Clinical and community-based studies indicate that gambling and gambling problems are common among many racial and ethnic groups (1). However, studies examining the etiology and treatment of pathological gambling have largely comprised white participants or have recruited insufficient members of minority groups for analyses of potential racial group differences among nonwhites (2), resulting in a knowledge gap. Studies in the United States have generally (3) but not uniformly (4) found increased risk of problem gambling and pathological gambling among blacks compared with whites. These findings suggest a health disparity that warrants addressing in order to optimize prevention and treatment strategies.
Identifying people with pathological gambling and engaging them in treatment during earlier stages of the illness are a challenge (5). Gambling helplines represent an important outreach strategy to guide primarily treatment-naïve individuals with problematic gambling into treatment. Gambling helpline findings complement those from general population and treatment samples and elucidate the characteristics of a group of problem gamblers who are likely to be in early stages of readiness for intervention (6). Despite the popularity of gambling helplines, few systematic studies have examined the race-related characteristics of problem gamblers using these services (7) and none have focused on blacks. An improved understanding of factors related to racial differences among callers to gambling helplines could enhance resource and program planning for gambling treatment programs.
This study investigated race-related differences between black and white callers to a gambling helpline. Given race-based differences in education and marital status in the general population (8), we hypothesized that black callers to the gambling helpline would be less likely than white callers to be married and to have education that extended past high school. Epidemiological findings indicate high proportions of black women with gambling problems (9), leading us to hypothesize that higher proportions of black callers than white callers would be women. Considering the high rates of gambling problems among blacks (9) in conjunction with hypothesized diminished treatment utilization (see below), we expected that black callers would demonstrate differences from white callers in patterns of gambling; for example, we hypothesized that black callers would have on average a longer duration of problem gambling and would report less frequently problems with forms of gambling (such as slot machines) that reportedly have rapid progression to problematic levels. Given findings suggesting that blacks are less likely than whites to have received mental health treatment (10), we hypothesized that black callers would be less likely than white callers to have utilized treatment services for gambling, substance abuse, and other mental problems. Because treatment seeking is influenced by problems in these domains, substance abuse and mental health problems were examined for race-related differences between blacks and whites. Although race-related differences in forms of gambling problems have been reported (for example, Asian-American gamblers are more likely than white gamblers to play baccarat) (11), studies have not specifically explored forms of gambling problems among blacks. Consequently, we examined possible race-related differences in forms of problem gambling among black and white callers to a gambling helpline.
Data were from deidentified recorded telephone calls to the Connecticut Council on Problem Gambling (CCPG) helpline, which were presented to the Yale Human Investigations Committee and exempted from review. The data were obtained from 2,741 calls received in the period of January 1, 2000, through December 31, 2003. As previously described (5,6), analyses were restricted to calls from "problem gamblers" (1,941 of the 2,741 calls), a term to describe callers who were seeking help for gambling problems. Of these, 1,797 callers provided information on race and age and were 18 years or older. Of these, 170 calls were excluded because the caller did not self-identify as either black or white, leaving 1,627 calls for analysis.
Helpline questionnaire items were grouped as in previous studies (5,6) into the following categories: demographic characteristics, gambling types and durations, forms of problematic gambling, psychiatric problems secondary to gambling, social problems secondary to gambling, financial problems, types of debt, substance use problems, and treatments received. Variables for years of gambling and anxiety secondary to gambling were removed because of colinearity with the variables years of problem gambling and depression secondary to gambling, respectively.
Logistic regression analyses were completed as described previously (5,6) for each of the nine categories of variables to determine relationships to the dependent variable of race (black or white). Nine regression models, one for each category, were generated. A conservative approach, with Bonferroni correction, was used to examine the nine models, with significance set at p<.006. Because black and white callers differed significantly on gender and education, odds ratios for nondemographic categories were adjusted for these two characteristics. If the overall model for a particular category was significant, individual variables within the model were examined with the results of the logistic regression analysis for significant relationship to black race. Before completion of the logistic regression analyses, independent variables in each category were examined for colinearity and multicolinearity by using correlation matrices and the equivalent model that was adjusted by a weight matrix. The SAS System was used for data coding, estimating models, and data analysis.
The sample included 151 black (9%) and 1,476 white (91%) adult problem gamblers. During logistic regression analysis, each of the nine categories of variables distinguished the groups of black and white callers (Table 1). For the categories of social problems secondary to gambling, financial problems, and types of debt, statistical significance appeared to be attributable to the inclusion of gender and education, which contributed significantly to each model. Each of the other categories (except possibly type of debt, which was at the p<.006 significance threshold) remained statistically significant (p<.05) after the analyses controlled for multiple comparisons. The variables of gender and education were not tabulated outside of the demographic characteristics category.
Although comparable proportions of black and white callers were married, black gamblers were less likely than white gamblers to have an education that extended past high school and were more likely to be female (Table 1). Compared with whites, blacks reported, on average, longer durations of problem gambling and were less likely to endorse a problem with casino slot machine gambling.
Whereas similar proportions of black callers and white callers reported gambling-related suicidal ideation or suicidal attempts, blacks were less likely to endorse depression secondary to gambling. Whites were more likely to report daily tobacco use. Although lower proportions of blacks than whites reported problems with alcohol or drug use, between-group differences were not significant (Table 1).
