Social Influences Involved in Children's Gambling Behavior

This is the html version of the file http://www.education.mcgill.ca/gambling/french/researche/PDF%20files/social3.pdf.
G o o g l e automatically generates html versions of documents as we crawl the web.
To link to or bookmark this page, use the following url:

http://www.google.com/search?q=cache:FpewQ1D0AOIC:www.education.mcgill.ca/gambling/french/researche/PDF%2520files/social3.pdf+%22children's+gambling%22&hl=en&ie=UTF-8

Google is not affiliated with the authors of this page nor responsible for its content.

These search terms have been highlighted:

children's

gambling

Page 1

Journal of Gambling Studies, Vol. 17, No. 3, Fall 2001 (

2001)

191

1050-5350/01/0900-0191$19.50/0 2001 Human Sciences Press, Inc.

Social Influences Involved in Children's Gambling

Behavior

Karen K. Hardoon

Jeffrey L. Derevensky

McGill University

Children (

130) in grades 4 and 6 from various schools in the Greater Montreal

Region completed a questionnaire concerning their gambling behavior and played a

computer-simulated roulette game individually (baseline trial & post-test trial) and in

groups (same and mixed gender dyads or triads: group trial). The purpose of this

design was to measure children's betting behavior (via average wagers) and to deter-

mine if any changes in betting occur as a result of playing in groups of two, three, same

and/or different gender peers. Results of repeated measures analyses reveal that dur-

ing individual and group play, males consistently exhibit higher average wagers than

females. Average wagers of females and mixed gender groupings appear to be most

affected by the group condition. Females were found to increase their average wagers

when playing with females and males. Female dyads' wagers increase significantly dur-

ing group play, indicating they are dramatically affected by the group game. Most

changes resulting from group play were generally maintained over a relatively short

period of time in the post-test condition. Results are interpreted with respect to the

importance of the influence of the peer group on children's gambling behavior. Future

directions for research are suggested.

KEY WORDS:

gambling; children; modeling; social learning; group effects.

Please address all correspondence to Karen Hardoon, McGill University, International Centre

for Youth Gambling Problems and High Risk Behavior, 3724 McTavish Street, Montreal, Quebec,

H3A 1Y2, Canada.

This is a revision of a paper presented at the 13

National Conference on Problem Gambling,

Detroit, MI, 1999, and has been adapted from Ms. Hardoon's M.A. thesis.

This research was partially supported by a grant from Loto Quebec to Dr. J.L. Derevensky.

Page 2

192

JOURNAL OF GAMBLING STUDIES

INTRODUCTION

In recent years, there has been growing concern over the increas-

ing occurrence of youth gambling problems. This is likely the result of

a widespread growth in the availability and accessibility of legalized

gambling. Gambling is legal in 48 out of 50 states, in all Canadian

provinces, and in approximately 90 countries worldwide (Lesieur &

Rosenthal, 1991; Stinchfield & Winters, 1998). Moreover, there is evi-

dence that gambling causes personal and financial difficulties in at

least 1%2% of the adult population (Ladouceur & Walker, 1996) and

is a significant burden on society (Ladouceur, Boisvert, Pepin,

Loranger, & Sylvain, 1994; Lesieur, 1998; National Opinion Research

Center Report, 1999).

Current prevalence rates of adolescent pathological gambling

(more than double adult gambling rates, see Gupta & Derevensky,

1998; Lesieur, Cross, Frank, Welch, White, Rubenstein, Moseley, &

Mark, 1991) indicate that between 4% and 8% adolescents exhibit

compulsive or pathological patterns of gambling activity (Fisher, 1993;

Gupta & Derevensky, 1998; Shaffer & Hall, 1996; Shaffer, LaBrie, Scan-

lan, & Cummings, 1994; Winters, Stinchfield, & Fulkerson, 1993;

Wynne, Smith, & Jacobs, 1996). Furthermore, 9.9% and 14.2% of ado-

lescents are at risk of developing or returning to serious gambling

problems (Shaffer & Hall, 1996). These findings are extremely worri-

some given the widespread availability of legalized gambling venues.

Gambling also appears to be prevalent in primary school. Among

, 5

, and 6

grade Canadian students, 86% had wagered money

and 40% reported gambling once a week or more (Ladouceur, Dub´e,

& Bujold, 1994b). Thus, gambling behavior appears to be established

early and to begin at the same time or earlier than other illicit behav-

iors (e.g., tobacco, alcohol, and other drug use). Retrospective studies

reveal that adolescent probable pathological gamblers report having

begun gambling at 9 or 10 years of age (Gupta & Derevensky, 1998;

Wynne, Smith, & Jacobs, 1996). Unfortunately, given that there are few

observable signs of gambling dependence among children, these prob-

lems have gone unnoticed compared to other addictions (e.g., alcohol

or substance abuse) (Arcuri, Lester & Smith, 1985; Lesieur & Klein,

1987).

Page 3

KAREN K. HARDOON AND JEFFREY L. DEREVENSKY

193

Social Acceptance of Gambling

Gambling is largely advertised, readily available to youth and of-

ten found in places which are glamorized, providing opportunities for

pleasant socializing (Stinchfield & Winters, 1998). Although betting in

casinos, on electronic gaming machines, and lotteries, in general, are

illegal for adolescents, the enforcement of these laws, as with under-

age drinking, can be difficult (Moore & Ohtsuka, 1997).

Family and Peer Influences.

Contact with gambling begins at an

early age. Thus, it is not surprising that several authors have placed an

emphasis on the relationship between gambling behaviors and gam-

bling in the family (Custer, 1982; Dell, Rozicka & Palisi, 1981; De-

revensky & Gupta, 1996; Gupta & Derevensky, 1997). Results of several

studies, in a number of countries, have indicated that the majority of

youth gamble with their family (40%68%) and friends (55%82%)

(Gupta & Derevensky, 1997; Derevensky, Gupta, & ´Emond, 1995; Hux-

ley & Carroll, 1992; Ladouceur & Mireault, 1988; Moore & Ohtsuka,

1997). Furthermore, parents appear to show little concern over their

children's gambling behavior. Between 80% and 90% of parents

readily acknowledge that their children gamble, 84% do not object

(Arcuri et al., 1985; Ladouceur & Mireault, 1988), while 78% of chil-

dren gamble in their own homes (Derevensky et al., 1995; Gupta &

Derevensky, 1997).

