Excitement-seeking gambling in adolescents: Health correlates and gambling-related attitudes and behaviors

Abstract

Motivational characteristics such as excitement-seeking are key components of models of addiction, including problem gambling. Previous studies have established associations between excitement-seeking and problem gambling in youth. However, these studies have employed dimensional psychological assessments which are unlikely to be routinely administered. Other approaches to conceptualize excitement-seeking could be of value. In the present study, we employed a single question (What are the reasons that you gamble?) to identify adolescents who reported excitement-seeking motivation for gambling. Cross-sectional data from 2030 adolescent gamblers who participated in a Connecticut high-school survey were examined. Gambling perceptions and correlates of problem-gambling severity were examined relative to excitement-seeking and non-excitement-seeking gambling. Gambling perceptions were more permissive and at-risk/problem gambling was more frequent among adolescents with excitement-seeking gambling versus non-excitement-seeking gambling. A weaker relationship between problem-gambling severity and moderate and heavy alcohol use was observed for excitement-seeking versus non-excitement-seeking gambling. Excitement-seeking gambling is associated with more permissive gambling-related attitudes and riskier gambling behaviors and may account for some variance in adolescent risk of heavy alcohol use. A single question may provide important information for identifying adolescents who are at elevated risk of problem gambling and associated negative outcomes, although the utility of the question in specific settings warrants direct examination, especially given the observed high prevalence of excitement-seeking motivations for gambling.

INTRODUCTION

Gambling is a common behavior among adolescents in the US (National Research Council 1999; Welte, Barnes, Tidwell & Hoffman 2008) and is associated with negative health and psychiatric outcomes, such as alcohol and drug use and depression (Lynch, Maciejewski & Potenza 2004). Adolescents are susceptible to developing persistent, recurrent patterns of gambling that are associated with significant distress and impairment (Chambers & Potenza, 2003). Although the Diagnostic and Statistical Manual (DSM-5) (American Psychiatric Association, 2013) recognizes such disruptive gambling with the term gambling disorder, the term problem gambling has been widely employed to refer to individuals who experience a wide range of gambling-related problems without perhaps formal diagnostic criteria (Potenza, Balodis, Derevensky, Grant, Petry & Vardejo-Garcia 2019). Previous research indicated that at-risk problem gambling (ARPG), in comparison to low risk gambling (LRG), is linked to negative health correlates such as poor school performance, depression and use of substances (tobacco, alcohol, cannabis and other drugs) among adolescents (Rahman et al. 2014; Weinberger et al. 2015; Yip et al. 2011).

Adolescents appear particularly vulnerable to developing problem gambling as estimated rates of problem gambling among adolescents – between 0.2% and 12.3% - are higher than those among adults – between 0.1% and 5.8% (Calado, Alexandre & Griffiths 2017; Calado & Griffiths 2016). It is possible that adolescents’ vulnerability to problem gambling is associated with decision-making and goal-directed behavior that is influenced by impulsivity and excitement-seeking during adolescence (Arnett 1992; Chambers, Taylor & Potenza 2003; Chambers & Potenza 2003; Greene, Krcmar, Walters, Rubin, Jerold & Hale 2000). Excitement-seeking, impulsive adolescents may be less responsive to natural rewards and less likely to postpone immediate gratification for larger payouts. In this respect, these adolescents may be prone to engage repetitively in behaviors that generate short-term rewards (e.g. substance use and gambling) at the expense of longer-term achievements, with some experiencing addictions and related negative consequences. Consistently, cross-sectional (Canale, Vieno, Griffihs, Rubaltelli & Santinello 2015; Nower, Derevensky & Gupta 2004; Vitaro, Arseneault & Tremblay 1999) and prospective (Dowling, Merkouris, Greenwood, Oldenhof, Toumbourou & Youssef 2017) studies have reported associations between excitement-seeking and impulsivity and problem gambling among adolescents.

