more reck-
lessly and poorly in a second tournament. In both experiments, the
first tournament was rigged so that participants assigned to the
Initial-Winning condition would win 80% of the hands, and par-
ticipants assigned to the Initial-Losing condition would lose 80%
of the hands. The second tournament was not rigged; and the
participants were given a fresh set of chips. They were told that
their performance in the second tournament would give them an
independent opportunity to win one of the large cash prizes.
The results from both experiments demonstrated that the stu-
dents in the Initial-Winning condition, as compared with those in
the Initial-Losing condition, bet significantly more chips on hands
that were likely to lose in the second tournament (i.e., they bet
recklessly). Moreover, Experiment 2 showed that participants in
the Initial-Winning condition experienced significantly more pos-
itive affect than did those in the Initial-Losing condition after the
first tournament, and that positive affect was significantly posi-
tively correlated with reckless betting in the second tournament. In
fact, the condition to which participants were assigned and how
positive they felt before playing in the second tournament ac-
counted for a fairly large percentage (30%) of the variability in
their reckless betting. However, our findings did not support the
idea that the effect of winning versus losing on reckless betting
was mediated by participants’ affect. We say this because the
condition to which participants had been assigned was still a
significant predictor of reckless betting even when the analyses
partialled out their positive affect scores. What our findings did
show is that, separately, winning was a strong predictor and
positive affect was a moderate predictor of reckless betting.
What implications do these findings have for predicting and
understanding the problem gambling to which young adults and
college students are particularly vulnerable? Our findings suggest
that if college students find themselves experiencing an unusually
292 CUMMINS, NARDORFF, AND KELLY
big winning streak and/or find themselves feeling particularly
good while gambling, these could be warning signs that they might
start betting recklessly. Indeed, winning in both experiments
caused the student participants to bet recklessly, perhaps be-
cause they expected their good luck to continue despite the odds
(see Johnson & Tversky, 1983; Nygren et al., 1996; Wright &
Bower, 1992). After all, previous research has demonstrated
that having a more optimistic perception of risk can lead indi-
viduals to make riskier choices on decision-making tasks
(Anderson & Galinsky, 2006). Our findings are particularly
informative given the recent emphasis in the positive psychol-
ogy literature on the benefits of positive emotions (see
Fredrickson, 2008, for a review). In contrast to the thrust of that
literature, we have provided clear evidence that being success-
ful and feeling good when gambling may actually set a person
up for making poor, risky betting decisions.
Understanding these risk factors may be essential to inform
interventions designed to prevent reckless betting and therefore
help short-circuit the problem gambling that could follow in
gamblers’ attempts to recoup losses from such reckless betting
(see Campbell-Meiklejohn et al., 2008). Recent research sug-
gests that merely trying to intervene by educating college
students on gambling odds and fallacies is insufficient to reduce
problem gambling (Williams & Connolly, 2006). Therefore,
we suggest that a two-pronged approach could be tested in
future research that might help in the following way. First,
college students can be educated that winning can cause them to
lose perhaps because it sets them up to expect that unlikely wins
will continue and thus to bet recklessly. They can be trained to
constantly reset their expectations for outcomes in subsequent
independent bets and to be especially vigilant of their betting
behavior after they have been winning. However, given Wil-
liams and Connolly’s (2006) findings on the insufficiency of
mathematically-based interventions, this intervention is un-
likely to be sufficient to change behavior without addressing the
role of emotions in reckless betting. Therefore, the second step
would be to use traditional negative mood management tech-
niques, such as those involved in cognitive– behavioral treat-
ment, that have been shown to be effective in experiments,
multiple-baseline studies, and case studies with a wide range of
clinical populations (see Chemtob, Novaco, Hamada, & Gross,
1997; Novaco, 1994). But rather than help gamblers become
more aware of negative feelings, perhaps these techniques
could be used help gamblers become more aware of the positive
feelings that they might experience while playing and help them
avoid letting their emotions cause them to bet recklessly.
Before closing, we note that a key limitation to the present
research was that there was variability in the reckless betting of
participants within each condition of these experiments. Thus,
although our experiments captured differences between the Initial-
Winning and Initial-Losing groups on the average, they were not
able to tell us about individual differences in reckless betting.
