Introduction

Problem/pathological gambling rates have increased in North America with the widespread increase in the number of gambling venues and government lotteries [1]. One of the groups most vulnerable to developing gambling-related problems are APIs [2, 3]. API communities are at risk for problem/pathological gambling for a number of reasons. First, language and cultural barriers do not prevent participation in gambling activities; in fact, major U.S. gambling establishments have recently begun to cater to the API market [4–7]. Second, financial difficulties may increase the perceived value of gambling as a way to get rich, particularly among immigrant APIs who may have fewer resources relative to long established U.S. residents. Third, social isolation from immigration makes gambling inviting, especially when peers are gambling. Fourth, gambling is culturally accepted and approved within many API cultures, with less stringent religious prohibitions regarding gambling and evidence of historical gambling traditions as documented by widespread pre-colonial era gambling in China, India, Korea, Vietnam, Thailand, Southeast Asia, and Japan [8].

An informal study conducted by the NICOS Chinese Health Coalition in San Francisco and a group of UC Berkeley graduate students found that as many as 21% of the Chinese community could be identified as pathological gamblers [9]. Furthermore, 16% of those surveyed considered themselves to be problem gamblers, or were told by somebody else that they were problem gamblers. In a more formal survey of 96 Southeast Asian refugees who emigrated to the United States from Laos, Cambodia, or Vietnam and were living in Connecticut [10], 59% scored 5 or more on the South Oaks Gambling Screen (SOGS) [11]. The respondents in this study completed SOGS scales that had been translated into their native languages and the three groups had mean scores of 5.6 (Laotian), 5.6 (Cambodian), and 7.9 (Vietnamese).

Problem/pathological gambling is associated with co-occurring substance use disorders, mood disorders, personality disorders [12], suicidality [13], and intimate partner violence [14, 15]. Problem/pathological gambling and its associated problems can take a major toll on Asian communities by impacting health, social, financial, and family well-being. Anecdotally, San Francisco API social workers estimate that: (a) 25–33% of divorce and domestic violence cases have their roots in pathological gambling; (b) close to 20% of the cases of child neglect in Santa Clara, CA have been linked to pathological gambling; and (c) community reports of bankruptcies in Westminster and Monterey Park are ‘legendary’ according to banking personnel and community leaders.

In addition to issues associated with problem/pathological gambling presented above, the practical impact of casinos on communities is related to attitudes towards gambling. For example, a factor related to the disapproval of the establishment of casinos in a community is the perception of negative consequences stemming from casino development, including traffic congestion [16].

Gambling is highly accessible in Southern California. Cardrooms, horse tracks, and Indian casinos are located within or very close to Los Angeles County. Las Vegas, Palm Springs, and Northern San Diego, all are destination resorts for gambling and are within a few hours drive. APIs have a high participation rate in cardrooms and casinos in Southern California and often are the target of casino marketing. One gambling club near Los Angeles County boasts a clientele that is approximately 50% API [4]. Also, there are numerous trips to Las Vegas and to the Indian casinos that originate from Los Angeles County areas with large populations of API residents.

Surveys throughout the state show that APIs desperately need resources to address problem/pathological gambling. According to the Chinese Community Health Study, a survey of 1,808 Chinese American adults in San Francisco [9], approximately 70% of respondents identified gambling as a problem in their community, making it the top social concern of the population. Similarly, a San Jose Mercury News survey of a Vietnamese community [17] showed that 34% of respondents regard gambling as a problem.

Very little research on gambling has been conducted among Asian-Pacific Islander (APIs) populations although they may be at higher risk for problem/pathological gambling. To date much of the information about problem/pathological gambling among APIs is in the form of newspaper reports or reports conducted by community stakeholders. This study examines knowledge about, community impact of, and consequences of problem/pathological gambling among male and female APIs and non-APIs attending API community events in Los Angeles County. We define APIs as individuals who self-identify as being of Chinese, Japanese, Korean, Vietnamese, Filipino, Thai, Southeast Asian, Indian Subcontinent, or Pacific Islander descent or origin. Thus, our definition of APIs includes both immigrant and U.S. born individuals. Non-APIs are defined as those who self-identify as any ethnicity other than those designated above as API.

The current data and analyses add to an underdeveloped area of the literature in that they provide preliminary community-based data collected by an independent research group on ethnic differences in knowledge about, community impact of, and consequences of problem/pathological gambling. Given the previous reports of potentially higher rates of problem/pathological gambling among APIs and the noted community concern about gambling problems, the goal of these analyses is to explore potential ethnic differences in knowledge, community impact of, and consequences of problem/pathological gambling. This study represents a starting point for research into cultural/gender differences that may be important in the design of community-wide problem/pathological gambling prevention interventions and treatment programs for APIs. Such research would be useful to county policy makers and advocates for prevention and treatment of problem gambling in multicultural communities like Los Angeles County and other metropolitan areas.

