2012-Prevention of problem gambling.pdf

36

Nower and Blaszczynski (2010) found a positive, linear relationship between gambling problem severity and specific limit-setting before the start of a gambling session. In essence, ‘at risk’ and problem gamblers were much less likely than regular gamblers to use a pre-commitment strategy when it was optional.

65

avenues for excessive gambling have been closed, is the degree to which the technique likely has broader utility for problem and non-problem gamblers.37

Eliminating Reward/Loyalty Cards or Changing Their Parameters An additional consideration in relation to smart cards and pre-commitment concerns their increasing incorporation within Reward Cards or Loyalty Cards (also sometimes known as Player Cards). In jurisdictions with a single monopolistic provider of gambling it is unclear why a “loyalty” card is needed. In these jurisdictions, Reward cards and ‘comps’ simply serve to reward people for their amount of gambling so as to induce them to gamble even more (which they successfully do) (Lucas & Bowen, 2002; Lucas & Kilby, 2008; Suh, 2012). This is obviously not conducive to responsible gambling, and is a practice that would not be acceptable for alcohol or tobacco consumption. Although there may be better justification for loyalty/reward cards when there are multiple providers for customers to choose between, rewarding people for the amount they gamble is still incompatible with efforts to constrain gambling behaviour through pre-commitment. Furthermore, in addition to promoting further gambling, these rewards/’comps’ likely helps heavy gamblers and problem gamblers to reconcile their losses. Thus, reward/loyalty cards should not be combined with pre-commitment technologies. Obviously, there are inefficiencies in having two cards rather than one.

The solution to this requires either a non-carding biometric ID system for both, or to change the typical parameters of Reward/Loyalty cards. If Reward/Loyalty cards are to be combined with pre-commitment, then they should be rewarding responsible play, rather than amount of play. One way in which this could be operationalized is for players to receive no points after exceeding a reasonable operator-set daily level of spending, and/or losing points after exceeding this threshold. Another possibility is for players to be unable to collect their reward points if they have exceeded their pre-commitment level (the problem with this latter strategy is that it will tend to encourage high pre-commitment levels). Providers could also award points for socially responsible and self-protective behaviour, such as viewing educational resources, completing personal risk assessments, or comparing personal consumption patterns to normative standards. Another useful strategy would be to use the data concerning the player’s gambling behaviour to proactively alert the player to patterns of behaviour that are associated with future problems (see Automated or Mandated Intervention Section later in this document). As things currently stand, providers collect such data from the Reward cards and actively respond to it through increasingly strong inducements to gamble more.

Within the data base are clear indicators of frequency, duration, average bet size, and net amount won or lost. If for no other reason, operators having direct knowledge of increased potential for harm and failing to act in 37

One of the defining features of problem gambling is ‘impaired control’ (Neal et al, 2005). Hence, it is naïve to think that simply providing problem gamblers with information about their gambling and/or ‘choice’ will be sufficient to curb their gambling.

66

the interests of protecting the patron fails to meet contemporary standards for corporate social responsibility (Hancock, Schellnick, & Schrans, 2008; Smith & Rubenstein, 2011). At worst, it might be grounds for legal liability from a ‘duty of care’ perspective.38 A final consideration is that widespread adoption of a Reward/Loyalty Cards would also allow these cards to be used for casino entry, so as to strengthen the self-exclusion programs for those few countries that currently do not require ID to enter casinos.

Operator Imposed Maximum Loss Limits In many gambling venues, there is no limit to the amount a patron can bet and/or lose over the course of a gambling session. When maximum loss limits are imposed on patrons it typically occurs in the context of government-owned/operated online gambling sites. However, in most cases the limits are very high, thus affording very little true protection for most people39. A few exceptions exist. For example, the New Zealand online lottery requires players to set weekly and monthly spending limits, with a maximum of NZ $150/week and $300/month (New Zealand Lotteries Commission, 2012). It is very rare for land-based venues to have maximum loss limits. An important historical exception to the use of operator imposed maximum loss limits is the state of Missouri, which from 1994 – 2008 had restricted each gambler’s losses to a maximum of $500 during two hour ‘excursions’ on its 11 riverboat casinos (patrons could buy no more than $500 in gambling chips for the slot machines and table games). There is no information on the effectiveness of this measure other than the Missouri casino industry repeatedly pointing out that their revenues were much smaller than competing riverboat casinos in neighbouring jurisdictions (Brokopp, 2006; Volkmann, 2008). The most important current example is found in Norway.

Since February 2009 the use of smart cards has become mandatory in Norway for EGMs, sports betting, and online gambling. Norsk Tipping limits the amount that can be put on the card to 400 kroner per day (~$67 U.S.) or 2200 Kroner per month. Although promising in theory, there has been no evaluation of this operator-imposed limit.

38

A 2009 newspaper story in the province of Ontario accessed Reward card information for a slots-at-racetrack facility and found that the highest recorded annual loss was $701,117. This member attended the venue 294 times over the course of the year. The 10 program members recording the highest losses averaged $455,555 for the year, or $38,796 per month. Each dollar lost was recorded in the loyalty program data base, and the operator applied direct knowledge of mounting losses to the awarding of ‘comps’ (Priest, 2009). 39

For example, British Columbia allows players to transfer $9,999 per week into their online gambling account.

67

Proactive Intervention with ‘At-Risk’ Gamblers Problem Gambling Awareness Training for Employees of Gambling Venues Problem gamblers in gambling venues are known to display reliable behavioural cues (i.e., anger, repeated withdrawals from ATMs, etc.) (Allcock et al., 2002; Delfabbro, Osborn, Nevile, et al., 2007; Delfabbro, Borgas, & King, 2011; Schellenick & Schrans, 2004). Accordingly, in the past 10 years problem gambling awareness training has been initiated in many countries to increase employee recognition of problem gambling among patrons and to direct these patrons to appropriate treatment resources. Programs are variously delivered by venue owners/operators, departments of health/addiction agencies, contracted companies, or combinations thereof. Program design tends to be based on collaborative consultation among government, gambling industries, and prevention/treatment agencies. Staff training for venue staff is mandatory in several jurisdictions, and is sometimes also required of EGM site holders/staff and lottery retailers. Front line employees at casinos typically receive a one-time knowledge and skill development session to understand and recognize problem gambling behaviours in patrons so as to alert their supervisors to these individuals.

