« Previous | Contents | Next »
Listen
CHAPTER FOUR: FACTORS INFLUENCING THE PREVALENCE AND INCIDENCE OF PROBLEM GAMBLING
4.1 Various factors have been found to influence the distribution of gambling problems, and these have been examined through studies of the incidence and prevalence of problem gambling rates. Prevalence refers to the proportion of people within the population suffering from a disorder at one point in time, while incidence refers to the proportion of people who acquire the disorder during a specific period. Prevalence is usually assessed using cross-sectional methods, such as surveys.
4.2 Assessing incidence requires longitudinal approaches involving panel or cohort studies and much longer time frames (Abbott & Volberg 1999). Studies of incidence can be useful in helping to determine the impacts of particular forms of gambling on groups who are newly exposed to it. A number of such studies have been carried out in North America, Australia and New Zealand, and indicate two main factors that contribute to the development of problem gambling, namely the structural characteristics of games themselves, and the availability of gambling.
Type of Game and Availability
4.3 Certain features of games are strongly associated with problem gambling. These include games that have a high event frequency (i.e. that are fast and allow for continual staking), that involve an element of skill or perceived skill, and that create 'near misses' (i.e. the illusion of having almost won). Size of jackpot and stakes, probability of winning (or perceived probability of winning), and the possibility of using credit to play are also associated with higher levels of problematic play. Games that meet these criteria are electronic gaming machines and casino table games.
4.4 The British Gambling Prevalence Survey found that casino games were associated with the highest rates of problem gambling in their sample, with a rate of 8.7% amongst those who had played in the past year. If activity was broken down by weekly activity, the figures were even higher, at 34.5%. However, only around 3% of the population actually participate in casino gambling, and the researchers emphasised that the relatively small numbers involved meant that this latter figure should be interpreted with caution. As a point of comparison, the next highest rates of problem gambling were found on betting with a bookmaker (8.1%), followed by betting on dog races (7.2%). Gambling machines were associated with a problem prevalence rate of 3.4%, although when playing on weekly basis was considered (which the researchers considered more accurate), the rate was 6.3%.
4.5 An earlier British study also found particularly high rates of problem gambling associated with casinos. The survey, the largest of its kind, interviewed over a thousand patrons of forty casinos and classified 2.2% of respondents as 'severe' problem gamblers and 5.2% as problem gamblers (Fisher 1996, 2000). It has been pointed out that these rates are likely to be an underestimate, as the researcher adjusted the figures downwards to take into account the frequency which respondents reported they visited casinos 2. Not taking this into account, the figures would be 7.7% for severe problem gamblers and 8.6% for problem gamblers.
4.6 In addition, the research demonstrated that visitors to casinos were made up of a small core of regulars (i.e. who visited at least once a week), who had higher rates of gambling problems, and a much larger group of less frequent visitors. Although the regulars made up only 7% of casino-goers, they accounted for 63% of the visits, and had high rates of problem gambling: 7% were classed as 'severe' problem gamblers, and 7% as 'problem gamblers', compared with 2% and 5% of the less regular players.
4.7 The study also found that regular visitors constituted a slightly different demographic group from visitors in general. Males, those aged over 40, people retired from work, those separated from their partners and people of non-white ethnicity were all more likely to be regular visitors. A highly significant finding was that problem players were three times more likely than non-problem players to be of Chinese ethnicity. Severe problem gamblers were also twice as likely to smoke as others, and twice as many had worried about their levels of alcohol consumption.
4.8 The author concluded that casino players appeared to be made up of a small core of regular players, who were demographically distinct from the population as a whole, and who possessed particularly high rates of problem gambling. However, problem gamblers who were not regular visitors to casinos tended to be younger (30 years of age or younger), single and unemployed. This suggests that there are two distinct groups who experience problems with casino gambling. One is young, single unemployed males whose primary form or forms of gambling are outwith the casino but who, as it were, 'bring their problems in' with them when they visit. The other is an older group of retired males who are also single and whose problematic behaviour may be predominantly confined to casinos. This latter group are also much more likely than non-problem players to be of Chinese ethnicity.
4.9 Britain is slightly unusual in its distribution of problem gambling, since almost every other country worldwide reports that the highest rates are associated with electronic gambling machines ( EGMs), where prevalence can be as high as 25% (Abbott & Volberg 2000; PC 1999). These rates are particularly associated with machines that are located outwith casinos, i.e. dispersed in convenient locations such as arcades, bars, and clubs and so on. This 'convenience' gambling (or 'ambient' gambling as it was named by the Budd Committee), is the kind that is most associated with impulsive and habitual play, and is of greatest concern to policy-makers and treatment providers alike.