Parents often serve as role models for gambling. There is ample

evidence that 20% to 25% of adult gambler's children gamble them-

selves and/or exhibit various addictions (Lesieur & Klein, 1987;

Lorenz & Shuttlesworth, 1983; Winters, Bengston, Door, & Stinchfield,

1998; Winters et al., 1993) and that 25% to 40% of adult pathological

gamblers' parents were problem gamblers (Custer, 1982; Jacobs, Mar-

ston, & Singer, 1985).

Family and peer models have been suggested as important eti-

ological factors in the development of high-risk behavior including

drug and alcohol use (Santrock, 1990), and now with gambling behav-

ior as well. Griffiths (1990) reported that 44% of adolescents initiated

gambling behavior because their friends were engaging in similar

practices. While a large proportion of children's gambling activities

occur with and around family members, as they get older they tend to

gamble with friends at their friends homes (Derevensky et al., 1995;

Page 4

194

JOURNAL OF GAMBLING STUDIES

Gupta & Derevensky, 1996; 1997; Ide-Smith & Lea, 1988), reinforcing

the fact that for many youth gambling is perceived as a socially ac-

cepted and entertaining pastime. These findings indicate that there

may be a relationship between gambling and other addictions as well

as a strong social learning component involved in the acquisition of

such behaviors (Derevensky, Gupta, & Della-Cioppa, 1996; Gupta &

Derevensky, 1997).

Social Learning Perspective of Gambling Behavior

Social Learning Theory and Gambling Behavior.

Bandura (1977),

among other social learning theorists, has indicated that observational

learning and modeling play an important role in shaping individual

behaviors. Further, children are more likely to imitate significant and

powerful role models such as parents, siblings, and peers. Given that

modeling has been shown to have a strong learning component for

school age children in a number of areas (Bandura, 1973; 1977, Ban-

dura, Ross, & Ross, 1963; Rushton, 1980; Staub, 1978), it is likely that

it would be a strong component involved in the acquisition and main-

tenance of gambling behavior.

It is well known that gambling activities are reinforcing, in and of

themselves, as they produce excitement, arousal, and enjoyment. How-

ever, gambling activities have social reinforcements as well; gambling

behaviors can be encouraged and strengthened by peers depending

on the individual's developmental level and social status (Gupta,

1994). From a social learning theory perspective individuals learn, ac-

quire, and persevere with behaviors that are attractive and reinforcing.

A social learning model of gambling behavior provides an explanation

for why many youth are attracted to the gambling environment

(Gupta, 1994). For example, being known as a gambler or risk-taker

leads to social recognition and often to a higher status among peers

(Opie & Opie, 1969; Smith & Abt, 1984). Further, gambling venues

have been found to be a social space for adolescents. Fisher (1995)

surveyed adolescents who visited Amusement Arcades in the U.K.

(commercial sites for the playing of coin operated fruit/gambling ma-

chines and video games). She discovered that the primary motivation

of frequenting the arcades was to "hang out or meet friends." Addi-

tionally, she found that regular arcade visitors differed sufficiently

from casual visitors, suggesting a distinct social group (Fisher, 1995).

Page 5

KAREN K. HARDOON AND JEFFREY L. DEREVENSKY

195

Kearney and Drabman (1992) have demonstrated that even pre-

school children will engage in a gambling-like situation longer and

initiate more risks when exposed to an appropriate "big win" model.

This study demonstrated that modeling could enhance risk-taking/

gambling-like behavior in young children. These findings support

modeling antecedents as viable links resulting in increased risk-taking

and parallels contemporary theories of compulsive gambling in adults.

Gender Differences in Gambling Behavior

Despite some inconsistent findings, much of the gambling litera-

ture suggests that gambling is more popular amongst males than fe-

males (Derevensky & Gupta, in press; Fisher, 1993; Ide-Smith & Lea,

1988; Ladouceur, Dub´e & Bujold, 1994a; 1994b; Lesieur et al., 1991;

Rosenstein & Reutter, 1980; Stinchfield & Winters, 1998). Pathological

gambling is twice as prevalent for males than females (Lesieur & Klein,

1987; Lesieur et al., 1991; Moore & Ohtsuka, 1997; Stinchfield & Win-

ters, 1998; Volberg, 1994; Volberg & Steadman, 1988). Males have

been found to make higher gross wagers and exhibit greater risk-tak-

ing behavior (Derevensky et al., 1995). It appears as though parents

encourage gambling in their sons (Ladouceur et al., 1994a; 1994b)

and that gambling is much more a part of the male culture (Huxley &

Carroll, 1992). Griffiths (1989) speculates that gambling allows boys to

display their masculinity in a social environment by exhibiting "cour-

age and bravery" and thus may be more popular with them.

The principal goal of the present study was to examine the social

influences involved in children's gambling behavior via interactive

play. Given that modeling has been shown to be a strong learning

component in school age children, it is likely that it is also instrumen-

tal in the acquisition and maintenance of gambling behavior. The fact

that gambling has many reinforcing qualities, especially those that are

social in nature, makes gambling a very rewarding pastime. With the

increase in child and adolescent gambling activities these notions are

worrisome.

The hypotheses of the present study were: (1) that peers serve as

models for gambling behavior, such that children playing together in a

group will have an influence on each others' betting patterns; (2)

larger groups will have a greater effect on gambling behavior, as larger

groups are believed to elicit more competition; (3) males will be more

Page 6

196

JOURNAL OF GAMBLING STUDIES

influential in gambling activities, given that they are socialized early to

gamble; and (4) males will wager more than females as they are

greater risk-takers by nature.

METHOD

Participants

The sample included 130 children (69 males, 61 females) in

grades 4 (

64) and 6 (

66) from three elementary schools and

one summer camp in the Greater Montreal region. The participants

were between 9 and 13 years of age (

10.68,

1.18). All stu-

dents volunteered to participate and obtained parental consent.

Groups.