More extensive training is typically provided for supervisory and management personnel at casinos, whose responsibilities include approaching the identified individual to offer immediate crisis management or treatment referral. In the late 1980s, Holland Casinos was the first gambling operator to provide an intensive training and education program for employees (d’Hondt, 2007). Manitoba, Canada was one of the first North American jurisdictions to implement an employee training program, beginning in 1998. Since then all Canadian provinces have implemented either mandatory or voluntary programs. Awareness training for employees also occurs in many other countries. In South Africa, the National Responsible Gambling Program (a collaborative public/private initiative), encourages and provides voluntary industry staff training. Harrahs, in the United States, has operated ‘Operation Bet Smart’ for several years (Harrahs Entertainment, 2004). As of 2003, staff training became part of the American Gaming Association’s Code of Conduct for Responsible Gaming. The European Casino Association’s Code of Conduct has similar provisions, ratified by the 20 member countries in January 2006. The World Lotteries Association Code of Conduct also includes a problem gambling employee training component.

New Zealand’s ‘Host Responsibility Training’ (a model of responsible gambling on which venue licensing is conditional) includes a mandatory staff training component. Research on the effectiveness of training programs is limited. The Addictions Foundation of Manitoba found that 98% of 1,550 video lottery terminal (VLT) site-owners and employees reported finding the training useful (Smitheringale, 2001). Employees of three casinos in Quebec reported improved knowledge about gambling and increased confidence in identifying gamblers in crisis (Giroux et al., 2008). However, this study was limited by a response rate of only 32%. Ladouceur et al. (2004) found that VLT retailers in Quebec reported greater confidence in recognizing and addressing problem gambling after receiving a 2 hour problem gambling awareness workshop, and also reported approaching problem gamblers more

68

frequently than new retailers who had not yet attended the workshop. In a later study, Dufour, Ladouceur & Giroux (2010) found that 2.5 hours after employees at VLT venues were exposed to a training session they showed improved attitudes toward problem gamblers, increased knowledge about how to help, and increased propensity to provide information pamphlets to VLT patrons and to talk to them about helpful resources. However, 8 months later, although the improved attitudes and knowledge were partly maintained, the behavioural changes were not. In a very recent study, LaPlante et al. (2012) found that training produced an improvement in responsible gambling knowledge among gambling industry employees one month following the training program (n = 217 at baseline and 116 at follow-up). A comparable, well researched initiative is training of alcohol servers to not serve intoxicated patrons. A systematic review of this evidence shows several instances where this training has resulted in the desired effect, but just as many instances where compliance with the training has been poor (Ker & Chinnock, 2006).

Some of the main factors interfering with the effectiveness of this training include the likelihood that intervention will compromise profits; the voluntary nature of the training (in some jurisdictions); the lack of enforcement by regulators; and a low-skilled work force with high turnover and personal drinking habits that are inconsistent with these interventions (Ker & Chinnock, 2006; Mosher, Toomey, Good, Harwood, & Wagenaar, 2002; Reiling & Nusbaumer, 2006). It is important to note that all of these barriers to compliance also apply to the gambling industry (Dangerfield, 2004; Hing & Nuske, 2011; Shaffer et al., 1999; Shaffer & Hall, 2002; Williams & Wood, 2004, 2007a; Guttentag, Harrigan, & Smith, 2012). Screening prospective gambling employees for problem gambling may be a policy initiative that could be helpful in addressing hesitancy to intervene because of inconsistency with personal practices. Automated or mandated intervention with ‘at risk’ players is another policy with potential to reduce harm (Hing & Nuske, 2011). Automated or Mandated Intervention Swiss casinos train employees in the use of a behavioural checklist to detect patrons who are likely experiencing problems with gambling. If checklist criteria are met they are obliged to approach the patron to ascertain whether either a visit limitation or ban (voluntary or involuntary) is warranted (Hafeli, 2009).

An evaluation of the Swiss program in 2006 found that there were 1,543 people approached by casino staff in the 6 casinos. However, a) this roughly translates into only 1 approach for every 1,000 casino admissions; and b) the majority of these ‘approaches’ consisted of a single conversation. Moreover, only 16.6% of people who registered for a self-exclusion ban in 2006 had been involved in this ‘early detection’ system (Hafeli & Lischer, 2009). A better on-site intervention system exists in the Netherlands. The requirement to show ID also allows Holland Casino to track the frequency of casino visits. If the computer indicates a significant increase in visitation frequency or 20 visits a month over three consecutive months, then the person is automatically approached to see whether they would like to sign a visit

69

limitation or self-exclusion contract (Bes, 2002). Only 18.5% of these approaches are perceived negatively by the patron (Bes, 2002). In 2004, a total of 21,360 interviews with patrons resulted in 3,155 visit restrictions and 4,423 admission bans (Holland Casino, 2009). Indirect evidence of the utility of Holland Casino’s approach lies in the fact that the number of people seeking help for problem gambling from the official social services is only 50% of what these numbers were in 1995 (Holland Casino, 2009); patron surveys show that only 5% of patrons at any given time are pathological gamblers (2% of all patrons) (d’Hondt, 2007); and the fact that the Netherlands has one of the world’s lowest rates of problem gambling (Williams, Volberg, & Stevens, 2012). To enter casinos in Austria, patrons have to use an issued ID card which is used to record the frequency and duration of casino attendance. If the person has entered the casino 90 days or more in the previous 180 days and he/she falls into the top 5% of people in terms of attendance frequency, the person is given a letter informing him that his/her gambling habits are deemed to be problematic and about the risks of continued gambling. Moreover, the visitor's ID card is invalidated and he/she is temporarily banned from entering casinos. The more often a visitor is banned, the longer each ban becomes.

After 7 bans the ban becomes permanent (EuroPriSe, 2011). The analysis of Reward/Loyalty Card data or online player activity provides additional opportunities for proactive intervention (Schellinick & Schrans, 2011). Several investigators have been researching behavioural markers of risky play from Reward Cards and online player activity for the purposes of proactively alerting the player or/or implementing some type of automated restriction/intervention (Braverman & Shaffer, 2010; Broda et al., 2008; Gainsbury, 2011; Tsogas, Dragicevic & Kudic, 2011):  The Saskatchewan Gaming Corporation (SGC) contracted with Focal Research to identify behavioural patterns associated with problem gambling in Reward Card data (Davies, 2007; Schellenick & Schrans, 2004; 2006). Beginning in 2006, these behavioural markers (iCare system) were used to allow designated staff to approach and interact with high risk gamblers in the two government-run casinos (Davies, 2007; Saskatchewan Gaming Corporation (SGC), 2006). (Interestingly, the program was portrayed as a voluntary assumption of the operator’s duty of care owed to its patrons). Subsequent research has demonstrated that these interactions have a significant effect on reducing player risk levels (SGC, 2008).

