Men & Women Playing Games: Gender and the Gambling Preferences of Iowa Gambling Treatment Program Participants

Transcription

1 Journal of Gambling Studies, Vol. 22, No. 1, Spring 2006 (Ó 2005) DOI: /s Men & Women Playing Games: Gender and the Gambling Preferences of Iowa Gambling Treatment Program Participants Debi A. LaPlante Sarah E. Nelson Richard A. LaBrie Howard J. Shaffer Division on Addictions, Harvard Medical School Historically, gambling has been a predominantly male pastime; however, as legalized gambling has expanded, female participation has increased. Nevertheless, some research suggests that a divide remains between the play patterns of men and women. For example, research suggests that men gravitate towards casino table games and track betting and women are attracted to games such as bingo and casino slots. Researchers have hypothesized that play pattern disparities exist because of inherent differences between the natures of men and women. Using data from 2256 (1309 male) problem gambling treatment participants, this research examines the influence of gender on play patterns. We tested the ability of gender and a series of demographic, economic, and health-related factors to discriminate among three groups of gamblers with different game preferences: casino preferred, slots preferred, and non-institutional preferred. The results of multiple discriminant function analyses indicated that gender provided a minimal contribution to discrimination beyond that of specific demographic, economic, and health-related factors. This finding suggests that for understanding gambling patterns, gender is less informative than descriptive gambler profiles. KEY WORDS: gambling; pathological gambling; gender differences; psychosocial factors. Please address correspondence to Debi A. LaPlante, Division on Addictions, Cambridge Health Alliance, 10 Presidents Landing, 2nd Floor, Medford, MA 02155, USA. debi_laplante@hms.harvard.edu /06/ /0 Ó 2005 Springer Science+Business Media, Inc.

2 66 JOURNAL OF GAMBLING STUDIES Over time men have been more active gamblers than women (Shaffer, Hall, & Vander Bilt, 1999). That is, men are more likely to be gamblers than women and also more likely to develop gambling related problems than women (National Research Council, 1999). This stratification might be due in part to longstanding differences in the cultural acceptability/unacceptability of male and female gamblers. However, a progressively more egalitarian social milieu has made it more acceptable for both men and women to participate in gambling. This permissiveness encourages researchers to consider investigations of the relationships between gambling and gender and increasingly researchers are doing this with vigor (e.g., Getty, Watson, & Frisch, 2000; Grant & Kim, 2002; Hing & Breen, 2001; Hraba & Lee, 1996; McDaniel & Zuckerman, 2003; Potenza et al., 2001). One pertinent topic that has recently elicited a great deal of scientific interest is play habits, patterns, and gambling preferences. There are several studies of general gender differences in play. Men and women have been found, for example, to differ in how much they gamble and the variability of the games that they play (e.g., Spunt, Lesieur, Liberty, & Hunt, 1996). Researchers also have demonstrated that men and women tend to progress toward disordered gambling differently (Potenza et al., 2001; Shaffer, LaBrie, LaPlante, & Kidman, 2002; Tavares, Zilberman, Beites, & Gentil, 2001). Men frequently begin gambling early in life, report slow emergence of problems, and seek help well after developing problems; alternatively, women seem to telescope, or start gambling later in life, then rapidly develop a problem and seek help more quickly. 1 Researchers have documented more specific differences as well. For example, many studies show that women are more likely to play gaming machines and slots and men are more likely to play casino games (e.g., Hing & Breen, 2001; Ladd & Petry, 2002; Potenza et al., 2001; Winters & Rich, 1998). Recently Welte and colleagues (2002) found that men bet more on games of skill than women. The importance of examining play patterns might not be clear immediately: bankruptcy is bankruptcy whether it is due to casino card playing or slots. However, some researchers suggest that, similar to substances of abuse, some games might elicit different responses from individuals (Fisher & Griffiths, 1995; Hing & Breen, 2001; Oliveira & Silva, 2001). For example, although alcohol and cocaine are both common substances of abuse, the sedating nature of alcohol and

