Serious physical fighting and gambling-related attitudes and behaviors in adolescents

Background and aims: Physical fighting and gambling are common risk behaviors among adolescents. Prior studies have found associations among these behaviors in adolescents but have not examined systematically the health and gambling correlates of problem-gambling severity amongst youth stratified by fight involvement. Methods: Survey data were used from 2,276 Connecticut high school adolescents regarding their physical fight involvement, gambling behaviors and perceptions, and health and functioning. Gambling perceptions and correlates of problem-gambling severity were examined in fighting and non-fighting adolescents. Results: Gambling perceptions were more permissive and at-risk/problem gambling was more frequent amongst adolescents reporting serious fights versus those denying serious fights. A stronger relationship between problem-gambling severity and regular smoking was observed for adolescents involved in fights. Discussion and conclusions: The more permissive gambling attitudes and heavier gambling associated with serious fights in high school students suggest that youth who engage in physical fights warrant enhanced prevention efforts related to gambling. The stronger relationship between tobacco smoking and problem-gambling severity amongst youth engaging in serious fights suggest that fighting youth who smoke might warrant particular screening for gambling problems and subsequent interventions.


INTRODUCTION
Both physical fighting and gambling are risk behaviors that occur often among adolescents and each represents a significant health concern. A recent study found that 33% of high school students reported involvement in physical fights in 2011, with 12% of such fights occurring on school property (Eaton et al., 2012). Fighting is associated with other high-risk behaviors including early sex, substance use, and lower academic achievement (Dukarm, Byrd, Auinger & Weitzman, 1996;Fraga, Ramos, Dias & Barros, 2011;Howard, Wang & Yan, 2007Pickett et al., 2005). Estimates of past-year gambling among adolescents are even higher, ranging from 50-90% (Gupta & Derevensky, 2000;Shaffer & Hall, 2001). Gambling, particularly at-risk or problem gambling (ARPG) which reflects greater problem-gambling severity, has been associated with poor academic functioning, violence, depression and substance abuse, and problems later in life Rahman et al., 2012). While parents and adolescents appear aware of problems associated with fighting (Hamburg, 1998;St. George & Thomas, 1997), data suggest that both groups may be less concerned about the risks associated with adolescent gambling (Campbell, Derevensky, Meerkamper & Cutajar, 2012).
Gambling and aggression, particularly extreme patterns of each (e.g., ARPG and propensities to get into physical fights) might each be conceptualized as expressions of impaired behavioral impulse control and thus be hypothesized to be associated. Gambling and fighting have been shown to co-occur amongst adults (Afifi, Brownridge, MacMillan & Sareen, 2010;Brasfield et al., 2012;Korman et al., 2008) and adolescents (Chaumeton, Ramowski & Nystrom, 2011;Goldstein, Walton, Cunningham, Resko & Duan, 2009;Potenza et al., 2011;Proimos, DuRant, Pierce & Goodman, 1998), suggesting that adolescents who fight might view gambling more permissively, and vice versa. However, little is known regarding their interaction and relative impacts on adolescent health. Despite associations between gambling and fighting, prior studies have not systematically examined health and gambling correlates of problem-gambling severity amongst adolescents based on their involvement in serious physical fights.
To address an existing gap in knowledge, we examined high school survey data to investigate the relationship between problem-gambling severity and health and gambling measures in adolescents who acknowledged or denied past-year involvement in serious fights resulting in physical injury. This subsample of adolescents, which may include victims of bullying, perpetrators of bullying, or individuals involved in both bullying and victimization, is a population associated with a variety of aggressive and risky behaviors. A number of studies have shown that victims of bullying, as well as adolescents defined as both bullies and victims, have reported a greater level of substance use than adolescents uninvolved in bullying behaviors (Radliff, Wheaton, Robinson & Morris, 2012;Tharp-Taylor, Haviland & D'Amico, 2009). Links between risky behaviors and victimization have also been observed in adults, with over 50% of one sample of problem gamblers reporting past subjection to physical and verbal intimate partner abuse (Korman et al., 2008). Often it is challenging to ascertain through self-report the extent to which adolescents involved in physical aggression may be perpetrators, victims or both. For these reasons, the group as being involved in physical fights was considered as a single entity.
In this current study, we hypothesized that problemgambling severity would be associated with fight involvement; adolescents involved in fights would view gambling more permissively and problem-gambling prevention efforts as less important; and health and functioning measures (poor academic performance, carrying a weapon, and substance use) and gambling measures (types and locations of gambling) would show differential relationships with problem-gambling severity in the fighting versus non-fighting groups (e.g., given propensities to fight on school grounds, different relationships with gambling on school grounds would be observed).