In fiscal 2006/2007 1,436 interactions occurred (SGC, 2007) and in fiscal 2008/2009 there were 26.4 interactions per day, exceeding the target of 26 per day (SGC, 2009).  In 2007, the Ontario Problem Gambling Research Centre launched an innovative initiative in collaboration with the Ontario Lottery & Gaming Corporation. In this system, high frequency (once a week or more over the past year) members of the Reward Card program received a mailed “brief motivational intervention”. The intention was to educate recipients about impaired control, offer a means for self-assessment, argue that early intervention (before debt becomes unmanageable) is preferable to later intervention, and provide a confidential link to a controlled gambling counselling program. Telephone surveys were conducted to test the effectiveness of the letter and the results were, in

70





general, positive, with the most favourable impact among those who identified with the atrisk behaviours described in the letter. In New Zealand, the Gambling Commission has mandated the Auckland SkyCity Casino to use reward card data to track EGM expenditure per session and visitation frequency and to potentially exclude anyone exceeding certain parameters (Collins, 2007). More specifically, someone who visits the casino at least 5 times a week and spends more than $300 on EGM machines per session, or visits at least twice a week and spends at least $500 a session, is deemed to be a potential problem gambler and must be excluded unless the casino considers they can continue to gamble without harm. In 2007 Sweden introduced a system called Playscan. Gambling behaviour on the government-owned online gambling site is monitored and a patron is alerted when he/she engages in behaviour that is statistically associated with the onset of problem gambling (they are told there are displaying either ‘at risk’ or ‘problematic behaviour’ (Svenska Spel, 2007). An internal evaluation of 820 Playscan users is contained on their website. Of those receiving notification of ‘problematic behaviour’, 48% reported that their gambling habits had been influenced by the feedback (12% had started to play less).

Of those receiving notification of ‘at risk behaviour’, 41% reported being influenced by the feedback, with 7% indicating they had started to play less.

Although proactive intervention with ‘at-risk’ gamblers has not received extensive evaluation, this type of intervention is almost always going to be less costly and more effective than treatment and rehabilitation of established gambling problems. The only caveat to this concerns the fact that the above-described interventions tend to be ‘Educational’ in nature, requiring the ‘at risk’ player to recognize his/her problematic gambling behaviour and voluntarily do something about it.

Restricting Access to Money In the United States, house credit is common practice in gambling venues. Cheque-cashing is considered to be a form of house-credit, and is also allowed in some states. Automated teller machines (ATMs) are commonly located in gambling venues, and there are often no limits on ATM withdrawal amounts. In Canada, the granting of house credit is banned in all jurisdictions except in Ontario’s commercial, resort-style casinos.40 ATMs are commonly located in casinos and EGM gambling venues throughout Canada both on and off the gambling floor. Venue-imposed ATM withdrawal limits do not appear to exist, and both debit and credit transactions are generally allowed. Manitoba is unique in banning debit card use for VLT gambling. Manitoba also prohibits cheque-cashing and credit card use in casinos.

40

These casinos also offer ‘front money accounts’, where patrons deposit cash reserves that can be accessed whenever desired by signing chits while gambling.

71

In Europe, house credit is generally banned. ATMs are generally available at gambling venues (not in Belgium), but information regarding withdrawal limit policies was not found in a search of electronically available information. Credit card use at ATMs appears to be allowed. The Republic of South Africa (2004) does not allow credit to be provided in gambling venues, and ATMs are not allowed in any gambling venue under South African law. Credit is banned in all Australian states and territories. ATMs and EFTPOS (Electronic Funds Transfer at Point of Sale) facilities are available at almost all gambling venues (not Tasmania and not Victoria since July 2012), although they are not permitted on the actual gambling floor. ATM withdrawals are limited to $200 a day in South Australia and $400 a day in Victoria. In all certain Australian states and territories (not Western Australia), winnings in excess of certain amounts (between $250 - $2,000 depending on the jurisdiction) are paid by cheque, and certain jurisdictions do not permit the venue itself to cash these cheques (Australian DFHCSIA, 2012). In New Zealand, credit card use to a limit of $200 per day is allowed. There appear to be no limits on ATM withdrawals, and regulations require that ATMs be located away from gambling areas within gambling venues.

In Asia, credit is permitted at Macau casinos but prohibited at Singapore casinos (Godinho, 2006; Ling, 2007). There is a lack of empirical research examining the effectiveness of monetary restrictions. However, anecdotal and existing survey data indicate that restricting ready access to cash is a potentially effective strategy. First, it is well established that problem gamblers access cash machines more frequently than regular gamblers (Allcock, 2002; Caraniche Pty Ltd., 2005; Independent Pricing and Regulatory Tribunal, 2004). Second, problem gamblers in treatment report that the most common reason for terminating a gambling session and leaving a gambling venue is because they have run out of money (Productivity Commission, 1999). Indeed, selfreports of problem gamblers consistently identify easy and immediate access to cash as exacerbating gambling-related harm (e.g., Caraniche Pty Ltd., 2005; McMillen, Marshall, & Murphy, 2004; South Australian Centre for Economic Studies, 2005). The majority of 418 EGM players in Victoria, Australia were of the view that ATMs should not be located in gambling venues at all. Among this same group, this measure was deemed to be the most effective harm minimization strategy available (Caraniche Pty Ltd., 2005).

A total of 72% of people who enrolled in Alberta’s casino self-exclusion program indicated that they felt that restricted access to ATMs would be a ‘very effective’ procedure, with another 10% reporting it would be a ‘somewhat effective’ procedure (Alberta Gaming and Liquor Commission (AGLC), 2007). Implementation of policies to ban credit, limit ATM withdrawals, or remove ATMs from or near gambling venues is often opposed by the gambling industry, due to the potential inconvenience to non-problem gamblers (McMillen, Marshall, & Murphy, 2004). While this may be true to

72

some extent41, it must be said that several problem gambling prevention policies have the same potential. As will be discussed in the Best Practices Section later in this document, the effective prevention of harm associated with potentially dangerous products (e.g., alcohol, tobacco, firearms) has always required some inconvenience to the general public in the interest of benefitting the population as a whole.