3 DEBI A. LAPLANTE ET AL. 67 energizing nature of cocaine elicit different psychoactive responses to each though some neurological responses are similar (Shaffer et al., 2004). Different types of gambling games might exhibit similarly disparate effects on individuals: some suggest that slots and video games accelerate the development of problem gambling (e.g., Fisher & Griffiths, 1995; Oliveira & Silva, 2001). This possibility coupled with men and womenõs tendencies to play different games suggests the need for closer scrutiny of game types and ultimately the consequences of play preferences. There are many explanations for gender play patterns. Among the explanations put forth are: genetics (Winters & Rich, 1998), social norms (Ladd & Petry, 2002), motivations (Potenza et al., 2001; Trevorrow & Moore, 1998), impulsivity (Langewisch & Frisch, 1998), and finances (Hing & Breen, 2001). Perhaps most common are explanations for differentiation that rely on stereotypes of men and women. For example, men prefer the thrill of gambling and hence play casino games, but women prefer to gamble to escape from reality and therefore like non-strategic games such as slots. It is important, however, not merely to explain gender differences by consulting our personal assumptions about the way men and women are. Such gross generalizations about men and women are unlikely to maintain any predictive power in specific settings or over time as gender roles evolve and change. One way to resolve this line of inquiry is to develop a more specific representation of the individuals who gravitate towards specific games. Wolfgang (1988) suggested that compared to personality factors, social norms related to gambling and gender, for example, have less to do with individualsõ interests in gambling. Researchers recently have begun to explore the complexities of play preferences. For example, marital status, race, legal problems, and income are all predictors of play preferences (Ladd & Petry, 2002; Potenza et al., 2001; Tavares et al., 2001). Petry (2003) recently published a catalog of demographic, economic, and psychological differences for five game types (i.e., track, sports, cards, slots, and scratch/lottery). Taking this kind of research a step further, it is possible to construct unique profiles of people with different game preferences based on individuating characteristics. Doing so allows for greater specificity and predictive ability than relying on gender differences alone. Gender differences might merely provide a semi-predictive heuristic that is not necessarily accurate for all women

4 68 JOURNAL OF GAMBLING STUDIES or all men. Indeed, some men prefer slots and some women prefer blackjack. We can develop a more complete understanding of the nature of supposed gender differences by creating gambling profiles based on demographic, economic, and health-related factors that might better explain individualsõ gravitation towards particular games. This will help prevention programs and treatment providers address the nature of what appear to be gender differences in play preferences, but might actually reflect important and predictable psychosocial precursors that operate across genders the gender as proxy hypothesis. That is, we will better understand whether gender is a direct risk factor or a proxy for factors commonly associated with gender. PRESENT STUDY Research and conventional wisdom frequently note that men and women tend to gravitate toward distinct games. To date, investigators have explained these differences by isolating stereotypical gender differences and engaging in post-hoc rationalization (e.g., men play interactive games because men like competition). However, a better route to understanding play patterns might be to identify unisex gambler profiles into which either men or women fit and then test those profiles for their ability to predict individualsõ play patterns: this strategy permits testing the extent to which gender influences play patterns compared to other factors. To accomplish such a test, in this study we compare the discriminative power of gender and demographic, economic, and health-related factors that are associated with the play patterns of Iowa Gambling Treatment Program (IGTP) participants. By focusing on treatment participants, we will better be able to modify treatment strategies to address personal needs. The IGTP is a network of local service providers that hold contracts with the Iowa Department of Public Health. This network provides an outpatient program with a variety of diagnostic and primary gambling treatment services. The IGTP amassed participant data for gamblers and concerned others (e.g., family members, significant others, or close friends of problem gamblers) over 4 years ( ). 2

5 DEBI A. LAPLANTE ET AL. 69 METHOD Participants The full IGTP dataset contains first admission records for 2356 identified gamblers who may have had more than one course of treatment with the IGTP. Of these 1367 are men and 989 are women. Because the subsequent analyses are pertinent to individualsõ gambling patterns, we excluded individuals who did not report gambling on any type of game (n = 52). We also restricted the analyses to individuals who reported gambling within the last 6 months, so we excluded an additional 48 self-identified non-gamblers. The total n for analyses was 2256 (1309 men and 947 women). Materials The IGTP completes an assessment within five working days of initial contact or service initiation. A person gambling excessively is determined in need of gambling treatment services if the person meets the criteria from any one of the following three assessment tools: the South Oaks Gambling Screen (SOGS) (Lesieur & Blume, 1987), the Gamblers Anonymous (GA) 20 Questions, or the diagnostic criteria for pathological gambling in the current American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association, 1994). Our data set was censored of individual identifiers but contained the other information collected by the IGTP at intake: demographic information (which includes date of birth, sex, race or ethnicity, and sufficient identification of the referral source), presenting problem, gambling history (which includes type, amount, frequency and duration of gambling activity), legal history (which describes any involvement with the criminal justice system), medical and health history, psychological history and mental status, family history, educational history, employment history, and other relevant information that would assist clinicians in formulating a treatment plan for the client. Procedures This article reports the results of an analysis of IGTP Assessment/ Intake data. Full details of the IGTP data collection and data