Compared with white callers, black callers were less likely to report having received mental health treatment. Fewer blacks than whites reported prior professional or 12-step gambling treatment, and this difference approached but did not reach statistical significance at p<.05. Comparable percentages of blacks and whites reported prior professional or 12-step substance abuse treatment (Table 1).
Multiple similarities were observed across racial groups. For example, similarly substantial proportions of black callers and white callers reported familial and financial problems related to gambling, with similarly high reports of credit-related financial problems and overall debt. These findings suggest that interventions that target these domains (family therapy, financial counseling, and interventions related to access to credit) are important for both black and white problem gamblers.
Our hypothesis that black problem gamblers would be less likely than white problem gamblers to be married and to have education past high school was partly supported. A smaller percentage of black gamblers compared with white gamblers were married (39% versus 45%, respectively), but this difference was not significant, possibly reflecting limited statistical power due to the study's sample size. Black callers were less likely than white callers to have education past high school. The extent to which education might influence involvement in gambling and the development of gambling problems across racial groups warrants consideration. College attendance has increased among minority groups. Although this may expose individual members of minority groups to an environment where gambling is common and treatment services are scarce (12), it may also better prepare them to manage their finances and reduce the long-term risk of engaging in risky financial ventures.
Higher proportions of black problem gamblers than white problem gamblers were women, which supports research indicating that women from racial minority groups may be at high risk of problem gambling (9). Also, black women may be more likely than their white counterparts to seek help. Attending to gender differences nested within cultural and racial groups is important in reducing the risk of stereotyping and enhancing treatment efficacies (13).
The findings support our hypothesis that black callers would show patterns of gambling different from their white counterparts, including a longer duration of problem gambling. Helpline callers reported having protracted gambling problems (average of more than seven years), with black callers endorsing a significantly longer duration than whites. This difference remained significant after we controlled for gender differences, suggesting that the finding is not attributable to the telescoping phenomenon described for female problem gamblers (14). Longer duration of problem gambling among blacks could reflect differences in treatment seeking or service utilization.
The finding that whites were more likely than blacks to report having gambling problems with casino slot machines is new. Although the precise basis of this difference is unclear, slot machine gambling has been described as "escape oriented" (5). Future studies on race and problem gambling may benefit from examining escape-oriented constructs such as sensation seeking and dissociation.
Our hypothesis that blacks would be less likely than whites to report having utilized mental health, gambling, and substance abuse treatments was partially supported. Blacks were less likely to report having received mental health treatment; this finding could be related to their reporting fewer mental health problems, particularly depression secondary to gambling. Future studies investigating the relationships between depression and other mental disorders, racial group status, gambling behaviors, and treatment seeking are warranted.
Although a higher proportion of whites reported receiving prior gambling treatment (attending professional treatment or Gamblers Anonymous), this finding approached significance at p=.05 but was not statistically significant. An improved understanding of treatment-seeking behaviors among minority groups at various stages of problem and pathological gambling is needed. Although similar proportions of both groups reported prior substance abuse treatment, a significantly higher proportion of whites than blacks reported daily tobacco use (57% versus 41%, respectively). This finding suggests that tobacco use may be an important target for resource and program planning in gambling treatment programs, especially among whites.
Limitations of this study include the use of self-report measures, the absence of formal diagnostic assessments of pathological gambling and other mental disorders, regional differences in the availability of forms of gambling, the telephone-based nature of data collection, and the potential for bias secondary to callers' subjective interpretations of questions or incomplete provision of data (5). A comprehensive assessment of psychiatric disorders would not only better define the sample with regard to psychological problems, but it would also elucidate further examination of race-related mental health utilization rates among black gamblers and white gamblers. This study investigated black and white callers only; although this strategy was used to limit variability, future research should examine gambling behaviors of other racial and ethnic groups. The extent to which gambling helplines introduce a sampling bias (that is, racial groups may differ in individuals' likelihood to use them) is unclear.
Despite these limitations, this study represents an important investigation of black-white racial differences in the characteristics of problem gamblers. It is the first, to our knowledge, to investigate specifically race-related differences among black and white problem gamblers who used a gambling helpline. Such helplines offer the possibility of directing large numbers of individuals with gambling problems to appropriate treatments and thus have substantial clinical significance (5). Because most callers had never utilized professional or self-help gambling treatments, they represent an important help-seeking clinical sample. The findings of differences in the characteristics of black and white callers have implications for program planning and outreach efforts.
This work was supported in part by grants K12-DA-00366, K12-DA-14038, K05-DA-00089, P50-DA-009241, UL1-RR-024925, RL1-AA-017539, R01-DA-020908, and R01-DA-019039 from the National Institutes of Health; by the Department of Veterans Affairs Integrated Service Network 1, Mental Illness Research, Education and Clinical Center; and by the Department of Women's Health Research at Yale.
Dr. Potenza consults for and is an adviser to Boehringer Ingelheim, has consulted for and has financial interests in Somaxon, and has received research support from the Mohegan Sun and from Forest Laboratories. The other authors report no competing interests.