Restrictions on Concurrent use of Alcohol and Tobacco Gambling and drinking often co-occur, particularly where gambling occurs at problematic levels (e.g., Crockford & el-Guebaly, 1998; Giacopassi, Stitt, & Vandiver, 1998; Grant, Kushner, & Kim, 2002; Lorains, Cowlishaw & Thomas, 2011; Welte, Barnes, Wieczorek, et al., 2001). The link between increased drinking and increased gambling has also been demonstrated. A study on consumption of alcohol during VLT play found that length of play, rate of double-up betting, and play of losing hands increased during moderate alcohol intoxication, especially for probable pathological gamblers (Ellery, Stewart & Loba, 2005). Kyngdon and Dickerson (1999) found that alcohol consumption prolonged gambling sessions, with the potential for greater financial loss arising from increased risk taking. Other research consistently replicates the finding that alcohol has a disinhibiting effect on gambling restraint and increases risk taking (e.g., Baron & Dickerson, 1999; Cronce & Corbin, 2010; McDonnell-Phillips Pty Ltd, 2006; Phillips, Triggs, Coman, & Ogeil, 2005; Sjoberg, 1969). Given this knowledge, restrictions on the use of alcohol while gambling have significant potential as a harm minimization strategy for problem gambling.

While policies regarding the sale of alcohol in gambling venues vary worldwide, responsible service practices (e.g., prohibiting continued sale of alcohol to intoxicated gamblers) are generally either legislated or otherwise entrenched in government policy. Policies concerning free drinks and other complementary goods and services are less likely to be included in responsible gambling codes. A brief review of alcohol-related policies in various gambling jurisdictions follows: Casinos in Canada may provide free goods and services, but not free alcoholic beverages. Alcohol service is prohibited in some British Columbia casinos, where municipal governments assume responsibility for such licensing decisions. In the United States, free drinks are provided to casino patrons in 13 of 22 states with commercial casinos (Colorado, Florida, Iowa, Louisiana, Michigan, Mississippi, Nevada, New Jersey, New York, Pennsylvania, Rhode Island, South Dakota, West Virginia) (American Gaming Association, 2012). Low-cost drinks are also common. Free and discounted alcoholic beverages are either banned or not commonly available in most European countries, except in some Eastern European countries (e.g., casinos 41

It could be argued that non-problem gamblers tend to predetermine their spending limits and are likely to bring the appropriate amount of cash with them. In this population, ATMs only serve the minority who may have forgotten to bring cash. At-risk patrons and those with impaired control, however, by definition consistently exceed pre-set limits and ‘in the heat of the moment’ will use ATMs to obtain cash they had not planned to spend and may not be able to afford. Inconvenience to these patrons is actually a benefit that reduces additional loss.

73

in the Ukraine). Some casinos in Australia (e.g., Casino Canberra; Crown Casino, Victoria) have historically provided low-cost or free drinks for customers. ‘Host Responsibility’ regulations in New Zealand prohibit free drinks, and include responsible practices for serving alcohol. Free drinks are available at some Macau casinos. Just as gambling and alcohol consumption are related, the association has been established between gambling and tobacco use. Public health campaigns beginning in the early 2000s have successfully led to implementation of ‘public place’ smoking bans in most jurisdictions around the world in order to reduce the well-known health risks associated with smoking and secondhand smoke. In Canada, smoking in casinos is banned in all provinces (with the exception of casinos on some First Nations reserves). In Australia, smoking is generally only permitted in private/premium gambling rooms (Australian DFHCSIA, 2012). As of July 2012, 27 U.S. states have enacted statewide bans on smoking in all enclosed public places. However, 9 of these states exempt casinos (Wikipedia, 2012). U.S. Indian casinos are also exempt. Smoking is allowed in Macau casinos. Smoking is banned in most casinos in Europe.

Smoking bans may inadvertently act as one of the more effective policies to reduce problem gambling, given that the majority of problem gamblers are smokers (e.g., Lorains, Cowlishaw & Thomas, 2011; McGrath & Barrett, 2009; Petry, Stinson, & Grant, 2005; Rodda, Brown, & Philips, 2004). It is no coincidence that gambling venues are the most common places to petition for and receive exemptions from public smoking bans. Indeed, significant 15% - 25% reductions in gambling revenues have followed gambling venue smoking bans in various jurisdictions, including Canada, Australia, New Zealand, Europe, and the United States (Atlantic Lottery Corporation, 2006; Garrett & Pakko, 2010; Hospitality Association of New Zealand, 2005; Pakko, 2008; Thalheimer & Ali, 2008). Reported revenue reductions are notable, considering that a large proportion of gambling revenue traditionally has derived from problem gamblers (Williams & Wood, 2004, 2007b). It is hypothesized that problem gamblers may be less likely to gamble for extended periods if they cannot smoke, thereby introducing a mechanism for reducing harm. Among a group of 418 EGM players in Victoria (49% of whom smoked, with significantly higher rates among problem gamblers), 67% regarded the restriction of smoking in gambling areas to be an effective gambling harm minimization strategy (Caraniche Pty Ltd., 2005).

Forty-nine percent of the smokers reported that they spent less time playing EGMs, and 5% reported they spent more time. These reports were corroborated by a survey of EGM venue operators, who reported that among all the harm minimization measures, the smoking ban was the most effective (Caraniche Pty Ltd., 2005). In New Zealand, a 15.8% drop was seen in the number of government-funded problem gambling service users after the first year of the smoking ban (New Zealand Ministry of Health, 2006). Interestingly, there is also evidence that EGM and casino revenues may return to their previous levels after smoking bans have been in place for some time (e.g., Buchanan, 2006, p. 13, citing Tattersall’s 2005 Annual Report: “Gaming revenue returned to full-year growth following the downturn after the introduction of smoking bans in Victorian gaming venues in 2002”). There is no empirical research to indicate whether this is due to smokers (and problem gamblers) having

74

adjusted to this requirement, or non-smokers patronizing gambling venues at higher rates because of the smoke-free environment.