The children were randomly assigned into one of seven

groups by age, which consisted of either same gender or mixed gender

dyads and triads. The groups included the following: (1) male dyads

(male, male) (n 20); (2) female dyads (female, female) (n 16);

(3) mixed gender dyads (male, female; female, male) (n 16); (4)

male triads (male, male, male) (n 21); female triads (5) (female,

female, female) (n 18); (6) mixed gender triads (male, male, fe-

male) (n 21), and (7) (female, female, male) (n 18). It is impor-

tant to note that while children were randomly assigned into their

respective groupings, the sample sizes are unequal. This is due to the

fact that there were more boys whose parents had consented to their

participation in the study. As such, there were a greater number of

male dyads and triads. Once the groupings were made, the children

were randomly placed in the order in which they played in the base-

line and group and post game play. These groups were selected to

determine whether any gender and/or size differences existed with

respect to influence during game playing. It is important to note that

every effort was made to counterbalance these groups. However, with

respect to the mixed gender triads, the order of the groups was kept

consistent (male, male, female; female, female, male) as there would

have been too many possible combinations (i.e., male, female, male;

female, male, male, etc.). Given the principal scope of the present

study, this triad was not counterbalanced with respect to all possible

gender combinations.

Page 7

KAREN K. HARDOON AND JEFFREY L. DEREVENSKY

197

Instruments

Gambling Activities Questionnaire.

(GAQ) (Derevensky et al.,

1996), is a brief questionnaire designed to ascertain the frequency and

type of gambling activities as well as with whom individuals gamble.

Gambling Activities. Caesar's Palace for Windows

(Masteller & Mas-

teller, 1993), a computer simulated game of American Roulette was

included in order to observe participants' playing behavior. An IBM

compatible laptop was used in order to render the game as realistic as

possible. Roulette, although a game involving chance, involves active

choices and provides participants with an element of perceived con-

trol. In order to keep the game as simple as possible, the children

were only permitted to bet on either black, red, even, and/or odd or

any possible combinations thereof. No individual numbers, boxing,

zeros, etc. were permitted. The minimum bet permitted was $2 while

the maximum was whatever the child had in their bankroll.

Procedure

After each child was administered the GAQ, all participants

played the computer task individually (7 minutes) in order to provide

a baseline of playing behavior. Following the individual game, partici-

pants played the group game in either dyads (12 minutes) or triads

(14 minutes). Immediately following the group game, the children

played individually (post game: 7 minutes) in order to ascertain

whether any possible change in their playing behavior was maintained.

The average number of spins played was 13, 11, and 15 for the base-

line, group, and post games respectively.

With the use of Grand Vision Pro software, the roulette game on

the computer screen was recorded onto a VHS cassette. The video-

tapes were then played back; information including the amount bet,

type of bet, outcome, and end balance was recorded for each spin of

the wheel and transcribed onto sheets for each player.

Due to ethical constraints, children did not play with real money

but were given a bankroll of $500 on the computer. Incentives were

provided to the children in order to approximate a real playing situa-

tion. At the conclusion of the game, the children were given monop-

oly money equivalent to their winnings over the initial $500 supplied,

which was placed into a draw for prizes at a local restaurant and movie

Page 8

198

JOURNAL OF GAMBLING STUDIES

theatre. This type of incentive has previously proven successful with

children (Derevensky et al., 1996; Gupta & Derevensky, 1996) and adults

(Powell, Hardoon, Derevensky, & Gupta, 1999). Further, money appears

to not hold as great a value with children as does other incentives, such

as candy, toys, and movie certificates. In fact, the verbalizations of the

children suggested that they wanted to increase their bankroll in order

to have a greater chance of winning one of the prizes.

Standardized instructions and demonstrations were given to each

child before they began to play roulette to ensure that each child

would have an equivalent understanding of the task (please see Ap-

pendix A). A research assistant was present at all times in order to

answer any questions and to ensure an understanding of the task.

RESULTS

Average Amount Bet

For each game (baseline, group, and post), the average amount

bet was calculated by summing the participants' bets for each spin and

dividing the total by number of spins. It is important to note that the

present distribution is highly skewed. Although the sample was ran-

domly distributed and assigned into their respective groupings, there

was large variability in the average bets placed by participants, espe-

cially by the males at the baseline. It is also important to note that the

mixed gender dyads (MF & FM) were collapsed into one group as

preliminary analyses indicated that there were no statistically signifi-

cant differences between these two groups.

In order to determine whether the participants' playing behavior

(i.e., risk) resulted from their success (i.e., winnings) in the task, sev-

eral analyses were performed on the participants' final bankrolls for

each condition (see Table 1).

Results of an analysis of variance indicated that there were no

significant differences for the final bankrolls of any of the groups in all

the trials. Further, paired samples t-tests indicated that the final bal-

ances of the various groupings did not significantly change over time.

Total Sample.

Results from the repeated measures analysis re-

vealed that a significant main effect was found for the average amount

Page 9

KAREN K. HARDOON AND JEFFREY L. DEREVENSKY

199

Table 1

Final Balances for the Different Groups Across Trials

Final Balances ($)

Baseline Trial

Group Trial

Post Trial

Group

Mean

587.80

726.43

351.65

294.07

560.75

565.84

MMM

474.14

239.50

458.48

293.70

400.57

301.91

499.44

147.87

465.69

150.14

357.94

253.06

FFF

425.22

133.44

439.61

168.79

459.39

170.97

MF & FM

428.69

200.88

471.12

212.84

422.06

246.97

MMF

444.57

193.58

432.14

219.80

498.52

241.65

FFM

543.11

142.26

426.72

276.81

428.56

234.14

bet across the baseline, group, and post games,

(2, 109) 4.34,

.01. More specifically, multivariate analyses indicated a significant

difference between average bets placed in the baseline, (

$51.34,

7.80) and the group (

$65.84,

7.55) conditions,

(1,

110) 4.08,

.05. However, no significant differences in average

wagers were found between the group (

$65.84,

7.55) and

post conditions (

$66.39,

6.38). It can be assumed that

there is a significant difference in average bets between the baseline

and the post games and since there were no significant differences

between the group and post conditions it appears as though the

change in betting behavior was maintained (Table 2).

Table 2

Average Amount Bet and Mean Differences Across Trials

Average Amount Bet ($)

Trials

Mean

Mean Difference

Baseline

51.34

7.80

Baseline to group

14.50 ( )

Group

65.84

7.55

Group to post

0.56 ( )

Post

66.40

6.38

Baseline to post

15.06 ( )

Page 10

200

JOURNAL OF GAMBLING STUDIES

Group Differences: Size and Gender.

Groupings were based upon

both size (dyad and triad) and gender (same sex and mixed sex) dif-

ferences. However, it remains difficult to separate the two factors since

they are integrated in the group category. As a result, both group size

and gender differences will be presented together.