Although the relationship between excitement-seeking and adolescent problem gambling is well documented, research has frequently considered excitement-seeking as a global trait evaluated by dimensional scales. Although this strategy is important, other approaches may also hold relevance, particularly since psychological assessments of excitement-seeking and impulsivity are not routinely administered to adolescents in school or clinical settings. A useful and simple approach may involve asking about excitement-seeking gambling (EG). Previously, a categorical approach was used to investigate the relationship between EG and other co-occurring behavioral problems in an adult sample (Pantalon, Maciejewski, Desai & Potenza 2008). In the study, individuals were classified as having EG if they acknowledged “excitement” or “challenge” as being either “important” or “very important” reasons to gamble. Currently, there are no similar studies evaluating whether a similar categorical approach could be useful in identifying adolescents who might be at increased risk of problem gambling and other co-occurring concerns.

Such a categorical approach could be a valuable tool to help identify youth at-risk of problem gambling. Although adolescent problem gambling is considered a public health problem (Messerlian, Derevensky & Gupta 2005), parents and teachers infrequently identify gambling problems in adolescents. Adolescents typically gamble with small amounts of money and therefore do not report large wins/losses (Wilber & Potenza 2006). Negative outcomes associated with adolescent problem gambling (e.g., poor academic performance and conflicts with parents) are also relatively non-specific. and therefore parents and teachers may associate them with normal development or other behavioral problems (e.g. alcohol use). Parents and teachers are unlikely to attribute possible behavioral issues to adolescent problem gambling as both parents and teachers usually do not consider gambling a serious problem among adolescents (Campbell, Derevensky, Meerkamper & Cutajar 2011; Derevensky, St-Pierre, Temcheff & Gupta 2014). Therefore, adolescent problem gambling frequently goes unnoticed (Derevensky 2012; Potenza, Balodis, Derevensky, Grant, Petry & Vardejo-Garcia 2019), and further research is required to aid in the identification of ARPG youth.

To address this gap, we interrogated a large sample of adolescent gamblers to examine differences in sociodemographic characteristics, gambling-related perceptions, health/functioning measures and gambling behaviors according to adolescents’ status as having EG or non-excitement-seeking gambling (NEG). We defined the two gambling groups according to adolescents’ self-reported motivation to gamble for “excitement”. We hypothesized that having EG would be positively associated with ARPG and more permissive views towards gambling. We also hypothesized the EG relative to the NEG group would demonstrate weaker relationships between problem-gambling severity and health/functioning measures (e.g., substance use) and gambling behaviors (e.g., more time spent gambling) given that EG would account for some of the variance in the relationships between these measures.

METHODS

Participants

The high-school survey and methodologies are similar to those in our previous publications (Desai, Krishnan-Sarin, Cavallo & Potenza 2010; Kundu et al. 2013; Leeman et al. 2014; Liu, Desai, Krishnan-Sarin, Cavallo & Potenza 2011; Potenza et al. 2011; Rahman et al. 2014; Slavin et al. 2013; Weinberger et al. 2015; Yip et al. 2011). Briefly, an invitation letter was sent by mail to all public four-year and non-vocational or special-education high-schools in the state of Connecticut; follow-up calls were also made to schools’ principals. Of 122 schools invited, 10 participated. As the initial response did not ensure representation of all geographical regions in Connecticut, additional targeted recruitment was conducted. Ultimately, the survey included schools from all geographical regions of the state of Connecticut as well as from each of the three tiers of the state’s district reference groups, which are clusters of schools based on familial socioeconomic status. Although this was not a random sample of high schools, the sample obtained was demographically consistent with the 2000 Census data of 14- to 18-year-old Connecticut residents. The high-school survey data were collected in 2006.