Perhaps future researchers could examine potential moderators of
the relationship between winning versus losing and reckless bet-
ting. Such moderators might include impulsiveness, sensation-
seeking, and competitiveness, given that these have been tied to
problem gambling in past research (e.g., McDaniel & Zuckerman,
2003; Parke, Griffiths, & Irwing, 2004; Vitaro, Arsenault, &
Tremblay, 1997).
Conclusion
Previous research had left open the question of whether winning
or losing is more likely to lead to risky, poor betting. These
experiments point to the conclusion that winning can cause losing,
and so gamblers need to be aware of the dangers of previous
winning streaks. The findings also point to the idea that managing
positive emotions could be an essential part of avoiding reckless
betting. Thus, although the findings from the present experiments
do not answer the question of why winning causes losing or why
feeling positive is linked to more reckless betting, they do help
researchers and clinicians by identifying these as key risk factors
in college students’ losing more money than they can afford.
References
Anderson, C., & Galinsky, A. (2006). Power, optimism, and risk-taking.
European Journal of Social Psychology, 36, 511–536.
Bechara, A., Demasio, A. R., Damasio, H., & Anderson, S. W. (1994).
Insensitivity to future consequences following damage to human pre-
frontal cortex. Cognition, 50, 7–15.
Campbell-Meiklejohn, D. K., Woolrich, M. W., Passingham, R. E., &
Rogers, R. D. (2008). Knowing when to stop: The brain mechanisms of
chasing losses. Biological Psychiatry, 63, 293–300.
Chemtob, C. M., Novaco, R. W., Hamada, R. S., & Gross, D. M. (1997).
Cognitive– behavioral treatment for severe anger in posttraumatic stress
disorder. Journal of Consulting and Clinical Psychology, 65, 184 –189.
Dalbar, Inc. (2003). Market chasing mutual fund investors earn less than
inflation. Retrieved April 1, 2008, from Http://www.dalbarinc.com/
content/showpage.asp?page2003071601&r/pressroom/default.asp&s
Return%2BTo%2BPress%2BReleases
Daughters, S. B., Lejuez, C. W., Lesieur, H. R., Strong, D. R., & Zvolen-
sky, M. J. (2003). Towards a better understanding of gambling treatment
failure: Implications of translational research. Clinical Psychology Re-
view, 23, 573–586.
Fredrickson, B. L. (2008). Promoting positive affect. In M. Eid & R. J.
Larsen (Eds.), The science of subjective well-being (pp. 449 – 468). New
York: Guilford.
Gerstein, D., Volberg, R. A., Murphy, S., Toce, M., Harwood, H., &
Christiansen, E. (1999). Gambling impact and behavior study. Report to
the National Gambling Impact Study Commission. Chicago: National
Opinion Research Center at the University of Chicago.
Griffiths, M. D., Parke, A., & Parke, J. (2003). Violence in gambling
environments: A cause for concern? Justice of the Peace, 167, 424 – 426.
Johnson, E., & Tversky, A. (1983). Affect, generalization, and the percep-
tion of risk. Journal of Personality and Social Psychology, 45, 20 –31.
Leopard, A. (1978). Risk preference in consecutive gambling. Journal of
Experimental Psychology: Human Perception and Performance, 4, 521–
528.
McDaniel, S. R., & Zuckerman, M. (2003). The relationship of impulsive
sensation seeking and gender to interest and participation in gambling
activities. Personality and Individual Differences, 35, 1385–1400.
Miller, E., Sklansky, D., & Malmuth, M. (2004). Small stakes Hold ’em:
Winning big with expert play. Henderson, NV: Two Plus Two Publishing.
Mulenman, R. L., DenOtter, T., Wadman, M. C., Tran, T. P., & Anderson,
J. (2002). Problem gambling in the partner of the emergency department
patient as a risk factor for intimate partner violence. Journal of Emer-
gency Medicine, 23, 307–312.
Novaco, R. W. (1994). Clinical problems of anger and its assessment and
regulation through a stress coping skills approach. In W. O’Donohue &
L. Krasner (Eds.), Handbook of psychological skills training: Clinical
techniques and applications (pp. 320 –338). Boston: Allyn & Bacon.