6 70 JOURNAL OF GAMBLING STUDIES management procedures and basic characteristics are available in Shaffer et al. (2002). Identifying Player Preferences. Individuals reported the percentage of money they spent in the past 6 months on each of 12 types of gambling (i.e., casino table games, slots, keno, video, non-casino cards, bingo, scratch tickets, lotteries, racetracks, sports, stocks, and all other). We labeled individuals who spent 90% or more of their gambling money on a single game as pure players of that game: pure player typology for these individuals matched their reported preferred game. For example, we labeled individuals who spent 90% or more of their money on slots, Slots Pure. We labeled individuals who reported spending 50% of their money on one game and 50% of their money on another game as 50/50. We labeled individuals who reported playing only two games, but spent between 51% and 89% of their money on one of the games, as Bi-game players: Bi-game typology for these individuals matched their preferred game. We labeled individuals who reported spending money on three or more games Multi-game players. For this article, we were particularly concerned with games on which men and women spent a disproportionate amount of time and money. Men and women reported the percentage of time they spent wagering in the last 6 months on the 12 gambling activities. We found that, compared to women, treatment seeking men spent more of their money wagering on casino games (t( ) = 13.57, p < 0.001), 3 playing non-casino cards (t( ) = 4.39, p < 0.001), playing the lottery (t( ) = 2.47, p < 0.05), betting at the race track (t( ) = 7.01, p < 0.001), betting on sports (t( ) = 8.32, p < 0.001), playing the stock market (t( ) = 4.06, p < 0.001), and playing non-specified games indicated as Other on the intake form (t( ) = 2.55, p < 0.05). Compared to men, treatment seeking women spent more of their money wagering on slot machines (t( ) = )15.31, p < 0.001) and bingo (t(1354.4) = )4.51, p < 0.001). Men and women did not differ in the amount of time spent on keno (t( ) = 1.73), video (poker/keno/blackjack) (t(2254) = 0.07), or scratch tickets (t(2254) = 0.35). Because we were concerned with explaining gender differences in play patterns, we did not include these three games (i.e., keno, video, scratch) in subsequent play preference analyses.

7 DEBI A. LAPLANTE ET AL. 71 Because the labeling process resulted in a large number of player preferences, we reduced the number using a combination of both theoretical reasoning and analytical evidence. Based on patterns of overlap (e.g., if the pure bingo playersõ second choice commonly was slots and if the pure slot playersõ second choice commonly was bingo, we combined these groups), we determined that we could create a composite variable combining non-casino card players, racetrack betters, sports betters, and stocks players. We called these individuals non-institutional players (i.e., gamblers who prefer engaging in gambling activities of perceived skill other than the skill games offered by casinos and other gaming institutions). This iteration resulted in a larger than desirable number of player preferences and prompted us to test whether or not we could combine pure players of a given game with bi-game players of the same game. Comparative analyses (i.e., t-tests and v 2 ) revealed scant demographic, economic, and health-related differences between pure and bi-game players who preferred the same game; consequently, we combined these types. This redefinition resulted in seven player preferences: Casino Preferred (CP), Slots Preferred (SP), Bingo Preferred (BP), Lottery Preferred (LP), Non-institutional Preferred (NP), 50/50, and Multi-game (MG). In all, there were 1180 SP (530 male), 400 MG (271 male) players, 244 CP (211 male), 109 NP (105 male), 69 50/50 (45 male), 16 BP (4 male), and 14 LP (10 male). Profile Variables. The IGTP intake/assessment procedure collects data for a large number of variables. To improve the focus of the study and increase reliability we created composite variables from conceptually similar items. A companion paper (Nelson, LaPlante, LaBrie, & Shaffer, in press) provides full details about the empirical procedures used to build these composite variables. In brief, we identified what we thought were important conceptual domains and identified variables from the intake/assessment form that might measure those domains. To test the adequacy of these domain constituents, we performed principal components analyses and reliability checks. For this paper, we utilized five composite variables that this process generated: socioeconomic status (i.e., highest grade completed, work status, personal income, household income, occupation), loss and debt (i.e., total debt, gambling debt, credit card debt, average lost weekly, most lost in a week), criminal history (i.e., ever arrested,