With respect to overall gender differences, the data presented in

Table 3 clearly demonstrates that males' average bets are higher dur-

ing all 3 trials. It should be noted that females largely increased their

bets from the baseline to the group trial, and decreased them slightly

from the group to the post trial, while males increased their bets

slightly across the 3 trials.

Planned comparisons indicated that males and females differ sig-

nificantly from each other in their betting behavior during the base-

line,

(1, 110) 10.23,

.002, group,

(1, 110) 5.55,

.020,

and post-test,

(1, 110) 10.94,

.001 conditions. More specifi-

cally, males bet significantly more than females in all three trials for

both grades 4,

(1, 110) 6.63,

.011 and 6,

(1, 110) 5.35,

.023. These overall differences are depicted in Figure 1.

Furthermore, results indicated that females differed significantly

in their average bets placed across time,

(2, 109) 3.70,

.028.

Bonferroni-corrected planned comparisons indicated that females' av-

erage bets increased significantly from the baseline (

$26.38,

11.07) to the post condition (

$45.30,

9.05). No sig-

nificant differences in average bets across time were obtained for

males.

In order to determine if overall group differences existed, a multi-

variate analysis was performed. Results revealed that there were signifi-

cant group differences in the baseline,

(6, 110) 3.80,

.002;

Table 3

Average Amount Bet by Males and Females Across Trials

Females ($)

Males ($)

Trials

Mean

Baseline

28.25

29.65

82.79

109.30

Group

51.01

86.13

90.69

76.58

Post

42.97

45.38

92.42

86.19

Page 11

KAREN K. HARDOON AND JEFFREY L. DEREVENSKY

201

Figure 1

Average Bets Placed by Males and Females Across Trials

group,

(6, 110) 3.32,

.005; and post-test,

(6, 110) 5.00,

.001 conditions. No significant interaction between group and av-

erage amount bet was found.

Interestingly, a look at wagers across conditions demonstrates that

all groups increased their average bets from the baseline to the group

game with the exception of the MF/FM dyad (see Table 4). Further, it

appears as though the MM and MMM groups remain consistently high

in their betting across the baseline, group, and post-test games. While

the MM group maintains a steady increase from the group to the post-

test, the MMM group decreased their average bets slightly in the post-

test. With respect to female dyads and triads, females appear to peak

during the group game with the FF group increasing their average

bets dramatically.

The only group that decreased their average wagers from the

baseline to the group game was the combined MF/FM group. How-

ever, they increased their average bets in the post game. Interestingly,

with respect to the mixed gender triads, the MMF group has greater

average bets in the baseline and post games which may be due to male

bets raising the mean. However, from the baseline to the group condi-

tion, both the MMF and FFM groups experience an increase in their

average bets by virtually the same amount. Overall the males groups

had the greatest average wagers, with the females and mixed gender

Page 12

202

JOURNAL OF GAMBLING STUDIES

Table 4

Average Amount Bet by the Different Groups Across Trials

Average Amount Bet ($)

Baseline Trial

Group Trial

Post Trial

Group

Mean

92.98

101.01

99.33

90.57

107.06

97.83

MMM

108.35

161.25

112.59

87.02

107.06

105.87

39.89

46.97

84.48

157.72

64.47

68.30

FFF

27.01

23.52

36.70

30.45

24.87

15.41

MF & FM

38.38

46.29

27.41

17.93

57.58

46.05

MMF

47.43

47.36

70.85

44.57

69.95

57.55

FFM

31.49

19.79

59.96

51.57

41.08

33.37

groupings' average bets about equal (see Table 4 and Figures 2 and 3

which graphically depict these results).

Bonferroni-corrected planned pairwise comparisons revealed that

during the baseline, the MMM group differed significantly from the

FFM group, with male triads wagering significantly more than mixed

gender groups. Furthermore, in the group game, results indicated that

the MM group bet significantly more than the MF/FM group. Finally,

in the post game, the MMM group was found to make significantly

greater average bets than the FFM group. Interestingly no differences

were found between the MMM and the MMF groups, perhaps because

there are two males in this group who are contributing to the larger

average bets. Additionally, no differences were observed when compar-

ing the females to the mixed gender groupings in any of the trials.

With respect to differences in average bets wagered between dyads and

triads, no significant differences were found between the MM and

MMM's or the FF and FFF's. However, the mean wagers indicate that

the FF group increased by a greater amount from the baseline to the

group game.

Although statistically non-significant, some interesting trends were

observed when examining the difference in the average amount bet

between males and females in the different groups for the baseline,

group and post play (see Table 5).

Page 13

KAREN K. HARDOON AND JEFFREY L. DEREVENSKY

203

Figure 2

Average Bets Placed by the Various Groups Across Trials

As noted previously, all groups increased their average bets from

the baseline to the group game, with the exception of the MF/FM

dyad. It can be observed that males in this group are the individuals

who decreased their average wagers, thus, lowering the overall group

mean. It appears as though playing with a female inhibited males

Figure 3

Average Bets Placed by the Mixed Gender Groups Across Trials

Page 14

204

JOURNAL OF GAMBLING STUDIES

Table 5

Average Bets Placed by Males and Females in the Various Groupings

Across Trials

Average Amount Bet ($)

Baseline Trial

Group Trial

Post Trial

Group

Mean

Females

39.89

46.97

84.48

157.72

64.47

16.46

FFF

27.01

23.52

36.70

30.45

24.87

15.52

MF &

15.98

14.38

20.92

14.34

44.76

23.28

MMF

20.62

10.42

40.35

31.46

57.44

25.15

FFM

27.25

19.83

54.13

41.66

34.98

19.01

Males

92.98

101.02

99.33

90.57

107.06

97.83

MMM

108.35

161.25

112.59

87.02

107.06

105.87

MF &

60.78

56.90

33.91

19.67

70.41

53.57

MMF

60.84

53.11

86.10

42.98

78.91

59.65

FFM

39.97

18.34

71.62

70.56

53.28

40.93

(their average bets drop in half) in this dyad. However, between the

group to post game their wagers increased (Figure 4). Males are great

risk takers either when playing alone or with other males. However,

males in the FFM triad appear to have increased greatly from the base-

line to the group game. The fact that they are paired with two females

seems to increase their betting rather than inhibit it.