Measures

The survey questionnaire included questions about sociodemographics (e.g., age, gender, race/ethnicity, grade level and family structure), academic (e.g., high-school grade averages and engagement in extracurricular activities), substance use (lifetime cigarette smoking, marijuana, alcohol and other drug use as well as current alcohol use and caffeine use), dysphoria/depression (past-year sadness or hopelessness for ≥ 2 weeks), and violence (carrying a weapon and getting involved with a serious fight). Questions also inquired about types of gambling (strategic, non-strategic, machine), age-of-onset of gambling (≤8 years old, 9–11 years old, 12–14 years old, ≥ 15 years old), frequency of gambling per week (≤ 1 hour, ≥ 2 hours), people with whom adolescents gambled (family, friends, other adults, strangers, alone), and gambling perceptions (i.e. the perceived importance of several gambling prevention approaches). These variables were categorized as depicted in the tables.

Problem-gambling severity was evaluated through the use of the 12-item Massachusetts Gambling Screen (MAGS) (Shaffer, Labrie, Scanlan, Cummings 1994), a validated instrument to evaluate DSM-IV pathological-gambling criteria among adolescents. Motivations to gamble were evaluated through one item asking, What are the reasons that you gamble? Check ALL that apply. Adolescents could check, (1) fun and entertainment, (2) excitement, (3) boredom, (4) to win money, (5) it’s a challenge, (6) to calm down, (7) to feel good about myself (e.g. feel like a winner), (8) to socialize with friends, (9) peer pressure (e.g. to fit in), (10) to support good causes, (11) it’s a hobby and (12) as a distraction from my problems.

Procedures

Adolescents who met at least one DSM-IV pathological-gambling criterion as assessed by the MAGS were classified as having ARPG while those who met no criteria were considered as having low-risk gambling (LRG). Adolescents who checked (2) excitement as an answer to the question, What are the reasons that you gamble? were classified as EG, while those who did not were classified as NEG.

Of the 4,523 students who participated in the survey, 2,030 were included in analyses. Individuals were excluded for the following reasons: (1) did not provide valid data for past-year gambling behavior (n = 168); (2) did not gamble over the past year (n = 454); and (3) did not provide answers to all the 12 items of the MAGS (n = 1,871).

Statistical analysis

Data were double-entered into an electronic system; random spot-checks of completed surveys and data-cleaning procedures were performed to ensure data were accurate and within range. All statistical analyses were conducted using the SAS system (Cary, NC). Several steps were employed to evaluate associations between EG and sociodemographic characteristics, gambling-related perceptions, health/functioning measures and gambling behaviors. Between-group differences (EG vs NEG) were examined using Pearson chi-square tests. All comparisons were two-tailed; a Bonferroni correction was applied and p-values of p < 0.0025 were regarded as significant. Next, regression models were constructed to provide odds ratios (ORs) with 95% confidence intervals (CIs) to measure magnitudes of associations. Logistic regression models and multinomial logistic regression models were employed for binary and categorical outcomes, respectively. All models were adjusted for gender, race/ethnicity, grade level and family structure.

Associations between problem-gambling severity and health/functioning measures and gambling behaviors stratified by the two gambling groups were evaluated using similar procedures as described above (e.g. Pearson chi-square, regression models). To determine whether EG moderated associations between problem-gambling severity and health/functioning measures and gambling behaviors, main effects for EG and ARPG, as well as the interaction term (excitement-seeking-status-by-problem-gambling-severity), were included in the appropriate logistic or multinomial logistic regression models.

RESULTS

Of the 2,030 adolescents included in analyses, 1,360 (67 %) reported “excitement” as a reason why they gambled and were classified as EG; the remaining 670 were classified as NEG. EG status was associated with problem-gambling severity (χ² = 100.76, p < 0.0001); ARPG was more common among EG in comparison to NEG (41.40% vs 18.96%).

Sociodemographic characteristics

Sociodemographic data are shown in Table 1. The EG group was predominantly male (68.16%) and white (74.19%), lived with two parents (70.52%) and reported grade average of A’s and B’s (55.37%). In comparison to NEG, EG was associated with male gender (χ² = 91.32, p < 0.0001) and white race/ethnicity (χ² = 10.96, p = 0.0009).

Table 1.