8 72 JOURNAL OF GAMBLING STUDIES arrested in past year, ever incarcerated), comorbid chemical addictions (i.e., illicit drug use, alcohol use, tobacco use), and comorbid behavioral addictions (i.e., food abuse, compulsive work, compulsive sex, compulsive spending). In addition to these composite variables, we examined the impact of other variables that we wanted to examine independently for theoretical reasons: gender, age, military experience (i.e., yes or no), percent of gambling money lost illegally, frequency of gambling, gambling background (i.e., family as first gambling partner or not), whether family accepted gambling growing up (i.e., no, neutral, yes), prior gambling treatment sought (i.e., yes or no), alcohol or drug treatment (i.e., yes or no), family responsibilities (i.e., marital and parental status), and game variety (i.e., number of games played). Finally, we also combined the information on age when the person first gambled, age when gambling first became a problem, and age at admission to the IGTP into a measure of the concentration of the course of problem gambling at different age levels (henceforth, Gambling Development). On this variable, low scores indicate the progress from gambling to treatment-seeking took place at an early age, high scores indicate the progress took place later in life, and intermediate scores result from the course spanning age groups. A factor analysis of the constituent variables revealed a single factor and a reliability analysis revealed an alpha of 0.82 for the composite variable. Analytic Strategy. We investigated whether demographic, economic, and health-related variables might effectively discriminate player preferences. Basic univariate differences by player preferences for gender and all profile variables are presented below. Gender by player preference analyses provide the information usually found in literature on gender differences. We sought to expand from this to develop multivariate player preference profiles. Our first step to achieve this profiling was to use a Stepwise Multiple Discriminant Function Analysis (MDA) to identify the demographic, economic, and health-related profile variables that successfully discriminate the player preferences. We did not include gender in this analysis. We compared the outcome of this analysis to an MDA in which simultaneously we entered gender along with the profile variables included in the final model of the stepwise analysis to determine how much gender improved the discriminative accuracy of the profile model. Doing so allowed us to

9 DEBI A. LAPLANTE ET AL. 73 determine the proportion of variance gender accounted for beyond other profile variables. Gender Differences by Player Preference RESULTS We first analyzed each of the identified seven player preferences by gender and found that gender was irrelevant to whether an individual was a 50/50 (v 2 (1) = 1.51, p > 0.1) or an LP (v 2 (1) = 1.04, p > 0.1) player. However, a greater proportion of men than women were identified as CP (v 2 (1) = 90.93, p < 0.001), NP (v 2 (1) = 69.01, p < 0.001), and MG (v 2 (1) = 18.89, p < 0.001) players. A greater proportion of women than men identified as SP (v 2 (1) = , p < 0.001) and BP (v 2 (1) = 7.22, p < 0.01) players. Subsequent multivariate analyses further examine gender differences for individuals who established a preference for a specific game. Because BP only had 16 people, MG established no clear preference for a single game, and 50/ 50 and LP distributed about equally across genders, we did not include these groups in the multivariate analyses. The analyses described below used the aforementioned demographic, economic, and health-related variables to discriminate CP, SP, and NP gamblers. Basic Profile Variable Differences by Player Preference Individuals who reported preferences for specific games (i.e., CP, SP, and NP) varied on a number of demographic, economic, and healthrelated variables. Using the appropriate univariate tests (i.e., F, v 2 ), we observed significant differences across player preferences for family responsibilities (F(2, 1529) = 18.82, p < 0.001), criminal history (F(2, 1530) = 5.9, p < 0.01), gambling development (F(2, 1530) = , p < 0.001), age (F(2, 1522) = 74.14, p < 0.001), percent lost illegally (F(2, 1528) = , p < 0.001), gambling background (v 2 (2) = 22.76, p < 0.001), and prior help sought for alcohol or drug use (v 2 (1) = 10.57, p < 0.01). Table 1 presents the details of these differences. Profiling Player Preferences The composite and selected profile variables (i.e., SES, family responsibilities, loss and debt, gambling background, criminal history,