Methods

Participants and Procedures

A total of 263 individuals who attended two community festivalsFootnote 1 in Los Angeles County in 2005 completed questionnaires. Twelve individuals were excluded from analyses due to incomplete or unusable data, leaving a final sample of 251 individuals. The final sample was 54.6% API (approximately 25% Chinese, 17% Filipino, 16% Japanese, 9% Korean, and 33% Other API), 21.1% White, non-Hispanic, 16.6% Hispanic, and 4.9% African-American, non-Hispanic; 56.6% female; nearly 70% employed; 69% with at least some college education; and had an average age of 43.6 years. The majority of respondents lived in areas (defined by Zip Code) that were more than 5 miles from a Zip Code in which a casino was located. Table 1 presents the sample demographic and background characteristics by API/Non-API ethnicity and gender.

Table 1 Mean (SD) or N (%) for demographics and background characteristics by ethnicity and gender
Full size table

The UCLA General Campus Institutional Review Board (GC-IRB) approved all study procedures prior to their implementation. Research assistants attending the two community festivals approached potential respondents and asked if they would be willing to participate anonymously in the study. After informed consent, those agreeing to participate completed a questionnaire that had no identifying information, which was returned to study personnel. Respondents were compensated with a bottle of water for their time.

Measures

Gambling Knowledge and Impact Survey

This survey included the following: (a) questions on demographic information (age, gender, ethnicity, education, and employment status); (b) one item assessing knowledge about problem/pathological gambling rated on a four-point scale ranging from 1 (almost nothing) to 4 (a lot), (c) one item each, rated on a three-point scale ranging from 0 (not at all) to 2 (a lot), assessing the impact of diabetes, depression, drug abuse, alcohol abuse, and problem/pathological gambling on the community; and (d) one item each, rated on a three-point scale ranging from 0 (not at all) to 2 (a lot), assessing the impact of problem/pathological gambling on domestic violence, traffic congestion, unemployment, and bankruptcy. For all items in the survey, except those assessing demographics, respondents could endorse a ‘don’t know’ response. This survey was developed in conjunction with the UCLA Health Services Research Center.

Statistical Analyses

Missing Data

Data from questions on knowledge, community impact, and consequences of problem/pathological gambling were considered missing in the case of non-response, but not in the case of a ‘don’t know’ answer. We employed group mean replacement for missing data on non-demographic items. The number of non-responses for all items analyzed was below 10%.

Preliminary Analyses

We conducted two preliminary one-way ANOVAs (data not shown; available from second author) for the main outcome variables by Asian subcategories (e.g., Chinese, Japanese, Korean, Filipino, etc.) and Non-Asian subcategories (e.g., White, Non-Hispanic, Hispanic, African-American) and found no differences between groups. Thus, we combined all Asian Subgroups into a single Asian category, and all Non-Asian subgroups into a Non-Asian category to maximize group N’s for analysis. Potential differences in demographic and background characteristics were evaluated between API and Non-API individuals using t-tests for continuous variables and χ2 for categorical variables.

‘Don’t Know’ Response Analyses

We excluded ‘don’t know’ responses from the main analyses, resulting in varied sample sizes for each analysis. We conducted two χ2 tests, one for ethnicity and one for gender, and three regression analyses for ‘don’t know’ responses using age, ethnicity, site, and gender as predictors to examine potential systematic patterns in ‘don’t know’ responses across all variables in our main analyses.

Main Analyses

Because the data were collected from two different festivals, site of data collection was included as a covariate in each of the main analyses. We conducted a 2 (male, female) × 2 (API, Non-API) ANCOVA for problem/pathological gambling knowledge. We examined differences in community impact and the consequences of problem/pathological gambling using 2 (male, female) × 2 (API, Non-API) MANCOVAs. Paired-samples t-tests were used to assess differences in ratings of community impact of diabetes, depression, drug abuse, alcohol abuse, and problem/pathological gambling for the entire sample.

Results

Preliminary Analyses

Asian-Pacific Islanders and non-API groups did not differ on mean age, gender distribution, level of education, employment, or percent living within 5 miles from a Zip Code in which a casino was located; however, females were significantly older (F [1,246] = 5.233, P < 0.05), and non-API females were more likely to be employed (χ2 = 8.780, P < 0.01).

‘Don’t know’ Response Analyses

Thirty-one individuals responded ‘don’t know’ on the problem/pathological gambling knowledge item. Ethnicity was unrelated to ‘don’t know’ responses; however, more females reported ‘don’t know’ than males on this item (χ2 = 4.469, P < 0.05). Fifty-five individuals responded ‘don’t know’ on at least one of the community impact items and 82 individuals reported one or more ‘don’t know’ responses on the items assessing problem/pathological gambling consequences. Ethnicity and gender were not related to ‘don’t know’ responses on these two sets of items.

Main Analyses

Means by gender and ethnicity for the 220 respondents included in the analysis of problem/pathological gambling knowledge are presented in Table 2. A 2 (male, female) × 2 (API, Non-API) ANCOVA, controlling for site, indicated a main effect for gender (F [1,215] = 7.27, P = 0.008), with females scoring slightly lower than males. Scores for males were closest to reporting knowing a ‘moderate amount’, and females’ scores were closest to reporting knowing ‘a little’ about problem/pathological gambling. Ethnicity and gender-by-ethnicity interaction effects were not significant (F [1,215] = 0.28, P = 0.595 and F [1,215] = 0.310, P = 0.578, respectively).