Survey
Cross-sectional, anonymous survey data from high school students were collected as described previously Grant, Potenza, Krishnan-Sarin, Cavallo & Desai, 2011a, 2011bKundu et al., in press;Liu, Desai, Krishnan-Sarin, Cavallo & Potenza, 2011;Potenza et al., 2011;Rahman et al., 2012;Schepis et al., 2008Schepis et al., , 2010Yip et al., 2011). Every public 4-year and non-vocational and special-education high school in Connecticut was invited to participate. The initial response from schools was not sufficient to ensure representation of all geographic regions in Connecticut so schools in targeted areas were re-contacted. The final sample included schools from each geographic quadrant and all three district-reference groups (a proximal link to socioeconomic status) and was consistent with the 2000 Census data of 14-to 18-year-old Connecticut residents. For the current study, 2,276 adolescents who completed the serious-fighting measure and all 12 questions corresponding to the inclusionary criteria for pathological gambling were included. A passive-consent procedure was utilized to obtain parental permission. Letters were mailed to parents outlining the study and instructing those not wanting their child participating to contact their child's high school, usually by calling the school's main office. From these phone calls, a list of students who were ineligible to participate was compiled for use on survey administration day. This consent procedure was approved by participating schools and Yale's Institutional Review Board.
Survey administration occurred on a single day at each school. Participation was voluntary, taking around 50 minutes. Reminders were given to keep information anonymous. Less than 1% of students refused to participate.

Measures
Problem-gambling severity and fight measures. Problem-gambling severity was defined (non-gambling, low-risk gambling [LRG], at-risk/problem gambling [ARPG]) using the 12 items from the Massachusetts Gambling Screen (MAGS) relating to the 10 inclusionary criteria for DSM-IV pathological gambling Shaffer, LaBrie, Scanlan & Cummings, 1994;Yip et al., 2011). The MAGS is a validated instrument designed to assess DSM-IV pathological gambling in adolescents (Shaffer et al., 1994). Specifically, participants endorsing gambling and no inclusionary criteria were classified as having LRG and those endorsing one or more criteria were classified as having ARPG.
Respondents were categorized into fight and non-fight groups based on a question from the Youth Behavior Risk Survey that stated, "During the past 12 months, how many times were you in a physical fight in which you were injured and had to be treated by a doctor or nurse?", with responses grouped as one or more versus none (Eaton et al., 2012).
Gambling perceptions. As described previously (Kundu et al., in press), adolescents indicated the importance (very/somewhat = "important" versus "not important") of the following gambling prevention approaches: checking identification for purchasing lottery tickets; hanging out with friends who do not gamble; participating in activities that are fun and free of gambling; fear of losing valuable possessions, friends or relatives to gambling; advertisements that show problems associated with gambling; not having access to Internet gambling at home; parent/guardian strictness about gambling; warnings about gambling from adults in the family; warnings about gambling from, or listening to, peers; having parents who do not gamble; learning about the risks of gambling in school; learning about the risks of gambling from parents; learning about the risks of gambling from peers; adults not involving kids in gambling; and parents/guardians not permitting card games (for money) at home.
Correlates of problem-gambling severity. Health and functioning measures were categorized as shown in the tables and included the following variables: grade average; extracurricular activities; lifetime tobacco smoking; lifetime marijuana use; ever and past-30-day alcohol use (categorized as none, light (1-2 drinking-days/month), moderate (3-9 drinking-days/month), and heavy (³9 days drinking-days/month)); lifetime use of other drugs; caffeine use; past-year sadness or hopelessness for ³2 weeks; and past-year carrying of a weapon such as a knife, club, or gun to school.
Dichotomous gambling variables (yes/no) were calculated among gamblers and included gambling types (strategic, non-strategic, machine), gambling locations (online, school grounds, casino), gambling triggers (pressure, anxiety), reasons for gambling (excitement, financial, escape, or social), usual company when gambling (family, friends, other adults, strangers, or alone), and weekly time spent gambling (<1 h, ³2 h).