Restricting Advertising Almost all countries have consumer protection legislation that requires ‘truth in advertising’, which should theoretically curtail attempts to portray gambling as harmless, safe or a good way to make money. Beyond this, many countries have specific legislation that place additional constraints on gambling advertising. For example, most European countries limit the amount of gambling advertising and/or make it mandatory to inform people of the associated dangers (European Commission, 2006; Research and Service Development Centre, 2007). In some jurisdictions the amount of money devoted to promoting gambling participation is many magnitudes greater than the amount of money devoted to problem gambling prevention and treatment. For example, the province of Ontario is reputed to spend more money on prevention, treatment, and research than any other jurisdiction in the world, amounting to $36 million in 2003/2004 (Sadinsky, 2005). However, in this same year, the Ontario Lottery and Gaming Corporation’s budget for marketing, advertising, and promotions was over $570 million. This amount did not include the additional advertising budgets of the three commercial casinos, which were estimated to exceed $200 million.

Consequently, it is the rare Ontario resident who is unaware of the omnipresent lottery and casino advertisements on television, radio, print media, and along public roadways. On the other hand, most Ontario residents are unaware of the province’s ‘responsible gambling’ initiatives (Turner, Wiebe, Falkowski-Ham, et al., 2005). The nature of gambling advertising is also of concern, as some of it is clearly misleading. Typical examples are lottery advertisements that suggest the chances of winning are better than they actually are, or that a person’s overall wellbeing will be substantially better after winning a jackpot (Griffiths, 2005b; Korn, Hurson, and Reynolds, 2005; McMullan & Miller, 2008, 2009a, 2009b). Many lotteries also have websites that provide players with information about the frequencies of winning lottery numbers, which deceptively conveys the impression that useful information might be gleaned from this data. There has also been an increasing usage of phrases such as ‘responsible gaming’, as opposed to ‘responsible gambling’. Indeed, a Google search shows the former to now be the more commonly used term. Similarly, engagement in gambling is typically described as ‘play’ (e.g., ‘playing slot machines’). This is potentially problematic.

Portraying your product as more benign than it actually is undermines efforts of alerting people to their over-involvement in it. An additional concern is that some of this advertising appears to be directed at underage youth (Derevensky, 2008; Friend & Ladd, 2009; Korn, Hurson, & Reynolds, 2005; Monaghan, Derevensky, & Sklar, 2008).

75

In general, the actual impact of advertising on consumer behaviour is complex and not well understood. What is clear is that the impact is dependent on many factors, including the quality and nature of the advertising, the nature of the product, the newness of the product to the marketplace, the nature of consumer, and the familiarity of the consumer with the type of product (Binde, 2007; Kim, 1992). A fair bit of research exists on the impact of alcohol and tobacco advertising, and generally indicates that advertising has some influence on market share, but little effect on overall consumption levels (Boddewyn, 1994; Duffy, 1996; Fisher, 1993; Nelson, 2001; Smart, 1988). Within these findings, however, there are some important subgroup effects. Recovered alcoholics, for example, commonly report that alcohol advertising is a significant trigger for relapse (Treise, Taylor & Wells, 1994). Second, longitudinal research in youth has found a relationship between exposure to tobacco or alcohol advertising and subsequent use of these substances (Anderson et al., 2009; Babor, Caetano, Casswell, et al., 2010; Ellickson, Collins, Hambarsoomians, & McCaffrey, 2005; Lovato, Linn, Stead, & Best, 2003). Furthermore, Weiss et al. (2006) found that anti-tobacco advertising was typically insufficient to counteract the effects of this pro-tobacco advertising.

Relatively little research exists on the effects of gambling advertising on gambling behaviour (Binde, 2007; Griffiths, 2005b). It is reasonable to speculate that gambling advertising may have helped contribute to the positive attitudes toward gambling and increased participation in gambling when it was first widely introduced (Binde, 2007). However, despite continuous high levels of advertising, attitudes toward gambling are gradually becoming more negative (Smith et al., 2011) and the prevalence of gambling and problem gambling is stable or decreasing in most Western jurisdictions (Williams, Volberg, & Stevens, 2012). Several studies have found that people who engage in gambling have a better recall of gambling advertising (e.g., Amey, 2001; Hani, 2008). However, this could simple be due to the fact that people with interest and involvement in any product typically have better attentiveness and recollection of information relevant to it. Several studies have found that advertising has some impact on intention to gamble. Lee, Lemanski, & Jun (2008) found that in a sample of 229 undergraduate students that media exposure to gambling influenced both gambling attitudes and gambling intentions, and that the direction of the effect was dependent on whether the message was positive or negative.

This same study found that anti-gambling media exposure was insufficient to counteract the effects of pro-gambling media exposure. Felsher, Derevensky, & Gupta (2004) found that, among Canadian youth between 10 and 18, of the 90% who reported recalling lottery advertising, that 39% thought it would influence them to buy lottery tickets. In a qualitative study, Derevensky (2008) found that 38% of youth ages 12 to 19 (n = 143) in Quebec reported that gambling advertisements affected both their perceptions of gambling (e.g., gambling can make you rich, winning is easy) and their desire to gamble, with this impact primarily occurring in individuals who already gambled or had an established interest in gambling (i.e., males, older youth, and problem gamblers). Similar findings were obtained in a follow-up quantitative study of 1,147 Quebec youth (Derevensky, Sklar, Gupta, & Messerlian, 2010).

76

The finding that gambling advertisements may have greater impact on propensity to gamble among people who are problem gamblers was also found by Grant & Kim (2001), who reported that 46% of a sample of 131 pathological gamblers indicated that advertising was a trigger (Grant & Kim, 2001). Twenty percent of female gamblers from Ontario with concerns about their gambling reported that ‘exposure to ads on TV, billboards and newpapers’ was ‘very’ or ‘extremely’ important in creating urges and temptations to gamble (Boughton & Brewster, 2002). Similar findings have been reported in a couple of Swedish studies (Binde, 2007, 2009; Jonsson et al., 2003) (although the large majority of problem gamblers in the Jonsson et al. study reported almost no influence of advertising on their behaviour). In a fair summary of the evidence, Binde (2007) concludes that pro-gambling advertising probably does have a small short-term impact on problem gambling, but with a magnitude much smaller than other causal factors. The mechanism of this impact is largely through a) the recruitment of new players for new games in an immature market (e.g., Internet poker), a small percentage of whom will go on to become problem gamblers, b) intensifying gambling behaviour in a portion of regular gamblers, and c) increasing the propensity for relapse in among existing problem gamblers.