Table 2 Mean (SD) problem gambling knowledge, community impact, and impact on other problems by ethnicity and gendera
Full size table

Paired-samples t-tests for the 196 respondents with complete data showed that the community impact ratings for diabetes, depression, drug, and alcohol abuse were equal to that of problem/pathological gambling. Forty-one percent of respondents reported that problem/pathological gambling impacted the community ‘a lot’ and this response category was not related to gender or ethnicity. The 2 (male, female) by 2 (API, Non-API) MANCOVA for the community impact items revealed no multivariate effects for gender (F [5,187] = 0.50, P = 0.779) or ethnicity (F [5,187] = 0.81, P = 0.543); however, a multivariate interaction effect was found for ethnicity and gender (F [5,187] = 3.61, P = 0.004), with API males reporting lower impact of drugs (F [1,187] = 8.00, P = 0.005) and alcohol (F [1,187] = 13.33, P = 0.000) on their community than API females and Non-API males and females.

A 2 (male, female) by 2 (API, Non-API) MANCOVA for the ratings of the impact of problem/pathological gambling on domestic violence, traffic congestion, unemployment, and bankruptcy using 169 respondents with complete data on these items yielded no significant multivariate effects for gender (F [4,161] = 1.268, P = 0.29), ethnicity (F [4,161] = 0.466, P = 0.76), or their interaction (F [4,161] = 2.128, P = 0.08) (see Table 2). The mean rating of the impact of problem/pathological gambling on domestic violence, traffic congestion, unemployment, and bankruptcy for the entire sample was nearest to 1, which reflects the response ‘a little’.

Discussion

Our analyses did not indicate significant ethnic differences in knowledge of, perceived community impact of, and consequences of problem/pathological gambling. We did find a gender effect in one domain. Women reported lower levels of problem/pathological-gambling-related knowledge and were more likely to respond ‘don’t know’ on this item than males. Recent work [18] suggest that in California men are more likely to report any past year gambling than women; however, past year gambling in casinos is not significantly different between Asian males and females. Asian females may be exposed to more casino style gambling than women of other ethnicities, but show lower levels of problem/pathological-gambling-related knowledge relative to men and similar levels of problem/pathological gambling knowledge relative to Non-API women.

Forty-one percent of the sample (similar for API and Non-API) reported that problem/pathological gambling impacted their community ‘a lot’. The mean rating for the community impact of problem/pathological gambling in the entire sample was equal to that of the mean ratings for the community impact of diabetes, depression, drug, and alcohol abuse, all of which have been the subject of community prevention and treatment efforts. At present, though, the availability of treatment programs for problem/pathological gambling are limited to specialty, fee-for-service providers and state and county funding for treatment is essentially non-existent. Evidence-based public policy development to support prevention and treatment efforts for problem/pathological gambling are needed.

Although we obtained a negative finding regarding ethnic differences, this finding is important in light of some data that suggest that problem/pathological gambling disproportionately impacts the API community and casinos aggressively market to the API community. Because APIs are at greater risk than non-APIs, increasing awareness about problem/pathological gambling in this community is necessary. Public service announcements geared towards the API community should be employed to raise awareness regarding the signs and symptoms of problem/pathological gambling, the relationship between gambling problems and physical/mental health, and the psychosocial consequences of gambling problems. These announcements should emphasize the specific targeting of API populations by the casino industry [4–7, 19].

Educational campaigns targeted towards women may be necessary to help to raise their awareness regarding the development, course, and consequences of problem/pathological gambling. Such campaigns are necessary given that as many as one half of problem/pathological gamblers may be women [19] and women may more quickly progress from social to problem/pathological gambling relative to males [20, 21].

Several issues should be considered when interpreting the results. First, data were from a convenience sample from two Los Angeles County festivals. The use of convenience sampling limits the generalizability of data from the current study. It likely resulted in a more homogenous sample than one would get from random selection. However, given the dearth of data related to attitudes about problem/pathological gambling among APIs in Los Angeles County, and the lack of cross-cultural studies in the problem/pathological gambling literature in general, our data have some utility. Second, our sample was highly educated, with over 50% of the sample having attended at least some college. Nevertheless, many respondents were uninformed about problem/pathological gambling as reflected by the number of ‘don’t know’ responses received to survey items and the overall ratings for problem/pathological gambling knowledge. This lack of knowledge was unrelated to ethnicity; however, it appears that females report less knowledge of problem/pathological gambling than men and women were more likely to report a ‘don’t know’ response for gambling related knowledge. Third, the study did not include a measure assessing the extent to which respondents themselves engaged in gambling behavior, which may influence their attitudes towards gambling [18]. Some data do suggest an increased prevalence and risk for problem/pathological gambling among APIs relative to Non-APIs [10], but documenting differences in gambling behavior in addition to knowledge, community impact, and consequences of problem/pathological gambling may have helped in providing another potential covariate that could be importantly related to the variables in this study.