10 74 JOURNAL OF GAMBLING STUDIES chemical addictions, behavioral addictions, gambling development, age, military status, game variety, percent lost illegally, frequency of wagers in past 30 days, family acceptance of gambling, prior help sought for gambling, and prior help sought of alcohol or drug use) were entered into an MDA in the order of their successive contribution to discrimination. After seven variables entered the MDA no other variable contributed significantly to discrimination. Discrimination required information on two dimensions (i.e., discriminant functions). In the order of entry, significant discriminators were: percent lost illegally, gambling development, age, gambling background, chemical addictions, SES, and family responsibilities. These seven variables correctly classified 59% of the sample. Classification accuracy varied by player preference. The model accurately classified approximately 71% of CPs, 66% of SPs, and 40% of NPs. Scrutiny of the functions at group centroids (i.e., the discriminant scores obtained from applying the functions to the central values, the group means, of the variables in the most typical group member) revealed that function 1 served to isolate NPs from both CPs and SPs but more so from SPs; function 2 isolated CPs from SPs and NPs but more so from NPs (see Figure 1). Variables that loaded highly on function 1 were percent lost illegally and substance use: Table 1 shows that NPs were much more likely to accumulate illegal losses and, though not statistically significant, less likely to report current chemical addictions. Variables that loaded most highly on function 2 were gambling development, age, family responsibilities, gambling background, and SES. Table 1 shows that CPs were more likely to have short consistent gambling development patterns. SPs were most likely to report being older, reported the most family responsibilities, reported first gambling with family, and reported lower SES. Gender as a Predictor When we entered the seven significant profile variables from the first MDA model and gender into an MDA simultaneously, the overall discriminative accuracy only improved 2% from 59% to 61%: accurately classifying 76% of CPs, 68% of SPs, and 40% of NPs. The two functions associated with the Enter analysis discriminated the player preferences similarly to the functions of the Stepwise analysis. In this analysis, gender loaded on function 2, but did not substantially improve the predictive accuracy of the aforementioned seven variables.

11 DEBI A. LAPLANTE ET AL. 75 Table 1 Means (SD) and Proportions for Profile Variables by Player Preference Variable CP (n = 244) SP (n = 1180) NP (n = 109) SES 0.15 (0.73) 0.02 (0.7) 0.15 (0.78) Family responsibilities** 3.22 (1.53) 3.71 (1.21) 3.25 (1.5) Loss and debt 0.12 (0.69) 0.08 (0.73) )0.1 (0.93) Criminal history* )0.01 (0.45) )0.04 (0.43) 0.11 (0.54) Chemical addiction )0.04 (0.69) )0.01 (0.6) )0.15 (0.56) Behavioral addiction )0.06 (0.5) )0.01 (0.58) )0.05 (0.6) Gambling development** )0.47 (0.71) 0.26 (0.82) )0.6 (0.76) Age** (11.6) (10.7) (11.41) Number of games 1.89 (1.09) 1.73 (1.05) 1.9 (1.22) Percent lost illegally** 2.08 (10.87) 1.47 (10.69) (47.37) Frequency of gambling (1.16) (1.16) 11.7 (1.44) in past 30 days Family acceptance 1.12 (0.9) 1.06 (0.9) 1.19 (0.87) of gambling Gambling background 28.3% 42.5% 28.4% (family first)** Prior drug 20.9% 18.4% 31.2% or alcohol treatment (yes)* Military experience 22.1% 18.5% 18.3% (yes) Prior gambling treatment (yes) 24.2% 18.6% 18.3% Note: N varies slightly depending on missing data; CP, Casino Preferred; SP, Slots Preferred; NP, Non-institutional Preferred. *p < 0.01; **p < DISCUSSION Gender does not hold as much discriminatory power for distinguishing gambling preferences as many have thought. The findings from these analyses suggest that personal demographic, economic, and health-related profiles provide essential distinguishing information for