Nygren, T., Isen, A., Taylor, P., & Dulin, J. (1996). The influence of
293
WINNING AND POSITIVE AFFECT
positive affect on the decision rule in risk situations: Focus on outcome
(and especially avoidance of loss) rather than probability. Organiza-
tional Behavior and Human Decision Processes, 66, 59 –72.
Parke, A., & Griffiths, M. D. (2004). Aggressive behavior in slot machine
gamblers: A preliminary observational study. Psychological Reports, 95,
109 –114.
Parke, A., & Griffiths, M. D. (2005a). Aggressive behaviour in adult slot
machine gamblers: A qualitative observational study. International
Journal of Mental Health and Addiction, 2, 50 –58.
Parke, A., & Griffiths, M. D. (2005b). Aggressive behaviour in adult slot
machine gamblers: An interpretative phenomenological analysis. Jour-
nal of Community & Applied Social Psychology, 15, 255–272.
Parke, A., Griffiths, M. D., & Irwing, P. (2004). Personality traits in
pathological gambling: Sensation seeking, deferment of gratification and
competitiveness as risk factors. Addiction Research and Theory, 12,
201–212.
Productivity Commission. (1999). Australia’s Gambling Industries (Report
No. 10). Canberra, Australian Capital Territory: AusInfo.
Shaffer, H. J., & Hall, M. N. (2001). Updating and refining meta-analytic
prevalence estimates of disordered gambling behavior in the United
States and Canada. Canadian Journal of Public Health, 92, 168 –172.
Skinner, B. F. (1963). Operant behavior. American Psychologist, 18, 503–
515.
Thaler, R. H., & Johnson, E. J. (1990). Gambling with the house money
and trying to break even: The effects of prior outcomes on risky choice.
Management Science, 36, 643– 660.
University of Chicago, Illinois, National Opinion Research Center. (1999,
April). Gambling Impact and Behavior Study: Report to the National
Gambling Impact Study Commission. Retrieved November 28, 2006,
from http://www.norc.org/new/pdf/gamble.pdf
Vitaro, F., Arsenault, L., & Tremblay, R. E. (1997). Dispositional predic-
tors of problem gambling in male adolescents. American Journal of
Psychiatry, 154, 1769 –1770.
Volberg, R., Abbott, M. W., Ro¨nnberg, S., & Munck, I. M. (2001).
Prevalence and risks of pathological gambling in Sweden. Acta Psychi-
atrica Scandinavica, 104, 250 –256.
Watson, D., Clark, L. A., Tellegen, A. (1988). Development and validation
of brief measures of positive and negative affect: The PANAS scales.
Journal of Personality and Social Psychology, 54(6), 1063–1070.
Williams, R. J., & Connolly, D. (2006). Does learning about the mathe-
matics of gambling change gambling behavior? Psychology of Addictive
Behaviors, 20, 62– 68.
Wright, W. F., & Bower, G. H. (1992). Mood effects on subjective
probability assessments. Organizational Behavior and Human Develop-
ment Processes, 52, 276 –291.
Received April 16, 2008
Revision received November 7, 2008
Accepted November 12, 2008
294 CUMMINS, NARDORFF, AND KELLY
... Gamblers exhibit different types of irrational behaviors that are not explicable from a rational or an economic perspective (Delfabbro, 2004;Fortune & Goodie, 2012;Ladouceur & Walker, 1996). Reckless betting is one type of irrational betting behavior in which people bet too much even though they are more likely to lose than win (Cummins et al., 2009). For example, B is more reckless when comparing two bets on a game with a 30% chance of winning, such that A is a 100-chip bet and B is a 200-chip bet. ...
... In other words, reckless betting is trying to gain more by taking risks in a gamble with negative economic utility. Cummins et al. (2009) determined the Expected Chips Lost (ECL) as an index of betting recklessness by considering the above operational definition. The ECL is calculated based on the value of a bet placed on a gamble (i.e., bet size) when the objective winning probability is less than 50%. ...
... The ECL is calculated based on the value of a bet placed on a gamble (i.e., bet size) when the objective winning probability is less than 50%. Cummins et al. (2009) conducted experiments using a gambling task and examined the effects of prior winning versus losing experiences on subsequent betting recklessness. Their results indicated that participants experiencing many wins in the first session of the task made more reckless bets in the second session than those who experienced many losses in the first session. ...