Sociodemographic characteristics stratified by excitement-seeking status

	EG		NEG		χ2 Statistics
	N	%	N	%	χ2	p
Gender					91.32	< 0.0001
Male	914	68.16	304	45.99
Female	427	31.84	357	54.01
Race/ethnicity African-American/black					6.14	0.0132
Yes	140	10.29	94	14.03
No	1220	89.71	576	85.97
Caucasian/white					10.96	0.0009
Yes	1009	74.19	450	67.16
No	351	25.81	220	32.84
Asian American					0.48	0.4888
Yes	64	4.71	27	4.03
No	1296	95.29	643	95.97
Hispanic					1.93	0.1647
Yes	203	15.77	118	18.27
No	1084	84.23	528	81.73
Other					5.46	0.0195
Yes	208	15.29	130	19.40
No	1152	84.71	540	80.60
Grade					6.11	0.1066
9th	415	30.60	197	29.58
10th	335	24.71	198	29.73
11th	352	25.96	159	23.87
12th	254	18.73	112	16.82
Family structure					5.11	0.0777
One parent	302	22.54	175	26.84
Two parents	945	70.52	428	65.64
Other	93	6.94	49	7.52
Grade average					9.38	0.0092
A’s and B’s	753	55.37	334	49.85
Mostly C’s	436	32.06	221	32.99
D’s or lower	171	12.57	115	17.16

EG: Excitement gambling; NEG: Non-excitement gambling

Gambling perceptions

Gambling-perception data are shown in Table 2. EG was associated with more permissive views towards gambling on items covering parental oversight. More EG relative to NEG respondents classified as unimportant the following gambling prevention approaches: (1) parental strictness about gambling (25.57% vs 19.06%, χ² = 9.58, p = 0.002), (2) having parents who do not gamble (27.27% vs 18.26%, χ² = 13.91, p = 0.0002) and (3) parents not permitting card games (for money) at home (42.76% vs 29.68%, χ² = 29.17, p < 0.0001). The EG (vs the NEG) group also reported more frequently having parents who were either neutral or positive towards gambling (χ² = 66.27, p < 0.0001). Lastly, the EG (vs the NEG) group was less likely to see as important learning the risks of gambling in school (χ² = 14.45, p = 0.0001) and advertisements showing problems associated with gambling (χ² = 10.65, p = 0.0011).

Table 2.

Gambling perceptions stratified by excitement-seeking status

	EG		NEG		χ2 Statistics
	N	%	N	%	χ2	p
Perceived parental perception of gambling					66.27	<0.0001
Disapprove	359	30.74	275	51.21
Approve	151	12.92	45	8.38
Neither	658	56.34	217	40.41
Perception of gambling prevention strategies Checking IDs for lottery					4.36	0.0368
Important	991	77.3	498	81.51
Not important	291	22.70	113	18.49
Hanging out with friends who do not gamble					4.51	0.0338
Important	865	68.22	444	73.03
Not important	403	31.78	164	26.97
Participating in leisure activities free of gambling					1.84	0.1752
Important	988	77.86	490	80.59
Not important	281	22.14	118	19.41
Fear of losing valuables, friends and/or relatives					0.74	0.3891
Important	1089	85.61	526	87.09
Not important	183	14.39	78	12.91
Advertisements of problems related to gambling					10.65	0.0011
Important	908	71.67	468	78.79
Not important	359	28.33	126	21.21
Not having access to Internet gambling at home					6.54	0.0105
Important	756	59.72	394	65.89
Not important	510	40.28	204	34.11
Parent/Guardian strictness about gambling					9.58	0.0020
Important	943	74.43	484	80.94
Not important	324	25.57	114	19.06
Warnings from adults in family					7.49	0.0062
Important	945	74.70	478	80.47
Not important	320	25.30	116	19.53
Warnings from, or listening to, peers					2.92	0.0872
Important	971	76.70	475	80.24
Not important	295	23.30	117	19.76
Having parents who do not gamble					13.91	0.0002
Important	923	72.73	478	80.74
Not important	346	27.27	114	18.26
Learning about the risks of gambling in school					14.45	0.0001
Important	886	69.98	465	78.41
Not important	380	30.02	128	21.59
Learning about the risks of gambling from parents					7.13	0.0076
Important	967	76.14	485	81.65
Not important	303	23.86	109	18.35
Learning about the risks of gambling from peers					6.46	0.0110
Important	923	72.73	464	78.25
Not important	346	27.27	129	21.75
Adults not involving kids in gambling					8.10	0.0044
Important	970	76.68	490	82.49
Not important	295	23.32	104	17.51
Parent/Guardian prohibition of card games for money at home					29.17	<0.0001
Important	727	57.24	417	70.32
Not important	543	42.76	176	29.68
Family concern					0.72	0.396
Yes	162	12.94	85	14.38
No	1090	87.06	506	85.62