Data analysis
As in prior analyses Desai et al., 2010;Grant et al., 2011aGrant et al., , 2011bKundu et al., in press;Liu et al., 2011;Potenza et al., 2011;Rahman et al., 2012;Schepis et al., 2008Schepis et al., , 2010Yip et al., 2011), data were double-entered and checked for accuracy. Analyses were conducted using SAS software (Cary, NC). Two-tailed, Pearson chi-square analyses (c 2 ) were used to compare characteristics and gambling perceptions of adolescents stratified by fight-involvement status. A Bonferroni correction was applied such that p-values of p < 0.0025 were considered significant. To produce odds ratios (ORs) and 95% confidence intervals (CIs) as a measure of the magnitude of the association between problem-gambling severity and dependent variables, logistic regression models were constructed for binary outcomes and multinomial logistic regression models for categorical outcomes, stratified according to fight-involvement status. To determine whether fight-involvement status moderated relationships with problem-gambling severity, the entire sample was utilized and main effects for fight involvement and problem-gambling severity, as well as the interaction term (fight-status-by-problem-gamblingseverity), were included in the appropriate logistic or multinomial logistic regression models. All models were adjusted for gender, grade level, race, Hispanic ethnicity, and family structure (living with one parent, both parents, other). Statistical significance was set at p < 0.05.

Gambling perceptions
Fighting adolescents displayed more permissive views towards gambling on all queried items (all p < 0.0012; Table 2). Amongst fighting adolescents, 50.47% to 65.09% viewed specific gambling prevention and other non-permissive measures as important, compared to a range of 65.03% to 89.85% in non-fighting adolescents (Table 2). These measures included items that queried the adolescents on the importance of gambling prevention measures that involved parental oversight of gambling activities, parental non-in-  learning about the risks of gambling from peers (c 2 = 26.73; p < 0.0001). In addition, fighting adolescents indicated significantly greater parental approval of gambling (c 2 = 72.20, p < 0.0001), as well as greater concern about the gambling of a close family member (c 2 = 23.18; p < 0.0001).

Health/functioning measures
Health and functioning data are displayed in Table 3  ing was over seven times stronger in fighting adolescents than in non-fighting adolescents.

Gambling characteristics
Gambling characteristics are displayed in Table 4 and Supplemental Interaction analyses did not identify any significant effects, suggesting that the gambling-related correlates of problem-gambling severity were similar across fighting and non-fighting groups.

DISCUSSION
To our knowledge, this is the first study to investigate differences in gambling perceptions, attitudes and behaviors and problem-gambling-severity correlates in adolescents stratified by past-year involvement in serious physical fights that required medical attention. Consistent with our first hypothesis, fight-involvement status was associated with problem-gambling severity, with a greater proportion of ARPG amongst fighting versus non-fighting adolescents. Consistent with our second hypothesis, fighting versus non-fighting adolescents reported more permissive attitudes toward gambling. Our third and fourth hypotheses were largely not supported as the relationship between problem gambling severity and health and gambling characteristics appeared sim-  ilar across the two groups of adolescents, with the exception of regular smoking. Clinical implications are described below.