Article
Full-text available
People who have experienced many gambling wins tend to make larger bets even when they are unlikely to win (reckless betting) than those who have experienced many losses. This study examined psychological factors underlying reckless betting when gambling from the perspectives of affect and risk–benefit perception. University students (N = 63) participated in an experiment using the Acey-Deucey Task, in which the number of wins and losses during the 1st session was experimentally manipulated such that there were either 24, 12, or 6 wins out of 30 trials. Positive–negative affect and perceived risk–benefit during the task were assessed by self-report. Betting recklessness during the 2nd session was calculated using winning probability and bet size data in each trial. The results indicated that experiencing few prior wins, that is, many prior losses decreased positive affect and perceived benefits of betting and increased negative affect and perceived risks of betting. Path analysis results suggested that gambler's positive and negative affect altered perceived benefits of betting, which influenced reckless betting. Although participants that experienced more prior wins made more reckless bets similar to previous studies, there were no statistical differences between the three groups. Time-series analysis revealed that participants who experienced many prior losses made increasingly reckless bets at the end of the gambling task. We have discussed other potential variables that might have influenced recklessness, and the time-series analysis’ implications on reckless betting and loss-chasing.
... More specifically, previous studies on gambling have focused mainly on whether winning or losing is more likely to lead to risky betting, and the results are mixed (e.g. Cummins et al., 2009;Kostek & Ashrafioun, 2014;Leopard, 1978;Mentzoni et al., 2012;Monaghan et al., 2009;Smith et al., 2009;Thaler & Johnson, 1990). For example, Leopard (1978) reported that individuals became more risk-taking following a gambling loss. ...
... They found that the participants were more likely to choose the risky option of gamble A after a win than after a loss. Another strand of evidence comes from an experiment reported by Cummins et al. (2009). They tested whether college students bet more recklessly following an experimentally induced winning or losing streak in a computerized game of cards. ...
... Specifically, some studies have exposed participants to predetermined outcomes (e.g. Cummins et al., 2009;Kostek & Ashrafioun, 2014;Mentzoni et al., 2012), and others have used hypothetical scenarios rather than an actual gambling task (e.g. Thaler & Johnson, 1990). ...
Article
Full-text available
Problem gambling is characterized by intense urges to repeatedly engage in gambling and is highly deleterious to individuals’ financial and social well-being. A fundamental issue in problem gambling is how repeated and risky betting behavior varies as a function of outcome history. We used empirical data on gamblers playing baccarat, one of the most popular casino games, to examine the effects of sequential winning versus losing on subsequent gambling behavior. Specifically, we analyzed data from 7,935,566 games played by 3,986 players at a land-based casino to examine changes in the betting amount and in the rate of betting on hands with different dividend rates according to prior consecutive wins or losses. The results revealed that the bet amount in baccarat gradually increased according to streak length, and this effect was more pronounced after sequential winning than after sequential losing. The proportions of multiple bets, including ‘longshots’ – hands with low winning percentages and high dividend rates – decreased after sequential losing but increased after sequential winning. The present study, as the first attempt to analyze a large dataset on baccarat betting, indicates that gamblers shift their gambling behavior to be more reckless after experiencing consecutive wins more than consecutive losses.
... This construct has been shown to have significant influence on many health-related intentions (Hagger et al., 2018;McEachan et al., 2011). Although not adopting he RAA, studies have also demonstrated the role of affect in gambling (e.g., Cummins et al., 2009;Simmons et al., 2016). For example, Cummins et al. (2009) found positive affect to relate to reckless betting. ...
... Although not adopting he RAA, studies have also demonstrated the role of affect in gambling (e.g., Cummins et al., 2009;Simmons et al., 2016). For example, Cummins et al. (2009) found positive affect to relate to reckless betting. The specific beliefs underlying affect are discussed below within the key belief section. ...