DISCUSSION

Peer Influences

As hypothesized, a significant increase in wagers was observed

from the baseline (individual play prior to group exposure) to the

group condition, and was maintained during the post-test (individual

Page 15

KAREN K. HARDOON AND JEFFREY L. DEREVENSKY

205

Figure 4

Average Bets Placed by Females and Males in the MF/FM Dyads

play following group exposure). One possible explanation for the find-

ing that wagering changed as a result of the group condition is that

betting naturally increases over time. The literature has demonstrated

that monetary risk-taking increases as a function of exposure to gam-

bling (Ladouceur & Mayrand, 1987) and that between gambling ses-

sions, risk-taking behavior does not return to the baseline, rather the

progressive effect is transferred to the next session (Ladouceur, May-

rand, & Tourigny, 1987). However, if this were the case then an in-

crease in average wagers from the group to the post game condition

would have also been found. Yet no significant overall differences were

observed between these 2 trials. Further, it is important to note that

while the possibility exists that greater average wagering could be at-

tributed to success (winning), results indicated that no significant dif-

ferences were found in the participants' final bankrolls and between

the groups in any of the three trials. Thus, since wagering did not

increase from the group to the post condition, and since no differ-

ences were found with respect to final bankrolls, it is likely that the

increase in betting from the baseline to the group game occurred as a

direct result of group exposure. Regardless of group (dyads and tri-

ads), the betting behavior of the entire sample significantly increased

from the baseline to the group game. Peers influenced participants'

Page 16

206

JOURNAL OF GAMBLING STUDIES

playing behavior and this effect was maintained, at a minimum, over a

relatively short period of time (e.g., post-test). These findings provide

preliminary support for the hypothesis that peers serve as models for

gambling behavior, such that bets increased from the baseline to the

group game (i.e., children's betting changed as a result of group expo-

sure). Furthermore, it appeared as though peers were trying to im-

press one another when playing by increasing wagers during group

play.Previous studies have demonstrated that gambling activities are

reinforcing to youth because they produce excitement, and arousal,

and are entertaining (Derevensky et al., 1996). Moreover, gambling

behaviors can be encouraged and strengthened by one's peers de-

pending on developmental level and social status (Gupta, 1994) and

may be considered a "rite of passage" into adulthood (Stinchfield &

Winters, 1998). It appears that in the present study some type of mod-

eling/peer influence is involved in increasing the participants' betting

behavior, such that increased wagering may lead to recognition as a

"gambler" among peers in the group condition, resulting in social rec-

ognition (Fisher, 1995; Opie & Opie, 1969; Smith & Abt, 1984). Fur-

ther, if the principles of social learning theory are applied directly to

gambling behavior, the question still remains as to whether children

can be influenced by their peers simply by observing them engage in a

gambling activity, by being part of a peer group which recognizes gam-

bling and risk-taking as socially acceptable, or if an interactive group

condition is necessary.

Gender Differences

As expected, the results indicated that males placed significantly

higher wagers than females in all three conditions. These findings are

consistent with previous research with children indicating that males

make higher gross wagers and have greater gross winnings than fe-

males, suggesting they are exhibiting greater risk-taking behaviors (De-

revensky et al., 1995; 1996). Gambling has also been found to be more

popular amongst males than females (Fisher, 1993; Ide-Smith & Lea,

1988; Ladouceur et al., 1994a; 1994b; Lesieur et al., 1991; Rosenstein

& Reutter, 1980; Stinchfield & Winters, 1998). This gender difference

was noted for children in both grades 4 and 6. Further, the overall

betting patterns of males remained consistent across trials, while fe-

Page 17

KAREN K. HARDOON AND JEFFREY L. DEREVENSKY

207

males' average wagers increased significantly over time. This is of par-

ticular concern as it appears that females are more susceptible to

group influences and/or multiple exposure to gambling activities.

Group Differences: Size and Gender

Group Size Differences.

One of the hypotheses of the present study

was that group size would have an effect on wager patterns, such that

individuals in larger groups (triads) would have greater mean bets as

compared with dyads. Results indicated that no significant differences

in wagers were found between the male dyads and triads (MM and

MMM) or between the female dyads and triads (FF and FFF). How-

ever, it must be noted that while not statistically significant, the FF

group largely increased their average wagers during the group condi-

tion (much more so than female triads). It appears as though female

dyads exhibit increased risk-taking behavior, as measured by total

amount wagered, in the group condition. It is possible that the FF

group's increase from the baseline to the group game did not reach

statistical significance due to intra-group variability of mean bets in the

group trial. Results of the present study seem to indicate that simply

gambling with another child serves to increase wagering behavior. The

findings do not support the hypothesis that having more children in a

group (triad) would have a greater effect on modifying pre-existing

gambling behavior (wagering). Essentially, the inclusion of one peer

appears to be a sufficient influence for increased betting behavior

amongst female dyads. However, additional research needs to be con-

ducted in order to further investigate the effects of the female dyads

and differences between dyads and triads.

Group Differences.

Analyses of group differences (i.e., how groups

changed over time) revealed that all groups increased their average

bets from the baseline to the group game, with the exception of the

MF/FM dyads. Interestingly, compared to all other groups, the male

dyads (MM) and triads (MMM) changed the least over time and

placed significantly higher wagers than female and mixed gender

dyads and triads across all trials and in both grades 4 and 6. Female

dyads and triads increased their wagers from the baseline to the group

game. However, as noted, when discussing size differences in the pre-

sent sample, female dyad groups (FF) increased their average bets dra-

Page 18

208

JOURNAL OF GAMBLING STUDIES

matically in the group game, denoting that female dyads appear to

induce greater wagering behavior than female triads, who do not seem

to have the same magnitude of effect (in terms of increasing their

average wagers). In addition, while both female dyads (FF) and triads

(FFF) responded to peer influences by increasing their average wa-

gers, both groups' average bets decreased during the post-test. Thus, it

appears that when females play individually in the third trial (post-

test), they return to more conservative betting patterns which were

observed during the baseline. Research, in general, on young chil-

dren's play styles suggests that females tend to play in dyads, while

males tend to play in larger groups (Benenson, 1993; Benenson, Apos-

toleris, & Parnass, 1997). The finding that female dyads risked more

money as a result of the group condition is worrisome, such that when

engaging in gambling activities together, females are possibly in jeop-

ardy of exhibiting increased risk-taking behavior. However, the finding

that females return to their regular betting patterns following the

group condition indicates that changes resulting from the group game

were not maintained. These findings appear to be encouraging for

female dyads given that the social influences which occur in the group

game do not appear to be generalized to individual play (post-test).