EG: Excitement gambling; NEG: Non-excitement gambling

Health/functioning measures

Health/functioning measures data are shown in Table 3 and Supplementary Table 1 (that also includes data on numbers of respondents and percentages). Among the EG group, ARPG was associated with occasional (OR = 1.64, 95%CI 1.24 – 2.17, p = 0.0006) and regular (OR = 2.19, 95%CI 1.57 – 3.06, p < 0.0001) smoking; lifetime marijuana use (OR = 1.75, 95%CI 1.36 – 2.25, p < 0.0001); heavy alcohol use (OR = 1.66, 95%CI 1.03 – 2.65, p = 0.0364) and lifetime other drug use (OR = 2.42, 95%CI 1.64 – 3.57, p < 0.0001). Among the NEG group, ARPG was associated with occasional (OR = 1.78, 95%CI 1.03 – 3.06, p = 0.0024) and regular (OR = 2.64, 95%CI 1.41 – 4.94, p = 0.0384) smoking; light (OR = 2.72, 95%CI 1.09 – 6.78, p = 0.0324), moderate (OR = 5.27, 95%CI 2.22 – 12.49, p = 0.0002) and heavy (OR = 4.95, 95%CI 1.7 – 14.39, p = 0.0034) alcohol use; and lifetime other drug use (OR = 2.89, 95%CI 1.49 – 5.60, p = 0.0017). Interaction ORs revealed a weaker association between ARPG and moderate (OR = 0.25, 95%CI 0.10 – 0.62, p = 0.0025) and heavy (OR = 0.34, 95%CI 0.12 – 0.99, p = 0.0484) alcohol use in the EG vs NEG groups.

Table 3.

Health/functioning measures and problem-gambling severity stratified by excitement-seeking status: regression analysis

	EG	NEG	Interaction OR (EG versus NEG)
	ARPG versus LRG OR (95% CI)	ARPG versus LRG OR (95% CI)	ARPG versus LRG OR (95% CI)
Substance use Smoking
Never	Reference	Reference	Reference
Occasionally	1.64 (1.24 – 2.17)	1.78 (1.03 – 3.06)	1.00 (0.56 – 1.79)
Regularly	2.19 (1.57 – 3.06)	2.64 (1.41 – 4.94)	0.94 (0.48 – 1.81)
Marijuana, lifetime	1.75 (1.36 – 2.25)	1.54 (0.947 – 2.52)	1.10 (0.66 – 1.85)
Alcohol, current
Never regular	Reference	Reference	Reference
Light	1.04 (0.70 – 1.54)	2.72 (1.09 – 6.78)	0.41 (0.16 – 1.04)
Moderate	1.09 (0.74 – 1.60)	5.27 (2.22 – 12.49)	0.25 (0.10 – 0.62)
Heavy	1.66 (1.03 – 2.65)	4.95 (1.7 – 14.39)	0.34 (0.12 – 0.99)
Other drug, lifetime	2.42 (1.64 – 3.57)	2.89 (1.49 – 5.60)	0.80 (0.39 – 1.64)
Caffeine use
None	Reference	Reference	Reference
1–2 per day	0.75 (0.54 – 1.06)	0.69 (0.39 – 1.22)	1.25 (0.66 – 2.34)
3 + per day	1.11 (0.78 – 1.59)	1.09 (0.58 – 2.05)	1.18 (0.6 – 2.35)