Article
Full-text available
University students demonstrate high rates of gambling behavior, which can have negative health consequences. Mobile gambling has rapidly expanded, and most students own a device. It is therefore important to understand the key psychological factors underlying students’ mobile gambling. This information could be used to inform interventions. Psychological determinants and beliefs were identified using the Reasoned Action Approach (RAA: Fishbein & Ajzen, 2010). Study 1 identified the modal salient behavioral, normative, and control beliefs. Thirty-five participants completed a questionnaire concerning each belief category. Content analysis identified 27 modal salient beliefs; Eleven behavioral, nine normative, and seven control. Study 2 assessed the predictive utility of the RAA and key beliefs using a cross-sectional design. Participants completed questionnaires assessing RAA determinants (N = 473) and mobile gambling behavior (N = 388). Constructs accounted for 66% and 56% of the variance in intention and behavior, respectively. Attitude (affective and instrumental) and injunctive norm significantly predicted intention, and intention and autonomy significantly predicted behavior. A number of key beliefs were identified including “Bonus offers and promotions” and “Access and availability of apps”. The studies provide important information regarding the psychology underlying students’ mobile gambling behavior. Interventions could target the identified key beliefs to change the behavior.
... somewhat static nature of flat design) of skeuomorphic games is more likely to be considered more active than a flat game. Finally, a change in general affect (Cummins et al. 2009) stemming from the previous factors could also occur. We intend to show that, while these variables may be influenced by the presence of skeuomorphic design elements, and in turn may influence choice of game, they are not the process through which skeuomorphism influences intended amount gambled. ...
Article
Full-text available
The current research focuses on how the use of “skeuomorphs” in digital design interacts with an illusion of control to influence gambling behavior. Skeuomorphism is a design concept in which an aspect of a modern item is made to represent its outdated counterpart, even if this representation serves no functional purpose, such as hands “dealing” cards or horses “racing” on the screen of a digital machine. It is proposed that the inclusion of these non-essential links to physical objects in digital gambling games interacts with a player’s illusion of control over the outcome to influence behavior. Shown across a pretest and three experiments, the inclusion of skeuomorphic elements in the design of gambling games, compared to a minimalistic “flat” design, increases amount gambled. Additionally, skeuomorphism and manipulated illusion of control interact to further increase the amount gambled. The manuscript concludes with a discussion on the practical, theoretical, and policy implications of this research.
... In that sense, it may be asserted that a difficulty to adequately manage positive emotions may lead to an excessive persistence of pleasant activities such as gambling. For instance, Cummins et al. (2009) found that winning and positive affect can lead to reckless gambling. Similarly, some authors tested the hypothesis that a proneness to act impulsively in response to positive emotional states is associated with GD. ...
Article
Full-text available
Emotion dysregulation is considered a central feature of gambling disorder. However, research has mainly focused on the examination of the role played by negative emotional states, neglecting processes related to positive emotionality. Recently, some authors developed a useful tool to measure dysregulation of positive emotions, the Difficulties in Emotion Regulation Scale-Positive (DERS-P), which has been successfully used in a study investigating substance use disorder. The role played by dysregulation of positive emotions in gambling disorder was investigated in order to bridge the gap in the literature. We administered the Impulsive Behavior Scale Short Form (UPPS-P) and the DERS-P to a sample of disordered gamblers (n = 95) using the South Oaks Gambling Screen (SOGS) as well as group of non-gamblers or non disordered gamblers (n = 105). We found that disordered gamblers, compared to controls, scored higher on the negative and positive urgency subscale of the UPPS-P as well as on the impulse and non-acceptance subscales of the DERS-P. Moreover, these scales positively correlated with the severity of gambling disorder. However, in a unique model of multiple regression analysis, only the positive urgency dimension was a significant predictor of the SOGS scores. Dysregulation of positive emotions seems to be involved in gambling disorder. Specifically, our study suggested that the difficulty to accept positive emotional states in non-judgmental way accounts for the disorder. However, the proneness to act rashly under the influence of positive emotional states appears to be the stronger predictor of gambling disorder severity in this sample.
... Previous research has examined the effects of priming on gambling, and from this work is it known that priming a positive mindset toward gambling, such as by cuing memories of prior wins, increases gambling behavior (Ludvig et al., 2015). While there are indications in the literature that positive affect is associated with more reckless gambling, such as the finding of a positive correlation between positive affect and reckless betting behavior (e.g., Cummins, Nadorff, & Kelly, 2009), to our knowledge, no studies have yet shown the kind of spillover effects reported in the present study. Somewhat similar to the halo effect (e.g., Morewedge & Kahneman, 2010), the patterns reported in the present study seem to suggest that mere concurrence with the afterglow of an unrelated positive-seeming event can influence risk-taking inclinations and behavior. ...