More research needs to be conducted in order to ascertain whether

female dyads are at heightened risk when gambling in dyads and

whether social influences are in fact maintained over time or gener-

alized in other gambling venues.

Interestingly, no differences in wagering were observed when

comparing the all female dyads and triads (FF & FFF) to the mixed

gender groupings (MF/FM, FFM, MMF) during any of the trials.

These findings partially confirm the hypothesis that groups involving

males would exhibit greater risk-taking behavior since findings were

specific to the all male groups (MM & MMM) and not to mixed gen-

der groups involving both males and females.

Given that there were two males and one female in the MMF

group, it was thought that this triad collectively would have larger wa-

gers than the FFM triad (irrespective of the individual performances of

the males and females), since males were found to exhibit greater wa-

gers overall. However, it was observed that both the MMF and FFM

triads (as a group) increase their average bets from the baseline to the

group game by almost the same amount. Interestingly, the MMF group

maintained the increase in the post-test, while the FFM group did not.

Page 19

KAREN K. HARDOON AND JEFFREY L. DEREVENSKY

209

Gender Differences by Group.

As previously noted, the MF/FM dyad

was the only group to decrease their average wagers from the baseline

to the group game. The results suggest that it was actually the males in

this group who decreased their bets as a result of group exposure

while females slightly increased their average wagers. Thus, it appears

as though playing exclusively with one female somehow inhibits males'

wagering. Nevertheless, the males' wagers increase in the post-test and

become larger than their baseline wagers. Males in the FFM triad ap-

pear to increase greatly from the baseline to the group game. Further-

more, while males in the MMF triad exhibited larger mean bets during

both the baseline and group games than males in the FFM triad, their

wagers did not make as large an amount as males in the FFM triad.

Thus, it appears as though being paired in a triad with one or two

females (MMF, FFM) seems to increase males' betting behavior rather

than inhibit it. Interestingly, males in both groups (MMF/FFM) do not

maintain increased wagers in the post-test. Perhaps an element of

competition is involved in the FFM triad, when one male is paired with

two females, accounting for the larger mean differences in the males'

average bets (compared to the MMF group). The male may feel as

though he wants to exert his masculinity by wagering more. Griffiths

(1989) hypothesized that gambling allows boys to display their mas-

culinity in a social environment by exhibiting "courage and bravery,"

and may provide an explanation for the present findings. Clinical ob-

servations of several FFM triads confirm this speculation, such that the

females form their own dyad and exclude the male. The female dyads

then compete against the male even though each individual has their

own bankroll. However, observation of MMF triads demonstrates much

less competition and much more cooperation. With respect to females

in the MF/FM dyad, they were found to increase their average wagers

slightly from the baseline to the group game, and much more so in

the post condition. In the MMF triad, females were observed to lin-

early increase their average wagers across the three trials. Finally, fe-

males in the FFM triad appear to be most affected by the group condi-

tion (increased wagers) and did not maintain any changes in the post-

test. These findings are further consistent with the clinical observa-

tions of the FFM triads, such that the girls appear to form their own

distinct group.

The aforementioned findings regarding group differences did not

provide support for the hypothesis that groups involving males would

Page 20

210

JOURNAL OF GAMBLING STUDIES

generate greater group influences. Given that no changes in average

wagers were observed in the male dyads and triads (MM & MMM)

across the three trials, and considering the finding that females in-

creased their average wagers playing in all groupings (i.e., same sex

and mixed sex dyads and triads), support is not provided for the hy-

pothesis that males are more likely than females to influence other

males and females during gambling activities. Furthermore, the find-

ings suggest that females appear to be more susceptible to social influ-

ences. Thus, for females, simply playing in a group, regardless of gen-

der, seems to increase wagering. The finding that males do not appear

to exert an influence on each other may indicate that they are already

high wagerers by nature (i.e., not only in the group condition) and are

not as susceptible to external influences. While females paired with

males in the mixed gender groups (MF/FM, MMF, and FFM) exhibit

an increase in their average bets over time, their wagers never reach

the magnitude of males' wagers. Thus, while females may be more

susceptible to group influences, the results indicate that they neverthe-

less remain more conservative than males. These findings again pro-

vide support for previous literature which has found that males dem-

onstrate greater wagers than females, suggesting that they are

exhibiting greater risk-taking behaviors (e.g., Derevensky et al., 1995;

1996).

The present findings suggest that some type of peer/social influ-

ence is involved in gambling behavior. Results have demonstrated that

changes in average wagers from the baseline to the group game were

found between and within the experimentally predetermined groups.

The predominant finding of the present research program was that

increased wagering resulted from playing in a group with one or more

children. Males were found to exhibit greater wagering behavior over-

all, while females were found to be most susceptible to external influ-

ences during the group game such that they demonstrated increases

in their average bets.

The present findings represent one piece of the puzzle with re-

spect to social influences involved in gambling behavior. There are

likely other factors which may influence gambling behaviors as well as

limitations inherent in the present study. For example, it is unknown

whether children's actual gambling behavior was changed merely as a

result of the group condition. Moreover, the present study was con-

Page 21

KAREN K. HARDOON AND JEFFREY L. DEREVENSKY

211

ducted in a laboratory setting and only one game (roulette) was em-

ployed. In addition, many researchers have questioned the use of in-

centives, rather than money in gambling research. However, this type

of incentive has proven successful with children (Derevensky et al.,

1996; Gupta & Derevensky, 1996) and adults (Powell et al., 1996) and

it appears as though money does not hold as great a value with chil-

dren as other incentives (e.g., candy, toys, and movie certificates). Al-

though the present sample was randomly distributed and assigned into

their respective groupings, there was huge variability in the average

bets placed by the participants, especially by the males at the baseline.

Other limitations of the present study include the fact that there

was no control group and that the post-test was played immediately

following the group game. In addition, some children may have bet

more money in the baseline, and thus had greater average wagers,

because they knew that they had two more additional trials and may

have reasoned that they could "afford" to take greater risks. Only fur-

ther research can determine what occurs when children play with their

peers during actual naturalistic and varying gambling activities and

can be aimed at clarifying the social influences involved in gambling

activities.