EG: Excitement gambling; NEG: Non-excitement gambling; ARPG: At-risk/problem gambling; LRG: Low-risk gambling; OR: odds ratio; 95% CI: 95% confidence interval

Gambling behaviors

Gambling-behavior data are shown in Table 4 and Supplementary Table 2 (that also includes data on numbers of respondents and percentages). Among adolescents with EG, ARPG status was associated with gambling alone (OR = 2.91, 95%CI 1.95 – 4.36, p < 0.0001), gambling with strangers (OR = 3.91, 95%CI 2.57 – 5.94, p < 0.0001), gambling with other adults (OR = 1.98, 95%CI 1.54 – 2.55, p < 0.0001) and gambling for more than 2 hours (OR = 3.69, 95%CI = 2.68 – 5.08, p < 0.0001); ARPG status was also inversely associated with onset of gambling at 15 years old or older (OR = 0.61, 95%CI = 0.41 – 0.90, p = 0.0142). Among adolescents with NEG, ARPG status was associated with gambling alone (OR = 3.54, 95%CI 1.49 – 8.41, p = 0.0041), gambling with strangers (OR = 6.92, 95%CI 2.44 – 19.60, p = 0.0003) and gambling for more than 2 hours (OR = 95.83, 95%CI 11.88 – 773.27, p < 0.0001); ARPG status was inversely associated with onset of gambling at 12–14 years old (OR = 0.43, 95%CI 0.19 – 0.98, p = 0.0453) and at 15 years old or older (OR = 0.36, 95%CI 0.16 – 0.84, p = 0.0179). Interaction ORs revealed weaker associations between ARPG status and gambling with friends (OR = 0.54, 95%CI 0.31 – 0.93, p = 0.0276) and gambling for more than 2 hours (OR = 0.04, 95%CI 0.01 – 0.32, p = 0.0024) in the EG vs. NEG groups.

Table 4.

Gambling behaviors and problem-gambling severity stratified by excitement-seeking status: regression analysis

	EG	NEG	Interaction OR (EG versus NEG)
	ARPG versus LRG OR (95% CI)	ARPG versus LRG OR (95% CI)	ARPG versus LRG OR (95% CI)
People gamble with
Family	1.0 (0.79 – 1.27)	0.65 (0.33 – 1.26)	1.08 (0.56 – 2.1)
Friends	0.93 (0.65 – 1.33)	1.41 (0.90 – 2.23)	0.54 (0.31 – 0.93)
Other adults	1.98 (1.54 – 2.55)	1.77 (0.91 – 3.42)	0.89 (0.46 – 1.74)
Strangers	3.91 (2.57 – 5.94)	6.92 (2.44 – 19.60)	0.73 (0.26 – 2.03)
Alone	2.91 (1.95 – 4.36)	3.54 (1.49 – 8.41)	0.82 (0.34 – 1.99)
Time spent gambling
1 hour or less	Reference	Reference	Reference
2 + hours/week	3.69 (2.68 – 5.08)	95.83 (11.88 – 773.27)	0.04 (0.01 – 0.32)
Age of onset gambling
8 years old or less	Reference	Reference	Reference
9 – 11 years old	1.01 (0.67 – 1.55)	0.63 (0.23 – 1.76)	1.40 (0.49 – 4.02)
12 – 14 years old	0.81 (0.56 – 1.17)	0.43 (0.19 – 0.98)	1.66 (0.71 – 3.89)
15 years old or more	0.61 (0.41 – 0.90)	0.36 (0.16 – 0.84)	1.65 (0.67 – 4.08)

EG: Excitement gambling; NEG: Non-excitement gambling; ARPG: At-risk/problem gambling; LRG: Low-risk gambling; OR: odds ratio; 95% CI: 95% confidence interval