Article
The tip-of-the-tongue (TOT) state—the feeling of being near accessing an as yet inaccessible word from memory—is associated with cognitive bias. For example, prior work has shown that TOTs are associated with a bias toward inferring positive qualities of the unretrieved information. People are biased during TOTs to indicate that the unretrieved target has a greater likelihood of being positively valenced and to have been associated with a higher value number earlier in the experiment. Additionally, when the TOT is for a pictured person’s name, that person is judged to be more likely to be ethical. The present study demonstrates that the TOT positivity bias extends to unrelated concurrent decisions and behavior. In Experiment 1, participants reported a greater inclination to take an unrelated gamble during TOTs than non-TOTs. Experiment 2 demonstrated the concurrent nature of this spillover effect. The TOT bias toward a greater inclination to gamble significantly diminished with a 10-second delay between the time of reporting the TOT state and the time to report the inclination. Finally, Experiment 3 showed that the increased inclination to want to take a gamble during TOTs translated to actual gambling behavior. Participants chose to gamble for points more often during TOTs than non-TOTs.
... Negative affect is generally associated positively with problem gambling (Moghaddam et al. 2014). However, while positive affect is negatively associated with problem gambler samples (Hwang et al. 2012), it is positively associated with higher betting among non-problem-gambler student samples (Cummins et al. 2009). Optimism is also found in non-problem-gambler student samples to be associated with greater expectations of winning and with post-bet-loss gambling continuance (Gibson and Sanbonmatsu 2004). ...
Article
Full-text available
This study offers both the first systematic investigation of the relationship between the five-factor personality model and general (ostensibly non-problem) lottery gambling, and the first application of Thompson and Prendergast’s (2013) bidimensional model of luck beliefs to gambling behavior. Cross-sectional analyses (N = 844) indicate the bidimensional model of luck beliefs significantly accounts for variance in lottery gambling that is discrete from and greater than that of the five-factor personality model. Moreover, the broad pattern of relationships we find between presumably harmless state-sponsored lottery gambling and both personality and luck beliefs tend to parallel those found in studies of problem gambling, suggesting implications for quality of life and public policy in relation to lottery gambling.
... As discussed further below, given the design of the primary experiment, it is possible that affective reactions trigger more risk seeking behaviour among participants (Cummins et al., 2009;Cheung and Mikels, 2011). Potentially, these changes in risk seeking behaviour could explain our results for the primary experiment. ...
Article
It is generally recognized that decisions about capital projects should be made by independent reviewers who select the economically strongest projects. However, prior research finds that reviewers’ choices can be biased by their affective reactions to a manager proposing a capital project. Potentially, this bias could be reduced by holding reviewers more accountable. We contend that holding reviewers accountable will lessen the effect of positive, but not negative, affective reactions on capital project choice. To provide evidence on our predictions, we conduct an experiment using highly experienced participants. Participants’ task is to select between two capital projects, each proposed by a different manager. Although one project is economically preferred relative to the other project, we manipulate whether there is a negative affective reaction to the manager proposing the preferred project or a positive affective reaction to the manager proposing the non-preferred project. We also manipulate the presence versus absence of reviewer accountability. As expected, participants were more likely to select the economically non-preferred project when proposed by a manager triggering a positive affective reaction, but this tendency was reduced by accountability. Also, as expected, participants were less likely to select the economically preferred project when proposed by a manager triggering a negative affective reaction, and accountability did not reduce this tendency. Implications of our findings for theory and practice are discussed.
Article
The research on the relationship between wins and gambling behaviour most often focuses on winning considerably large amounts of money. It seems, however, that it is not the amount of the win that exerts a decisive influence on gambling behaviour but the significance that the player assigns to the win. Therefore, we adopted the concept of “significant win”, a win perceived by gamblers as important to them. The research aimed to discover what kind of wins are experienced as significant and what factors explain experiencing wins as significant. The research conducted in Poland (N = 3,143) and France (N = 5,692) also had a comparative goal: discovering intercultural differences in experiencing significant wins. The computer-assisted web survey was conducted among gamblers engaging in pure-chance gambling, selected from representative samples in both countries. Logistic regression models were used to examine predictors of significant win experience in both countries and cross-countries differences between them. The results demonstrated that Polish gamblers more frequently considered wins significant when accompanied by strong, often negative emotions and were objectively higher than French gamblers. A significant win was more frequently associated with a positive experience in the view of French gamblers. The common predictors of a significant win experience in both countries were: being in debt, experiencing the win of a close person, gambling in a game of pure chance other than lotteries, more systematic pursuit of gambling, enhancement and coping gambling motivations. The age of the initiation into gambling was a significant predictor only in the French sample, while financial motivation – in the Polish one. The results confirmed that the subjective perception of gambling wins is only partially related to the amounts of wins, which has practical implications for planning prophylactic strategies.