Problem-gambling severity
Associations between violent behaviors and problem-gambling severity have been reported previously (Afifi et al., 2010;Brasfield et al., 2012;Chaumeton, 2011;Goldstein et al., 2009;Korman et al., 2008;Potenza et al., 2011;Proimos et al., 1998). Our finding of an association between ARPG and serious physical fights suggests that youth engaging in fights may be at risk for gambling problems; alternatively youth engaging in gambling may potentially be at risk for fighting. As physical fighting among adolescents is commonly visible to adults who may witness the act or observe the injuries, it should help provide insight into less observable risk behaviors, such as gambling, in which the adolescent may be involved. Schools might consider targeting aggressive behaviors with educational interventions; e.g., teaching individuals detained for physical fighting about the potential risks of gambling. School policies could also consider educating parents of adolescents who fight about the relationship between fighting and gambling in adolescents.
Although reasons behind an association between violent behaviors and gambling are undetermined, several possibilities include motivational factors (e.g., gambling as a method of escape from distressing situations) or shared behavioral tendencies (diminished self-control; Boughton & Falenchuk, 2007;Crisp et al., 2004;Ledgerwood & Petry, 2006;Li, 2007;Walker, Hinch & Weighill, 2005). Poor impulse control should be further investigated in fighting adolescents, which could link to both perpetrators and victims of physical violence. The extent to which adolescents involved in fights gamble to escape should also be further studied, particularly amongst youth who are bullied, as our data indicate a greater percentage of fighting adolescents reporting gambling to escape (48%) than non-fighting adolescents (27%). This finding indicating more frequent motivations relating to gambling to escape amongst fighting adolescents reporting resulting injuries may suggest a greater likelihood of losing the fight or being the victims of bullying. Additional research is needed to investigate directly this possibility, and if this hypothesis is upheld, it may be particularly useful to assess gambling behaviors in adolescent victims of bullying.