DISCUSSION

The present study provides sociodemographic characteristics, gambling-related perceptions, health/functioning measures and gambling behaviors in a large adolescent sample stratified by self-reported motivation to gamble for excitement. Our hypotheses were partially confirmed. Consistent with our a priori hypotheses, adolescents who reported gambling for excitement were more likely to report ARPG and report more permissive views towards gambling. EG status moderated relationships between ARPG and: 1) moderate and heavy alcohol use; 2) gambling with friends; and, 3) time spent gambling. Similar associations with other hypothesized measures (e.g., age of gambling onset) were similar across the EG and NEG groups. Implications are discussed below.

EG adolescents were more likely to exhibit ARPG. Motivational characteristics are important components of theoretical models of addictive behaviors, including problem gambling (Chambers, Taylor & Potenza 2003; Chambers & Potenza 2003; Potenza 2007). In line with our findings, previous studies have indicated excitement-seeking is associated with problem gambling among adolescents (Gupta, Derevensky & Ellenbogen 2006; Estevez, Herrero-Fernandez, Sarabia & Jauregui 2015; Dowling, Merkouris, Greenwood, Oldenhof, Toumbourou & Youssef 2017). Our study extends results from previous research as it showed that a simple question querying about gambling for excitement may help in identifying adolescents who might be at an increased risk of problem gambling. Considering parents and teacher are unlikely to identify youth who experience gambling-related problems and to administer psychological assessments that measure excitement-seeking and related constructs, questioning youth about EG might be a reasonable alternative to help identify ARPG youth. However, the high frequency of EG in adolescents may limit its specificity. Future research is required to determine potential treatment and prevention implications of using this question to identify ARPG youth, especially considering the dichotomous question identified the majority of adolescents who gamble (~67%) in our sample.

EG versus NEG adolescents were more likely to be white and male. Previous research on adult gambling has shown that men typically score higher than women on measures of sensation-seeking (Cross, Cyrenne & Brown 2013; Zuckerman & Neeb 1980). Yet, previous research investigating gender-related differences in adolescent gambling using a nationally representative U.S. sample reported that boys and girls did not differ on the self-reported motivation to gamble for excitement (Desai, Maciejewski, Pantalon & Potenza 2005). Given that our study employed a larger sample (n = 2030 vs n = 236), it is possible that the previous study was underpowered to demonstrate such gender-related differences. It is also possible that the previous study failed to demonstrate such differences given it only included adolescents of 16 and 17 years of age and, during this period of girls’ lives, excitement-seeking usually peaks (Romer & Hennessy 2007). The association between white race and excitement-seeking may be related to sociocultural aspects and may be conserved across different addictive behaviors. A previous meta-analysis reported that white race is associated with stronger effect sizes between excitement-seeking and alcohol use (Hittner & Swickert 2006). When taken together, the current findings suggest psychotherapeutic interventions aimed at controlling sensation-seeking and stimulating self-efficacy could target particularly white males as a strategy to prevent and treat adolescent problem gambling, although this is a speculative claim that requires empirical examination by future research.

EG versus NEG adolescents were more likely to classify gambling prevention strategies as not being important, which suggests EG youth were more likely to display a more permissive view towards gambling in comparison to NEG youth. Specifically, EG adolescents classified as unimportant several gambling prevention strategies that involve parental control; the group also reported more frequently having parents who did not have a negative opinion towards gambling. Our findings are in line with previous research reporting an association between sensation-seeking and permissive parenting in children and adolescents (Hayes, Hudson & Matthews 2004; Xu, Farver & Zhang 2009). Previous research has indicated that perceived parental permissiveness toward gambling was more strongly related to gambling and risky behaviors (i.e. substance use and related problems) in sensation-seeking youth (Leeman et al. 2014). We hypothesize that perceived parental permissiveness could account for some of the increased risk of ARPG in EG youth and that increased parental monitoring of EG youth could contribute to decrease gambling-related problems in this group (Magoon & Ingersoll 2006), although empirical research is required to examine directly this point.