Gambling Venue Design When choosing among commercial venues, customers will always preferentially patronize establishments that are more attractive and that offer a greater range of products and services. The same is true for casinos (Friedman, 2000). It is also self-evident that people will tend stay longer in more pleasant surroundings and thereby, potentially spend more (Mayer & Johnson, 2003). Consequently, there is a fair amount of research to identify the types of casino environment that patrons prefer (e.g., Johnson, Mayer & Champaner, 2004; Lam et al., 2011; Lucas, 2003; Mayer, Johnson, Hu, & Chen, 1998). However, beyond these basic tenets of consumer behaviour, the issue is whether there are additional design elements of gambling venues that increase gambling time and/or expenditure. As discussed in the EGM Parameters section, providing comfortable seating, increased privacy, and low machine lighting are features that appear to facilitate EGM play. Specific casino design elements that people have speculated may be related to continued play are the lack of windows, absence of clocks, maze-like interiors, low ambient lighting punctuated by the bright colorful lights of EGMs, and the constant background noise of EGMs, particularly the sounds of winning (there is no sound of losing) (Griffiths, 2009).

Subjectively, many gamblers believe that these design characteristics do promote different patterns of play (e.g., Finlay, Marmurek, Kanetkar, & Londerville, 2007; Griffiths & Parke, 2003; Marmurek, Finlay, Kanetkar & Londerville, 2007), and that alteration of these features would be useful harm minimization strategies (Caraniche Pty Ltd., 2005; Hing, 2003). Some jurisdictions mandate that venues have clocks and/or proper lighting (i.e., certain states in Australia; Australian DFHCSIA, 2012) and that EGMs are inactive when not being played (e.g., Netherlands, Spain; Gourdriaan, de Bruin, & Koeter, 2009; Becona, 2009).

77

However, there is very little research on whether gambling behaviour is actually altered as a function of casino environment, and no research on whether these things facilitate excessive gambling behaviour. The presence of a certain aroma was associated with higher amounts spent on slot machines in a Las Vegas casino compared to a different aroma and no odor at all (Hirsh, 1995). The presence of music and the type of music played had no impact on bet size or amount spent in a laboratory-based roulette game (Dixon, Trigg & Griffiths, 2007). However, a faster music tempo was associated with a faster speed of betting. A similar finding was obtained by Spenwyn, Barrett, & Griffiths (2010) where fast tempo music under red lighting was found to increase speed of bets (but not amount bet) in an online version of roulette. Other research has also found tentative evidence that people gamble more under red lighting (Griffiths & Swift, 1992; Stark, Saunders, & Wookey, 1982). However, even if it were well established that these elements promoted gambling behaviour among current gamblers, a plausible mechanism might be their prior conditioned association to gambling, as lights and sounds are very salient, easily conditionable stimuli (i.e., a stronger case is made if these effects can be demonstrated in novice gamblers).

The final observation relevant to the issue of the impact of gambling venue design is that EGMs have no difficulty generating significant revenues across a myriad of environments, including convenience stores, bars, clubs, hotels, arcades, restaurants, racetracks, and boats. Thus, while certain environmental features of gambling venues may facilitate time and money spent gambling, it seems unlikely they are powerful factors.

Increasing the Cost of Gambling Increasing the cost of alcohol and tobacco (through taxation) is one of the most effective policies to reduce alcohol and tobacco use and abuse (e.g., Babor, Caetano, Casswell, et al., 2003; Cnossen, 2005; Cook, 2007; Wagenaar, Salois, & Komro, 2009). However, it is unclear whether increasing the cost of legal gambling would effectively deter problem gambling (Clotfelter, 2005). Substantial economic ‘costs’ are already built into gambling products. Moreover, population participation in gambling is inversely related to the payback rate: the highest patronage is for lotteries and instant win tickets, which offer the lowest payback rate, and there is comparatively low patronage of casino table games and sports betting which have higher payback rates. Variation of payback rates within a particular gambling format (e.g., roulette, sports betting), does sometimes influence spending on that format (i.e., lower spending on variants providing lower payback rates) (Harvey, Swayze, Walls, 2004; Paton, Siegel, Vaughan-Williams, 2004). That said, a general review of this literature by Parke (2010) found very little empirical evidence that payback percentage was an important feature of gambling behaviour or related to the development of problem gambling.

However, it is important to recognize that when consumers purchase a product their perception of cost has more to do with the initial ‘sticker price’ relative to the perceived value of the product. Applied to gambling, this explains why there is a preference for cheaper lottery

78

tickets (e.g., Turner & Ferentzy, 2010), lower denomination EGMs (Schwartz, 2010), and larger jackpots (e.g., DeBoer, 1990; Forrest, Perez, & Baker, 2010; Garrett & Sobel, 2004; Turner & Ferentzy, 2010). Hence, the minimum bet size relative to the maximum win size will generally be more important to the gambler than the overall payback percentage or actual odds of winning. (People who are knowledgeable and concerned about payback percentage and odds tend to avoid commercial gambling in the first place). Accordingly, raising the minimum bet size and lowering the maximum win size are likely to have the greatest potential for harm minimization.

Government Provision of Gambling There is wide divergence concerning how gambling is provided and who provides it. There are four basic approaches:  A free market approach that allows private commercial operators to provide gambling services in a manner similar to other commercial products (although sometimes taxed at a much higher rate).  Restricting the provision of gambling to one or two government-owned or governmentcontrolled providers.  Restricting the provision of gambling to private monopolies.  Having a mixture of government controlled monopolies and private commercial offerings. The free market approach tends to be favoured in jurisdictions such as the United States and Australia, whereas the remaining approaches are more common in places such as Canada and Europe. There is also significant variability as a function of game type. In most countries it is common for lotteries and lottery products to be delivered by government-controlled monopolies, whereas government involvement is rare in the delivery of horse racing or bingo. Casino delivery is quite variable. In the United States they are provided by private operators, whereas in several European countries only a public operator or and operator closely supervised by the State can manage a casino (Germany, Greece, Finland, Hungary, the Netherlands, Sweden, and Slovenia).