12 76 JOURNAL OF GAMBLING STUDIES Figure 1 Plot of functions at group centroids by player preference. Structure matrix profile variables associated with discriminant functions indicated in parentheses. 0.6 Non-Institutional 2.251, Discriminant Function 2 (gambling life event periods, age, family responsibilities, gambling background, & SES) Slot , Casino 0.309, Discriminant Function 1 (percent lost illegally & chemical addiction) gamblers who prefer specific games. Importantly this finding encourages scientists and public policy makers to avoid the tendency to overgeneralize the importance of specific demographic characteristics, such as gender. Such over-generalization values simplicity at the cost of precision in the identification of individual preferences, risk factors, and tendencies related to gambling. This loss of precision ultimately could contribute to poor treatment planning and prevention development because some of the specific mechanisms related to play preferences are masked. For example, some men whose profiles more closely resemble slot players than non-casino players erroneously might receive treatment that focuses more heavily on presumable male concerns (e.g., the roles of illegal gambling and chemical addictions), rather than the impact of family and age characteristics. Such individuals might benefit more from greater emphasis on the latter factors, despite the conventional wisdom that commonly associates family and age factors with female gamblers.

13 DEBI A. LAPLANTE ET AL. 77 Our analysis revealed that using more comprehensive player profiles resulted in good discriminative accuracy overall and discriminated more specific player preferences simultaneously; however, despite a fair amount of unexplained variance remaining after the entry of the seven profile variables, gender did not contribute substantially to classification accuracy over and above detailed player profiles. The primary implication of these findings is that although gender seems to provide a simple post-hoc, indirect and rationalized division by which we might predict play patterns and tailor prevention or treatment efforts, more comprehensive profiles contain important individuating information related to game choice that gender alone cannot capture. As a result, by examining profiles, researchers and treatment providers will gain a better understanding of people who already have gambling problems than by invoking gender based characterizations. They also will develop a better idea of how to identify and potentially prevent people who might be at-risk for gambling problems from developing any or severe gambling problems. Some limitations of this strategy are obvious. For example, time and resources might argue against comprehensive profile development in practice. Individuals in need are often in immediate need and might not have the luxury of waiting for profile development. Similarly, treatment providers frequently have little time and financial resources. If a single predictor provides an approximation for risk, even if less adequate and accurate than other methods, some treatment providers might believe that using that single predictor still might be more practical. Despite its expediency, however, this strategy will necessarily limit the development of a nascent field and the effectiveness of clinical practice. Although clinicians might not be able to create full profiles, what they do choose to measure ought to be as informative as possible. This study reveals that gender is less informative than the conventional wisdom suggests. Finally, the link between specific games and the progression of disordered gambling is tenuous at best. Although a large amount of anecdotal evidence identifies slot machines and similar electronic gambling devices as a quick route to unhealthy gaming, more research is necessary to examine the veracity of that assumption. More information is necessary to better determine the value of knowing individualsõ gaming preferences to specific prevention and treatment strategies.

14 78 JOURNAL OF GAMBLING STUDIES Methodologically it is important to keep in mind that treatment seekers generally, and IGTP treatment seekers specifically, might not be representative of gamblers in general. Even so, the findings presented here are seminal: this study presents the first player profiles and establishes a first step towards understanding the value of building universal player profiles. That said, other profile variables not measured by the IGTP intake/assessment forms might be even more important than those identified in this paper. Moving forward, we suggest that researchers and clinicians further explore the theoretical and clinical usefulness of player profiles. It would be interesting to determine whether such profiles might prospectively predict gaming preferences. However, it will be important to identify the durability of profilesõ discriminatory power. Although gender has taken on an important role in gambling research and social research in general, its own durability for discriminating is questionable. It is important that the context and nature of what gender represents is identified. Until then, gender is and will remain an ambiguous construct that itself is quite malleable over time and across social settings. The usefulness of gender as a marker for risk and means of organizing treatment only remains for as long as gender roles and representations remain fixed. ACKNOWLEDGMENTS This project required the collaboration of many people. We extend thanks to Christine Reilly, Christine Thurmond, Gabriel Caro, Lymari Graciano, Mike Stanton, Alexa Rubenstein, and the Iowa Gambling Treatment Program. This project was supported, in part, with funding from the Iowa Department of Public Health and the National Center for Responsible Gaming. NOTES 1. Nelson, LaPlante, LaBrie, & Shaffer (in press) indicate that this effect might primarily relate to age of initiation, rather than gender. 2. Some descriptions of the IGTP methods and procedures were derived from Shaffer et al. (2002). 3. Equal variances not assumed in analyses.