Research
Full-text available
This project focused on gamblers whose choose “pure chance” gambling, that is to say games based exclusively on “chance” (scratch, draw, slot machines, etc.) in contrast to games that include a share of skill or expertise. This decision is justified by the place these games occupy in the overall activities of gambling: 9 players out of 10 practice lottery games (drawing and scratching) according to the study of Costes et al, 2020. The elements of the literature review carried out prior to this study convinced us to let the players define themselves, the notion of significant win, that is to say one or more wins that marked their course. This completely subjective concept, goes from the financial value, to the use they make of it or the moment of its realization... The initial assumption was that significant win is one of the predictors of excessive gambling. The central question is therefore to measure the impacts of these significant wins in pure chance players, that is to say to make an inventory of contexts, reactions, cognitive and emotional impacts, behaviors following a significant win and differences in behavior types of players. To do this, we conducted a literature review on winnings, and then conducted a qualitative study with 30 moderate to excessive players, with a dual objective: enrich the quantitative questionnaire, which was to follow and deal with specific, easier-to-use topics in semi-structured interviews.
Article
Full-text available
Investigated the role of affect in judgments of risk in 4 experiments. 557 Ss were recruited on college campuses and read paragraphs modeled after newspaper reports that described fatal or nonfatal accidents or (Exp III) positive events. Ss were later asked to estimate the chances of specific fatal or nonfatal accidents happening to them and/or to the population in general. Experimental manipulations of affect induced by report of a tragic event produced a pervasive increase in Ss' estimates of the frequency of many risks and other undesirable events. Contrary to expectation, the effect was independent of the similarity between the report and the estimated risk: An account of a fatal stabbing did not increase the frequency estimate of homicide more than the estimates of unrelated risks such as natural hazards. An account of a happy event that created positive affect produced a comparable global decrease in judged frequency of risks. (12 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
Full-text available
Gambling is acknowledged to have many negative effects on the individual. However, from a psychological perspective, aggression as an outcome of gambling has been overlooked. This paper investigates the dynamics of the relationship between aggression and slot ma-chine gambling. A non-participation observation study observed a small group of aggressive individuals (n = 8; identified in a previous study) over a period of eight weeks. Four catego-ries of aggressive behaviour were confirmed from previous research (verbal aggression to-wards the gambling arcade staff; verbal aggression towards the slot machines; verbal aggres-sion towards other slot machine players; and physical aggression towards the slot ma-chines). From the in-depth observations, possible reasons motivating these types of aggres-sive behaviour are discussed. It is suggested that the frustration, guilt and embarrassment of losing are the prime causes of such aggression.
Article
Full-text available
How is risk-taking affected by prior gains and losses? While normative theory implores decision makers to only consider incremental outcomes, real decision makers are influenced by prior outcomes. We first consider how prior outcomes are combined with the potential payoffs offered by current choices. We propose an editing rule to describe how decision makers frame such problems. We also present data from real money experiments supporting a "house money effect" (increased risk seeking in the presence of a prior gain) and "break-even effects" (in the presence of prior losses, outcomes which offer a chance to break even are especially attractive).