Essentially, the main issues are whether gambling a) should be provided by the government or private commercial provider(s), and b) whether it should be provided by a single entity (monopoly) or multiple providers. An extensive literature exists on the benefits and drawbacks of each of these approaches in the delivery of various consumer products. It is beyond the scope of this paper to provide a review of this more generic literature. However, it is worthwhile to at least identify the main issues as they relate to problem gambling. A conflict of interest exists when government adopts the role of both the regulator and provider of gambling or when the government is a major financial beneficiary of gambling. This conflict of interest potentially compromises the regulator’s ability to implement truly effective prevention policies, and to effectively regulate the operator. Effective prevention and

79

treatment will typically negatively impact revenues, introducing a policy conflict between the protection of public health and the maximization of gambling revenues (Adams, Raeburn, & de Silva, 2009; Orford, 2009). The actual effects of this conflict of interest are difficult to determine, as this situation tends to be confounded with other things. Many of the jurisdictions where government is the provider of gambling and/or receives a significant portion of the revenue (e.g., Canada, several European countries) are involved in gambling ostensibly to provide a safer and more controlled product to the public. With this greater concern for public welfare, these governments also tend to offer more in the way of problem gambling prevention and treatment initiatives. That being said, a) the creation of these initiatives is partly spurred on by their sensitivity to this conflict of interest criticism, and b) the initiatives that most of these governments have chosen to implement have tended to be the least effective ones42. In contrast, jurisdictions where the government is primarily involved in the regulation rather than provision of gambling (e.g. United States, Australia), tend to have less in the way of protective measures, as these governments put more responsibility on the individual to govern their own behaviour.

The power of the gambling lobby/industry is also much stronger in these types of countries and often deters the adoption of protective measures (Grinols, 2004; Kindt, 1998). There are also some important lessons from the alcohol field, where the evidence suggests that monopolistic and/or government involvement in alcohol provision is associated with less harm to the public (e.g., Miller, Snowden, Birckmayer, & Hendrie, 2006; Popova et al., 2011; Wagenaar & Holder, 1996). However, these data are largely correlational, and therefore subjects to similar confounds as mentioned above. A stronger methodology involves examining the effects of privatizing alcohol sales in jurisdictions where government monopolies previously existed. Privatization has occurred in several U.S. states (Idaho, Iowa, Maine, Virginia, Washington, West Virginia), 3 Canadian provinces (Alberta, British Columbia, Quebec), and certain countries (New Zealand), although in some cases it just involved elimination of certain types of alcohol provision (e.g., retail wine monopoly), and in some cases there were still restrictions on private retail (e.g., no provision in grocery stores).

Nonetheless, research has generally found that privatization is associated with an increase in the number of retail outlets, longer opening hours, and increases in overall alcohol consumption (with overall consumption level having a statistical relationship to overall level of harm) (Holder, Agardh, Hogberg, et al., 2008; Stockwell et al., 2009; Wagenaar & Holder, 1995; 1996; cf. Trolldal, 2005). Hence, although theoretically sound, in practice, eliminating or constraining government involvement in the provision of gambling may not achieve the desired effect. In lieu of this, another solution is to enact legislation that targets this conflict of interest or that requires gambling providers (government or otherwise) to effectively mitigate the harm from the provision of gambling. While it is true that many jurisdictions have legislation requiring gambling providers to provide gambling responsibly, only a few have enacted specific 42

Most of these initiatives have focused on the fairly weak strategy of better education of consumers as opposed to more effective policy initiatives that constrain the availability of gambling and how it is provided.

80

legislation to ensure this result. A notable exception is Germany, which has enacted legislation that among other things a) prohibits the pay of gambling provider executives to be tied to gambling revenue; b) requires that the monitoring of compliance with gambling regulations to be done by authorities not connected to the fiscal interests of the state; c) requires all new gambling products to be reviewed by an advisory board of gambling addiction experts prior to introduction; and d) requires gambling providers to detect and exclude problem gamblers from gambling venues (Meyer & Hayer, 2009).

81

SUMMARY OF THE EFFECTIVENESS OF EDUCATIONAL AND POLICY INITIATIVES The following general themes arise from the evidence that has been reviewed to this point. 1. First, a very large number of different prevention initiatives exist. This reflects the many different jurisdictional approaches to problem gambling, the considerable interest and effort that is being put into mitigating gambling’s harm, and the large number of ways in which gambling behaviour can be potentially influenced. 2. Some jurisdictions have little or no prevention strategies, whereas others have an extensive array. There is tendency for more socially progressive jurisdictions to also have more progressive and effective prevention strategies (which makes inter-jurisdictional comparisons of the effectiveness of different policies difficult). 3. There are a few initiatives where almost no direct evidence exists concerning their efficacy, many initiatives where some direct evidence exists, and a small number of initiatives where extensive evidence exists. That being said, most of these initiatives have analogues in allied health fields that have been thoroughly researched. 4. While certain initiatives are more effective than others, there is almost nothing that is not helpful to some extent and, conversely, there is almost nothing that by itself has substantial potential to prevent harm. 5.

The most commonly adopted prevention measures tend to be among the least effective ones (e.g., awareness/information campaigns, responsible gambling features on EGMs, casino self-exclusion, etc.). Furthermore, when potentially more effective initiatives are implemented, they are typically done in such an inconsequential or perfunctory fashion as to virtually ensure lack of impact (e.g., small reductions in number of gambling venues or numbers of EGMs, minor restrictions on access to money, etc.). Table 6 summarizes the evidence concerning the demonstrated or estimated effectiveness of various educational and policy initiatives to prevent problem gambling.

82

    ? 

POLICY INITIATIVES Restrictions on the General Availability of Gambling Restricting the Number of Gambling Venues (casinos/racinos) Restricting More Harmful Types of Gambling Limiting the Number of Gambling Formats Restricting Gambling to Dedicated Gambling Venues Restricting the Location of Gambling Venues Limiting Gambling Venue Hours of Operation Restrictions on Who can Gamble Prohibition of Youth Gambling Increasing the Legal Age for Gambling Restricting Venue Entry to Non-Residents Restricting Venue Entry to Higher Socioeconomic Classes Casino Self-Exclusion Restrictions or Alterations on How Gambling is Provided Modifying EGM Parameters Player Pre-Commitment Eliminating Reward/Loyalty Cards or Changing Their Parameters Operator Imposed Maximum Loss Limits Problem Gambling Training for Employees of Gambling Venues Automated or Mandated Intervention for At-Risk Gamblers Restricting Access to Money Restricting Concurrent use of Alcohol and Tobacco Restricting Advertising Gambling Venue Design Increasing the Cost of Gambling Government Provision of Gambling

low



EDUCATIONAL INITIATIVES Childhood Interventions Information/Awareness Campaigns On-Site Information/Counselling Centres (RGIC) Statistical Instruction School-Based Prevention Programs

moderately low

moderate

moderately high

high

Table 6. Estimated Effectiveness of Problem Gambling Prevention Initiatives.