Gambling attitudes and perceptions
Fighting versus non-fighting adolescents reported more permissive views toward gambling across a broad range of measures. More adolescents in the fight (versus non-fight) group indicated that their parents approved of gambling, possibly suggesting a permissive parenting style amongst these parents. Fighting versus non-fighting adolescents rated the multiple gambling-related efforts involving parental oversight as less important including parent/guardian strictness about gambling, warnings from adults in family, learning about the risks of gambling from parents, and not having access to Internet gambling at home. Permissive parenting, characterized by a lack of parental monitoring (Ginsburg, Durbin, Garcia-Espana, Kalicka & Winston, 2009), is linked to internalizing and externalizing behavioral problems in youth (Alizadeh, Talib, Abdullah & Mansor, 2011), and may represent an underlying factor in adolescents who engage in physical fighting and gambling. Low parental monitoring has been linked to both gambling and physical fighting in adolescents while high levels protect against these risk behaviors (Curtner-Smith & MacKinnon-Lewis, 1994;Magoon & Ingersoll, 2006;Rudatsikira, Mataya, Siziya & Muula, 2008).
Fighting adolescents also rated measures reflecting parental involvement in gambling, such as having parents who do not gamble and parents who do not involve their children in gambling activities, as significantly less important than adolescents who did not fight. Parents may directly influence their child's engagement in risk behaviors by openly participating in these activities themselves and/or involving their child in such behaviors. Children of parents who gamble and of parents who endorse violent behaviors have been shown to engage in similar behaviors (Farrell, Henry, Mays & Schoeny, 2011;Lesieur & Klein, 1987;Ohene, Ireland, McNeely & Borowsky, 2006;Winters, Bengston, Dorr & Stinchfield, 1998;Winters, Stinchfield & Fulkerson, 1993). These associations should be considered in the development of school disciplinary policies for adolescents who fight, as a greater emphasis on the parent-child relationship may be helpful in reducing other risk behaviors in these adolescents. For example, school administrators or teachers could educate parents of fighting adolescents on the protective role of parental monitoring and the potential risks of openly engaging in gambling or violent behaviors.
In addition, fighting versus non-fighting adolescents rated as less important those measures reflecting friend/peer involvement in gambling, such as hanging out with friends who do not gamble, warnings from or listening to peers about gambling, and learning about the risks of gambling from peers. Such responses suggest a greater propensity for fighting youth to consider gambling less risky or problematic, including with respect to peer advice about gambling-related risks. Data indicate strong links between delinquent peer associations and problem behaviors that include gambling and physical aggression (Brown & Wolfe, 1994;Farrell et al., 2011;Hardoon & Derevensky, 2001;Kearney & Drabman, 1992). Developing approaches to change gambling attitudes amongst fighting youth, including with respect to peer involvement, represents an important effort.
Fighting adolescents more frequently acknowledged concerns about a family member's gambling, suggesting that adolescents who witness parents with poor control over their gambling may be more likely to get into fights at school. The extent to which this relationship underlies the observed findings, as well as the extent to which other related factors (e.g., stress or trauma exposure, each of which has been linked to gambling and violence [Bergevin, Gupta, Derevensky & Kaufman, 2006;Kaplan, Madden, Mijanovich & Purcaro, 2012;Kausch, Rugle & Rowland, 2006;Schiff et al., 2012]) might mediate such a relationship, warrants additional investigation. Additionally, the extent to which individual differences relating to impulse control might mediate relationships between stress and gambling (as has been observed between stress and hazardous drinking in adults [Hamilton, Ansell, Reynolds, Potenza & Sinha, 2013]) warrants further investigation. Such information could inform the development and implementation of interventions (e.g., mindfulness-based stress reduction) to prevent youth violence and gambling problems (de Lisle, Dowling & Allen, 2011;Robins, Keng, Ekblad & Brantley, 2012). Behavioral Addictions 2(3), pp. 167-178 (2013) | 173

Fighting and gambling
Correlates of problem-gambling severity With the exception of regular smoking, the correlates of problem-gambling severity with measures of functioning and gambling characteristics were largely similar amongst fighting and non-fighting adolescents. The stronger association between problem-gambling severity and tobacco smoking amongst fighting adolescents appears consistent with reported associations between gambling, violent behaviors, and alcohol and substance use (Brasfield et al., 2012;Wanner, Vitaro, Charbonneau & Tremblay, 2009). Although the nature of these associations is undetermined, similarities in personal dispositions, familial qualities, and peer influences have been shown in adolescents who engage in gambling, compulsive substance and alcohol use, and delinquent behaviors including physical violence; these factors include impulsivity, poor parental supervision, and deviant peers (Wanner et al., 2009). Adolescents may also be especially likely to engage in high-risk behaviors such as cigarette smoking, physical fighting and gambling for social reasons such as peer pressure and improving popularity (Brady, Song & Halpern-Felsher, 2008;Johnson, Frattaroli, Wright, Pearson-Fields & Cheng, 2004;Langhinrichsen-Rohling, Rohde, Seeley & Rohling, 2004).