15 DEBI A. LAPLANTE ET AL. 79 REFERENCES American Psychiatric Association (1994). DSM-IV: Diagnostic and statistical manual of mental disorders. Fourthnd ed. Washington, D.C: American Psychiatric Association. Fisher, S., & Griffiths, M. (1995). Current trends in slot machine gambling: Research and policy issues. Journal of Gambling Studies, 11(3), Getty, H. A., Watson, J., & Frisch, G. R. (2000). A comparison of depression and styles of coping in male and female GA members and controls. Journal of Gambling Studies, 16(4), Grant, J. E., & Kim, S. W. (2002). Gender differences in pathological gamblers seeking medication treatment. Comprehensive Psychiatry, 43(1), Hing, N., & Breen, H. (2001). Profiling Lady Luck: An empirical study of gambling and problem gambling amongst female club members. Journal of Gambling Studies, 17(1), Hraba, J., & Lee, G. (1996). Gender, gambling and problem gambling. Journal of Gambling Studies, 12(1), Ladd, G. T., & Petry, N. M. (2002). Gender differences among pathological gamblers seeking treatment. Experimental Clinical Psychopharmacology, 10(3), Langewisch, M. W., & Frisch, G. (1998). Gambling behavior and pathology in relation to impulsivity, sensation seeking, and risk behavior in male college students. Journal of Gambling Studies, 14(3), Lesieur, H. R., & Blume, S. B. (1987). The South Oaks Gambling Screen (SOGS): A new instrument for the identification of pathological gamblers. American Journal of Psychiatry, 144(9), McDaniel, S. R., & Zuckerman, M. (2003). The relationship of impulsive sensation seeking and gender to interest and participation in gambling activities. Personality & Individual Differences, 35(6), National Research Council (1999). Pathological gambling: A critical review. Washington D.C: National Academy Press. Nelson, S. E., LaPlante, D. A., LaBrie, R. A., & Shaffer, H. J. (in press). The proxy effect: Gender and gambling problem trajectories of Iowa Gambling Treatment Program participants. Journal of Gambling Studies. Oliveira, M. P. M., & Silva, M. T. A. (2001). A comparison of horse-race, bingo, and video poker gamblers in Brazilian gambling settings. Journal of Gambling Studies, 17(2), Petry, N. M. (2003). A comparison of treatment-seeking pathological gamblers based on preferred gambling activity. Addiction, 98(5), Potenza, M. N., Steinberg, M. A., McLaughlin, S. D., Wu, R., Rounsaville, B. J., & OÕMalley, S. S. (2001). Gender-related differences in the characteristics of problem gamblers using a gambling helpline. American Journal of Psychiatry, 158(9), Shaffer, H. J., Hall, M. N., & Vander Bilt, J. (1999). Estimating the prevalence of disordered gambling behavior in the United States and Canada: A research synthesis. American Journal of Public Health, 89(9), Shaffer, H. J., LaBrie, R. A., LaPlante, D. A., & Kidman, R. C. (2002). The Iowa department of public health gambling treatment services: Four years of evidence. Boston: Division on Addiction, Harvard Medical School(No ). Shaffer, H. J., LaPlante, D. A., LaBrie, R. A., Kidman, R. C., Donato, A., & Stanton, M. V. (2004). Toward a syndrome model of addiction: Multiple manifestations, common etiology. Harvard Review of Psychiatry. Spunt, B., Lesieur, H. R., Liberty, H. J., & Hunt, D. (1996). Pathological gamblers in methadone treatment: A comparison between men and women. Journal of Gambling Studies, 12(4), Tavares, H., Zilberman, M. L., Beites, F. J., & Gentil, V. (2001). Gender differences in gambling progression. Journal of Gambling Studies, 17(2), Trevorrow, K., & Moore, S. (1998). The association between loneliness, social isolation and womenõs electronic gaming machine gambling. Journal of Gambling Studies, 14(3), Welte, J. W., Barnes, G. M., Wieczorek, W. F., Tidwell, M. -C., & Parker, J. (2002). Gambling participation in the U.S. Results from a national survey. Journal of Gambling Studies, 18(4),

16 80 JOURNAL OF GAMBLING STUDIES Winters, K. C., & Rich, T. (1998). A twin study of adult gambling behavior. Journal of Gambling Studies, 14(3), Wolfgang, A. K. (1988). Gambling as a function of gender and sensation seeking. Journal of Gambling Behavior, 4(2),

17 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.