1

 1 1 ? ?  ?2 ? ?3  ?4 ? 5  6 7 ? ? 8  ?  9 ? 10

? 

83

1. If the reductions are substantial. 2. Unless the time reduction is substantial. 3. A more effective strategy might be to model responsible gambling to youth prior to the

legal gambling age. 4. Prevention benefits are limited to residents rather than non-residents. 5. Even with effective detection mechanisms, this initiative has limited preventive value because of its tertiary nature. 6. Decreasing maximum bet and win size, slower speed, reducing frequency of near misses, reducing number of betting lines, eliminating bill acceptors, reducing their interactive nature, presenting pop-up messages, and absence of seating. 7. Only if mandatory. Also more effective if irrevocable, applicable to all EGMs or all gambling in a jurisdiction, with non-swappable ID. 8. The value of this measure may be increased if staff intervention is mandatory and compliance is enforced. 9. More important for preventing relapse in problem gamblers and preventing involvement in gambling among young people. 10. Greater harm minimization potential with raising minimum bet size and lowering maximum prize size. Note. Question mark (?) indicates uncertainty due to insufficient evidence.

84

BEST PRACTICES TO PREVENT PROBLEM GAMBLING The foregoing summary indicates some level of positive empirical support for a number of prevention initiatives. The purpose of the present section is to provide more direct guidance about the most effective strategies that would constitute ‘best practices’ for the prevention of problem gambling. Some might contend that it is premature to identify best practices without a stronger base of direct evidence. We respond that, it is not only possible to speculate on best practices, it is necessary that this speculation occur. As has been seen, dozens of prevention initiatives have already been implemented in hundreds of jurisdictions around the world. People are not waiting for ‘all the evidence to be in’. Unfortunately, history tells us that once a prevention program (or treatment program) becomes established, it often becomes very difficult to supersede it, even when more effective alternatives become available. This is the situation found in the substance abuse field, where the most commonly used (and entrenched) educational initiatives for prevention and treatment tend to be the less effective ones (Babor, Caetano, Casswell, et al, 2010; Miller, Wilbourne & Hettema, 2003; Tobler et al., 2000).

BEST PRACTICE 1: Strive for Optimal Design and Evaluation of New Initiatives. 1.1 Employ Social Scientists in the Design, Implementation, and Evaluation. The inspiration, design, and implementation of almost all existing problem gambling prevention initiatives have primarily been led by people from government and the gambling industry without backgrounds in either prevention or the social sciences. When social scientists have been utilized, they are usually confined just to the evaluation of these initiatives. Undoubtedly, this is an important contributing factor to why many of the existing problem gambling prevention initiatives have tended to be fairly ineffective. 1.2 Use Proven Theoretical Models of Behavioural Change to Guide the Development of New Initiatives. It is difficult to change people’s behaviour. However, there are some well established theories that are able to predict when an initiative is likely to be successful in causing people to take action (Noar & Zimmerman, 2005). These include the Theory of Planned Behaviour (Ajzen & Madden, 1986); the Transtheoretical Model (Prochaska, DiClemente, & Norcross, 1992); and the Health Beliefs Model (Becker, 1974; Janz et al., 2002).

The Health Beliefs Model, in particular, is very clear about the circumstances that cause people to change their behaviour: when the information being communicated is personally relevant; when the consequences of not changing behaviour are potentially severe; when the benefits of adopting the behaviour are quite beneficial; when the person perceives that the behavioural change required is relatively easy for them to achieve; and when the person receives plenty of cues for action.

85 1.3 Always use Behavioural Change as the Primary Measure of Effectiveness. The best measure of actual effectiveness will always be empirical evidence that the initiative reduces the number of new cases of problem gambling in the population (i.e., incidence). Admittedly, this type of evidence is difficult to obtain as it requires follow up of large group of individuals over an extended period of time. In lieu of measuring incidence, there needs to be focus on meaningful behavioural change within the individual as the measure of effectiveness. Improvements in awareness, knowledge or attitudes are of value as intermediate steps in the right direction, but of very limited importance if not accompanied by behavioural change. Similarly, the perceptions or opinions of at-risk gamblers toward a particular prevention policy is useful information, but is never a substitute for actual changes in behaviour. ‘Knowing something’ and ‘doing something about it’ are always two different things, particularly in the field of addictions. 1.4. Conduct Regular Population Assessments of Problem Gambling to Monitor the Impact of New Initiatives. Regular evaluations of overall population prevalence of problem gambling is an important and efficient way to gauge the impact of the array of different initiatives that are typically implemented in the same time period.

Annual jurisdiction-wide treatment numbers is another useful statistic.

86

BEST PRACTICE 2: Recognize that Effective Problem Gambling Prevention requires Decreased Revenue and some Inconvenience to Non-Problem Gamblers. As indicated earlier, the most commonly implemented prevention initiatives have also tended to be the least effective ones. Furthermore, when potentially more effective measures are implemented (e.g., reducing number of EGMs, reduced venue hours, etc.), the reductions have been too minor to have a significant impact. What this reflects is the fact that policy makers have an unrealistic desire to implement effective prevention policies that do not inconvenience non-problem gamblers or reduce revenues. Unfortunately, the reality is that the effective prevention of problem gambling prevention is only likely to occur with some level of inconvenience to non-problem gamblers and necessarily involves a loss of revenue because of the significant contribution problem gamblers make to overall gambling revenue (Williams & Wood, 2004, 2007a). The effective prevention of harm associated with potentially dangerous products or activities has always required some inconvenience to the general public and loss of revenue. The general public accepts the fact that aggressive government policies restricting tobacco advertising and consumption have resulted in reduced tax revenue.