EG status moderated the relationship between ARPG and moderate and heavy alcohol use such that a weaker association was observed relative to the NEG group. Excitement-seeking has previously been linked to both problem gambling (Gupta, Derevensky & Ellenbogen 2006; Estevez, Herrero-Fernandez, Sarabia & Jauregui 2015; Dowling, Merkouris, Greenwood, Oldenhof, Toumbourou & Youssef 2017) and alcohol use (Hittner & Swickert 2006) among adolescents, and the behaviors usually co-occur (Leeman et al. 2014; Rahman et al. 2014). These findings suggest that excitement-seeking may account for some of the variance in the co-occurrence of ARPG and moderate and heavy alcohol use, thereby weakening the relationship between them. ARPG and moderate and heavy alcohol use in the EG group may be in part explained by factors such as impulsivity (Liu et al. 2013; Leeman et al. 2014), depression and/or anxiety (Keough, Penniston, Vilhena-Churchill, Michael Bagby. & Quilty 2018; Medeiros, Sampaio, Leppink, Chamberlain & Grant 2016). Future studies should examine further relationships between EG, ARPG and moderate and heavy alcohol use.

EG status moderated the relationship between ARPG and gambling with friends and spending more than two hours per week gambling such that a weaker association was observed relative to the NEG group. Social aspects of gambling likely differ relative to ARPG status (Yip et al. 2011). Adolescents with EG may prefer gambling alone leading to problematic levels of gambling. Additional research into more isolative patterns of gambling are needed, particularly in the current gambling environment that includes gambling on the internet (Potenza, et al. 2011). Longer durations of gambling may be particularly relevant to ARPG among NEG adolescents. This possibility warrants additional investigation. However, given the wide 95% confidence intervals associated with NEG respondents gambling for more than 2 hours per week, this interpretation should currently be considered particularly cautiously.

Strengths and limitations

The large sample size analyzed, which is similar in composition to Connecticut census data, and the evaluation of a wide range of gambling attitudes and perceptions and co-occurring behaviors (e.g. substance use) are strengths of this study. However, important limitations should be noted. In our study, we were unable to characterize further gambling for excitement as answers were provided in a dichotomous yes/no manner. It is likely that there exists heterogeneity within youth that self-identified as excitement-seeking gamblers, and different findings could have been obtained if a more detailed description of EG had been available. As data were collected more than a decade ago and there have been changes in gambling and other behaviors (e.g., relating to internet and sports gambling), additional studies are warranted to examine relationships between gambling and excitement-seeking in the current environment. However, the current study provides important information against which future findings may be compared. In contrast to the Pantalon et al. (2008) study in which gambling for excitement or challenge was examined, we focused on gambling for excitement here. While we believe that this measure is more precise, the differences may complicate direct comparisons across studies. There was a large number of individuals with past-year gambling who were excluded from our analyses (N = 1,871) as they did not provide complete data on the DSM criteria for pathological gambling; as such, the findings may be biased and not be representative of the entire sample. The relatively small number of individuals classified as having ARPG in the EG and NEG groups may have contributed to some of our findings being underpowered, particularly in identifying interaction effects. Further, studies of clinical samples rather than high-school students may have identified different relationships. As the current study was cross-sectional, it had limited examination of the natures of associations, and could not examine longitudinal or potentially causal relationships. Future longitudinal studies are needed to examine relationships between excitement-seeking and gambling behaviors and to translate the findings into improved policy, prevention and treatment strategies.

CONCLUSION

The current study makes several important contributions to our understanding of the relationship between EG motivations and adolescent gambling and problem gambling. This study reported a significant association between EG and ARPG in a large adolescent sample and more permissive attitudes towards gambling in association with EG. EG status moderated associations between ARPG and multiple measures including moderate and heavy alcohol use, gambling with friends and age-of-gambling onset. Taken together, these findings highlight the importance of identifying EG among adolescents. Future studies should investigate the utility of screening for EG among youth.