Strengths and limitations
This study has multiple strengths including a large sample size that is similar in composition to Connecticut census data, as previously described Grant et al., 2011aGrant et al., , 2011bKundu et al., in press;Liu et al., 2011;Potenza et al., 2011;Rahman et al., 2012;Schepis et al., 2008Schepis et al., , 2010Yip et al., 2011). Limitations also exist. First, the number of adolescents involved in serious physical fights was relatively small, limiting the power to detect interaction effects. Second, although the question determining fight-involvement status is derived from the widely used Youth Child Risk Behavior Survey (thus facilitating comparisons across studies), it does not differentiate between levels of violence, such as the amount or extent of physical fights, or whether these adolescents were perpetrators or victims of physical violence. Future research should investigate these areas with respect to problem-gambling severity. Such information might be best obtained with the input of involved school officials as adolescents involved in fights may be unwilling to admit to either perpetration or victimization, or may be inaccurate in their assessments thereof. Third, as the sample is from Connecticut, it is not nationally representative and findings may not generalize. Fourth, the study was cross-sectional, limiting the ability to fully examine the nature of the observed associations. For example, it cannot be determined whether fighting in adolescence leads to gambling or gambling leads to fighting behaviors; thus, longitudinal studies are needed. Fifth, multiple measures, including assessments of depressive and aggressive features, used non-diagnostic and dichotomous measurements. Future studies using more clinically valid measurements may be valuable to better understand relationships between problem-gambling severity and health/ functioning measures.

CONCLUSIONS
The current study demonstrates that adolescents involved in serious physical fights are more likely to report more per-missive gambling-related perceptions and attitudes, exhibit more risky/problematic gambling and demonstrate stronger associations between tobacco smoking and problem-gambling severity than adolescents who do not physically fight. Such findings highlight the need for more research into the etiologies of these relationships. Improved educational prevention and interventional efforts for adolescents who fight that also incorporate teachings on more discrete risk behaviors such as gambling, may be useful in targeting commonalities of both risk behaviors.
Funding sources: This work was supported in part by the NIH (RL1 AA017539), the Connecticut State Department of Mental Health and Addictions Services, the Connecticut Mental Health Center, The Connection, an unrestricted research gift from the Mohegan Sun casino, and the Yale Gambling Center of Research Excellence Award grant from the Institute for Research on Gambling Disorders. The funding agencies did not provide input or comment on the content of the manuscript, and the content of the manuscript reflects the contributions and thoughts of the authors and do not necessarily reflect the views of the funding agencies.
Authors' contributions: M Slavin generated the initial draft of the manuscript. CP conducted analyses. RH, SK-S, M Steinberg and MP developed the survey, with RH, SK-S and MP overseeing data collection and entry. All authors edited the manuscript and approved the submitted work.

Conflicts of interest and disclosure:
The authors report no conflicts of interest with respect to the content of this manuscript. Dr. Potenza has served as a consultant or advisor to Boehringer Ingelheim, Somaxon, Lundbeck, Ironwood, gambling businesses and organizations, law offices, the federal defender's office in issues regarding impulse control disorders. He has received research support from the National Institutes of Health, Veteran's Administration, Mohegan Sun Casino, the National Center for Responsible Gaming and its affiliated Institute for Research on Gambling Disorders, Psyadon, Forest Laboratories, Ortho-McNeil, Oy-Control/Biotie, and GlaxoSmithKline. He has participated in surveys, mailings, or telephone consultations related to drug addiction, impulse control disorders, or other topics. He has provided clinical care in the Connecticut Department of Mental Health and Addiction Services Problem Gambling Services Program. He has performed grant reviews for the National Institutes of Health and other agencies. He has guest-edited journal sections, has given academic lectures in grand rounds, continuing medical education events, and other clinical and scientific venues, and has generated book or book chapters for publishers of mental health texts. Dr. Hoff has received research support from the National Institutes of Health (NIH), Veterans Administration Clinical Research and Development, the National Center for Responsible Gambling and its affiliated Institute for Research on Gambling Disorders, and the National Center for PTSD; has participated in surveys, mailings, or telephone consultations related to psychiatric illness, ethics in medical research, or other health topics; has performed grant reviews for the NIH and other agencies; has guest edited journal sections; has given academic lectures in grand rounds, CME events, and other clinical or scientific venues; and has generated books or book chapters for publishers of mental health texts. Drs. Pilver, Steinberg, Rugle, and Krishnan-Sarin and Ms. Slavin report no biomedical financial interests or potential conflicts of interest.