By developing a better understanding of individual and social fac-

tors which influence gambling behavior in children and adolescents,

effective prevention and intervention programs can be developed and

can target the appropriate age and at-risk groups. The development of

prevention and intervention programs are greatly needed, as gambling

is the most frequently reported potentially addictive behavior engaged

in by children and adolescents (Gupta & Derevensky, 1998; 1997).

APPENDIX A: STANDARDIZED INSTRUCTIONS

General Instructions

You will be playing one game against the dealer in order to win a

prize. First you will be playing by yourself for a few minutes, then you

will be playing with some of your friends, and then you will be playing

alone again. For each game we are going to give you $500 of fake

money to play with. When you play alone, you will be playing for 7

Page 22

212

JOURNAL OF GAMBLING STUDIES

minutes. In that time, you should try to win as much money as you can

because for all the money that you make over the $500 you will earn

the equivalent in monopoly money that we will put into a draw. The

draw is for 3 prizes, you can either win $30 in movie certificates, $20

and $10 gift certificates at McDonalds. The more money that you win

the better the chance you have of winning a prize. For example, if at

the end of 7 minutes you have $550 then we will give you $50 in mo-

nopoly money which will be put into this box for the drawing of $10 in

McDonald's certificates, but if you have $450 we can't give you any

money, you have to have more than $500. You don't have to play the

whole game, you can stop whenever you want and take your money

and put it into the draw. Remember, the more money you have the

better your chances of winning one or more of the prizes. This is the

same for when you will be playing with your friends, when you will be

playing for about 1214 minutes.

Individual Instructions

You will be playing a casino game called roulette. In this game, you are

betting on what number or color the little ball will land on after the

wheel is spun. For this game you can only bet

evens, odds, black, and/or

red

. For example, if you have bet black, you will only win if the number

that the ball lands on is black. To bet, you place the chip on the right

spot. The chips can be worth

$1, $5, $25,

and

$100

. You can make the

wheel spin faster by moving the arrow near the wheel and clicking the

left mouse button. But this doesn't change anything. To make a bet

click on play, then you click on whatever amount you want your chips

to be. Then click on the table in the spot where you want to put your

bet. The more times you click, the more chips you are putting on the

table. So if you choose $5 chips and click 3 times on red then you are

betting $15 on red. If you make a mistake and want to take off a chip

then click on the right mouse button. If you win, the money that you

win will be left on the table, to pick them up click

cash in

Don't forget that your goal is to end up as much money over the $500

dollars we lent you so you can win as much money as you can for the

draw. You have 7 minutes to play, or until you run out of money. Don't

forget that you can stop at any time.

Page 23

KAREN K. HARDOON AND JEFFREY L. DEREVENSKY

213

Group Instructions

You have all played this game before by yourselves. Now you will be

playing it together, with (another child/ or two other children). Each

of you will take a turn placing your bet before we spin the wheel.

Remember that you are playing against the dealer and not against

each other.

You can talk and help each other out if you want

. You will each

have $500 in your bank to play with and you will play for 12

(dyad)

/14

(triad)

minutes. You can win more tickets in this game, if you end up

with more than $500 at the end of the game. Remember, for all the

money over $500 you will "win" the equivalent in monopoly money to

enter into the draw. You can also stop whenever you want.

REFERENCES

Arcuri, A. F., Lester, D., & Smith, F. O. (1985). Shaping adolescent gambling behavior.

Adolescence,

20,

935938.

Bandura, A. (1977).

Social learning theory.

Englewood Cliffs, NJ: Prentice-Hall.

Bandura, A. (1973).

Aggression: A social learning analysis.

Englewood Cliffs, NJ: Prentice-Hall.

Bandura, A., Ross, D., & Ross, S. (1963). A comparative test of the status envy, social power, and

secondary reinforcement theories of identificatory learning.

Journal of Abnormal and Social Psy-

chology, 67,

527534.

Benenson, J. F. (1993). Greater preference among females than males for dyadic interaction in

early childhood.

Child Development, 64,

544555.

Benenson, J. F., Apostoleris, N. H., & Parnass, J. (1997). Age and sex differences in dyadic and

group interaction.

Developmental Psychology, 33,

538543.

Custer, R. L. (1982). An overview of compulsive gambling. In P. A. Carone, S. F. Yolles, S. N.

Kieffer, & L. W. Krinsky (Eds.),

Addictive disorders update

(pp. 107124). New York: Human

Sciences Press.

Dell, L. J., Ruzicka, M. E., & Palisi, A. T. (1981). Personality and other factors associated with

gambling addiction.

International Journal of the Addictions, 16,

149156.

Derevensky, J. L., & Gupta, R. (in press). Youth gambling: A clinical and research perspective.

The

Electronic Journal of Gambling Issues.

Derevensky, J. L., & Gupta, R. (1996, October).

Familial and social influences associated with children's

early gambling behavior.

Paper presented at the Interprovincial Conference on Problem Gam-

bling, Winnipeg, Canada.

Derevensky, J. L., Gupta, R., & Della-Cioppa, G. (1996). A developmental perspective of gambling

behavior in children and adolescents.

Journal of Gambling Studies, 12

(1), 4966.

Derevensky, J. L., Gupta, R., & ´Emond, M. (1995).

Locus of control, video game playing and gambling

behavior in children and adolescents.

Poster session presented at the American Psychological Asso-

ciation Annual Convention, New York.

Fisher, S. (1993). Gambling and pathological gambling in adolescents.

Journal of Gambling Studies, 9

(3), 277288.

Fisher, S. (1995). The amusement arcade as a social space for adolescents: An empirical study.

Journal of Adolescence, 18,

7186.

Griffiths, M. D. (1989). Gambling in children and adolescents.

Journal of Gambling Behavior, 5,

83.

Griffiths, M. D. (1990). The acquisition, development, and maintenance of fruit machine gambling

in adolescents.

Journal of Gambling Studies, 6,

193204.

Page 24

214

JOURNAL OF GAMBLING STUDIES

Gupta, R. (1994).

The relationship between video game playing and gambling behavior in children and

adolescents.

Master's thesis, McGill University, Montreal, Quebec, Canada.

Gupta, R., & Derevensky, J. L. (1998). Adolescent gambling behavior: A prevalence study and

examination of the correlates associated with problem gambling.

Journal of Gambling Studies, 14

(4), 319345.

Gupta, R., & Derevensky, J. L. (1997). Familial and social influences on juvenile gambling behavior.

Journal of Gambling Studies, 13

(3), 179192.