Article
In recent studies of the structure of affect, positive and negative affect have consistently emerged as two dominant and relatively independent dimensions. A number of mood scales have been created to measure these factors; however, many existing measures are inadequate, showing low reliability or poor convergent or discriminant validity. To fill the need for reliable and valid Positive Affect and Negative Affect scales that are also brief and easy to administer, we developed two 10-item mood scales that comprise the Positive and Negative Affect Schedule (PANAS). The scales are shown to be highly internally consistent, largely uncorrelated, and stable at appropriate levels over a 2-month time period. Normative data and factorial and external evidence of convergent and discriminant validity for the scales are also presented. (PsycINFO Database Record (c) 2010 APA, all rights reserved)
Article
"Reinforcement may be contingent, not only on the occurrence of a response, but on special features of its topography, on the presence of prior stimuli, and on scheduling systems. Operant techniques are important in defining the behavioral effects of physiological variables—surgical, electrical, and chemical—in specifying what aspects of behavior are to be attributed to hereditary endowment, in tracing features of mature behavior to early environment, and so on. They are important in clarifying the nature of defective, retarded, or psychotic behavior." Within the field of human behavior "the contingencies of reinforcement which define operant behavior are widespread if not ubiquitous. In its very brief history, the study of operant behavior has clarified the nature of the relation between behavior and its consequences and has devised techniques which apply the methods of the natural science to its investigation." (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
Studied changes in risk preference over a series of consecutive decisions as a function of outcome history and financial state. 40 undergraduates chose and played 250 2-outcome gambles from 1 of 4 lists of gambles designed to differ only in amount of risk. In each list, risk was manipulated by variance or skewness but not both. All analyses were done separately for each S. Most Ss (67%) took more risk when they had fallen behind than when they were ahead. The amount of risk taken was also influenced by patterns of runs of wins and losses; the direction of the influence depended on the individual. The risk order between 2 gambles with zero expectation and the same variance but differing in skewness was not independent of the variance level, in contrast to the findings of an earlier study by C. H. Coombs and D. G. Pruitt (1960). (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
To date, research into personality traits among gamblers has been largely inconsistent. The purpose of this study was to assess the predictive values of three personality traits on pathological gambling (sensation seeking, deferment of gratification and competitiveness) – two of which (deferment of gratification and competitiveness) have never been investigated before. A questionnaire was administered to 114 gamblers of whom 38% were classified as pathological gamblers according to the DSM-IV criteria. The questionnaire included the Sensation Seeking Scale (Zuckerman, M. (1984). Sensation seeking: a comparative approach to a human trait. Behavioural and Brain Sciences, 7, 413–471.), the Deferment of Gratification Scale (Ray, J.J. and Najman, J. (1986). The generalisability of deferment of gratification. Journal of Social Psychology, 126, 117–119.) and the Gambling Competitiveness Scale constructed by the authors specifically for this study. Results showed that competitiveness had a strong positive predictive value for pathological gambling, and that deferment of gratification had relatively strong negative predictive value. Sensation seeking was shown not to be a significant predictor of pathological gambling. This is the first ever study to show empirically that competitiveness and deferment of gratification appear to be important risk factors in pathological gambling.
Article
A growing number of studies have reported a link between gambling and aggressive behaviour. The aim of this study was to contextualize objective findings of a previous observational study regarding slot machine gambling and aggressive behaviour (Parke & Griffiths, Psychological Reports, 95, 109–114, 2004). Interpretative Phenomenological Analysis was applied using the Idiographic Case Study method. The data revealed three superordinate themes regarding slot machine gambling-induced aggression (i.e. Competitive Advantage Reduction, Self-esteem Reduction, and Cognitive Regret). Within these superordinate themes, subordinate themes emerged identifying how environmental factors and structural characteristics of slot machine gambling, along with the consequences of losing, produced aggressive behaviour. It is concluded that gambling-induced aggression is a manifestation of the underlying conflict of engaging in dysfunctional behaviour while consciously acknowledging its detrimental effects. Copyright © 2005 John Wiley & Sons, Ltd.
Article
Five studies investigated the hypotheses that the sense of power increases optimism in perceiving risks and leads to more risky behavior. In Studies 1 and 2, individuals with a higher generalized sense of power and those primed with a high-power mind-set were more optimistic in their perceptions of risk. Study 3 primed the concept of power nonconsciously and found that both power and gain/loss frame had independent effects on risk preferences. In Study 4, those primed with a high-power mind-set were more likely to act in a risk-seeking fashion (i.e., engage in unprotected sex). In Study 5, individuals with a higher sense of power in a face-to-face negotiation took more risks by divulging their interests. The effects of power on risk-taking were mediated by optimistic risk perceptions and not by self-efficacy beliefs. Further, these effects were attenuated when the high-power individual felt a sense of responsibility. Copyright © 2006 John Wiley & Sons, Ltd.