Gupta, R., & Derevensky, J. L. (1996). The relationship between gambling and video-game playing

behaviour in children and adolescents.

Journal of Gambling Studies, 12

(4), 375394.

Huxley, J., & Carroll, D. (1992). A survey of fruit machine gambling in adolescents.

Journal of

Gambling Studies, 8,

167179.

Ide-Smith, S. G., & Lea, S. E. (1988). Gambling in young adolescents.

Journal of Gambling Behavior, 4

(2), 110118.

Jacobs, D. F., Marston, A., & Singer, R. (1985). Testing a general theory of addiction: Similarities

and differences among alcoholics, pathological gamblers, and overeaters. In J. J. Sanchez-Soza

(Ed.),

Health and clinical psychology

(Vol. 4). Netherlands: Elsevier Science.

Kearney, C. A., & Drabman, R. S. (1992). Risk-taking/gambling-like behavior in preschool chil-

dren.

Journal of Gambling Studies, 8

(3), 287297.

Ladouceur, R., Boisvert, J. M., P´epin, M., Loranger, M., & Sylvain, C. (1994). Social cost of patho-

logical gambling.

Journal of Gambling Studies, 10

(4), 399409.

Ladouceur, R., Dub´e, D., & Bujold, A. (1994a). Prevalence of pathological gambling and related

problems among college students in the Quebec metropolitan area.

Canadian Journal of Psychi-

atry, 39,

289293.

Ladouceur, R., Dub´e, D., & Bujold, A. (1994b). Gambling among primary school students.

Journal

of Gambling Studies, 10

(4), 363370.

Ladouceur, R., & Mayrand, M. (1987). The level of involvement and the timing of betting in

gambling.

Journal of Psychology, 121,

169175.

Ladouceur, R., Mayrand, M., & Tourigny, Y. (1987). Risk-taking behavior in gamblers and non-

gamblers during prolonged exposure.

Journal of Gambling Behavior, 3,

115122.

Ladouceur, R., & Mireault, C. (1988). Gambling behaviors among high school students in the

Quebec area.

Journal of Gambling Behavior, 4

(1), 312.

Ladouceur, R., & Walker, M. (1996). A cognitive perspective on gambling. In P. M. Salkovskis

(Ed.),

Trends in cognitive behavioural therapies

(pp. 89120). New York: Wiley.

Lesieur, H. R. (1998). Costs and treatment of pathological gambling.

Annals of the American Academy

of Political and Social Sciences, 556

, 153171.

Lesieur, H. R., Cross, J., Frank, M., Welch, M., White, C. M., Rubenstein, G., Moseley, K., & Mark,

M. (1991). Gambling and pathological gambling among university students.

Addictive Behav-

iors, 16,

517527.

Lesieur, H. R., & Klein, R. (1987). Pathological gambling among high school students.

Addictive

Behaviors, 12,

129135.

Lesieur, H. R., & Rosenthal, R. J. (1991). Pathological gambling: A review of the literature (pre-

pared for the American Association Task Force on DSM-IV committee on disorders of impulse

control not elsewhere classified).

Journal of Gambling Studies, 7

(1), 439.

Lorenz, V. C., & Shuttlesworth, D. E. (1983). The impact of pathological gambling on the spouse

of the gambler.

Journal of Community Psychology, 11,

6776.

Masteller, R., & Masteller, B. (1993). Ceasar's palace for windows [computer software], Irvine, CA:

Virgin Games.

Moore, S. M., & Ohtsuka, K. (1997). Gambling activities of young Australians: Developing a model

of behaviour. Unpublished Manuscript.

National Opinion Research Center at the University of Chicago. (February, 1999).

Gambling impact

and behavior study: Report to the National Gambling Impact Study Commission.

Chicago, IL: National

Opinion Research Center at the University of Chicago.

Opie, I., & Opie, P. (1969).

Children's games in street and playground.

Oxford: Oxford University

Press.

Powell, G. J., Hardoon, K., Derevensky, J., & Gupta, R. (1999). Gambling and risk taking behavior

of University students.

Substance Use and Misuse, 34

(8), 11671184.

Page 25

KAREN K. HARDOON AND JEFFREY L. DEREVENSKY

215

Rosenstein, R. J., & Reutter, R. (1980). Gambling: An adolescent activity.

Journal of Adolescent Health

Care, 1,

180.

Rushton, J. P. (1980).

Altruism, socialization, and society.

Englewood Cliffs, NJ: Prentice-Hall.

Santrock, J. W. (1990).

Children.

Dubuque, IA: Wm. C. Brown.

Shaffer, H. J., & Hall, M. M. (1996). Estimating the prevalence of adolescent gambling disorders: A

quantitative synthesis and guide toward standard gambling nomenclature.

Journal of Gambling

Studies, 12

(2), 193214.

Shaffer, H. J., LaBrie, R., Scanlan, K. M., & Cummings, T. N. (1994). Pathological gambling among

adolescents: Massachusetts gambling screen.

Journal of Gambling Studies, 10

(4), 339362.

Smith, J. F., & Abt, V. (1984). Gambling as play.

Annals of the American Academy of Political and Social

Sciences, 474,

122132.

Staub, E. (1978).

Positive social behavior and morality

(Vol. 1.). New York: Academic Press.

Stinchfield, R., & Winters, K. C. (1998). Gambling and problem gambling among youth.

Annals of

the American Academy of Political and Social Sciences, 556,

172185.

Volberg, R. A. (1994). The prevalence and demographics of pathological gamblers: Implications

for public health.

American Journal of Public Health, 84,

237241.

Volberg, R. A., & Steadman, H. J. (1988). Refining prevalence estimates of pathological gambling.

American Journal of Psychiatry, 145,

502505.

Winters, K. C., Bengston, P, Door, D., & Stinchfield, R. (1998). Prevalence and risk factors of

problem gambling among college students.

Psychology of Addictive Behaviors, 12,

127135.

Winters, K. C., Stinchfield, R., & Fulkerson, J. (1993). Patterns and characteristics of adolescent

gambling.

Journal of Gambling Studies, 9,

371386.

Wynne, H., Smith, G., & Jacobs, D. (1996).

Adolescent gambling and problem gambling in Alberta.

report prepared for the Alberta Alcohol and Drug Abuse Commission. Edmonton, Alberta:

Wynne Resources Ltd.

Received October 1, 2000; final revision April 20, 2001; accepted May 11, 2001.