4.10 This is most apparent in Australia, where EGMs or 'pokies' were introduced in large numbers in New South Wales and Victoria in the 1990s, and are now widely dispersed in a range of locations throughout communities. This form of gambling is by far the most problematic in that country. They are particularly concentrated in low income areas, and easily accessible through their wide dispersal in local neighbourhood venues. The distribution of EGMs is largely responsible for the different rates of problem gambling throughout Australia.
4.11 For example, in Western Australia, EGMs are banned outwith the state's one regional casino, and as a consequence, the ratio of machines per 1000 adults is 0.9. In contrast, in Victoria, large numbers of EGMs are widely dispersed, and the ratio of machines to adults is 8.1. These differences are reflected in rates of problem gambling: Western Australia has one of the lowest in the country, at 0.7%, while Victoria one of the highest, at 2.14%. In Victoria, nearly 90% of problem gamblers receiving counselling state EGMs as their main problem; in Western Australia, the number is 18%. New South Wales, which has the greatest density of machines in the country overall, also has the highest rate of problem gambling: 2.55% 3 ( SA Centre for Economic Studies 2005). In other words, it is not only the type of game, but also its availability that combine to create problems in the Australian case.
4.12 It has been suggested that these types of electronic machines tend to displace other forms of gambling in terms not only of overall expenditure, but also of share of problem playing. As one researcher put it, the great majority of problem gamblers who appear in both community surveys and treatment settings worldwide report having problems exclusively or primarily with gaming machines (Abbott, 2001). EGMs are the fastest growing and most profitable sector of gambling markets: between 70 and 80% of revenues come from them (Abbott 1999).
4.13 In addition, they have wide appeal. They are popular with young people, already familiar with computer video games; with women, who can be uncomfortable with more traditional casino table games; and with a range of more traditional bettors, who may use them to fill in time between games at, for example racetracks or in betting shops. Internationally, it has been found that where such machines proliferate, the distinctions between male and female rates of problem gambling have tended to decline, in a 'feminisation' of problematic behaviour.
4.14 The rates of individuals presenting for treatment can also be taken as an indication of the association between particular forms of gambling and problem behaviour. The great majority of individuals calling gambling telephone helplines and presenting for treatment report that their problems are primarily with electronic gaming machines, track betting and casino table games. In the last ten years, increased expenditure on non-casino gaming machines has been accompanied by increases in the proportion of people contacting helplines who cited this as their major form of gambling throughout Australia, New Zealand and a number of other countries (Paton-Simpson et al. 2004; PC 1999) 4.
4.15 Similarly in the U.K., reports of problem gambling are most often associated with electronic machines. Over one third of calls to the national gambling helpline, GamCare, and over half of counselling sessions, are predominantly related to them (GamCare 2003). In the past two years, problems associated with a new form of betting machine - Fixed Odds Betting Terminals ( FOBTs) - a kind of remote 'electronic roulette' - have increased dramatically. A recent, unpublished survey by Mintel for the Association of British Bookmakers has revealed dramatic increases in the use of FOBTs by problem gamblers (Europe Economics-Mintel 2006). The research examined patterns of gambling participation amongst betting shop visitors, and found that, between 2004 - 2005, the use of FOBTs among problem gamblers had escalated from 12% to 40%. Given that new casinos in Britain, and especially Regional developments, will have relatively large numbers of EGMs, including FOBTs, this is a trend which should be watched closely.
4.16 The Australian case provides clear evidence of the link between the availability of certain types of gambling, and problem gambling. This relationship will be explored further in the remainder of this section, when surveys of the impacts of the availability of various forms of gambling on problematic behaviour are examined.
4.17 However, it should be pointed out here that availability is not simply a matter of the number of gambling outlets in a particular area. A range of other factors are involved. Some of these include ease of access, such as opening hours, restrictions on entry, public transport and/or private parking facilities, and general location. Such features can make a visit to a venue relatively time consuming and something which must be planned - such as a trip to an out of town casino - or they may make it very easy, such as dropping coins in a machine in a bar.
4.18 Among the major forms of gambling, gambling machines and lotteries are most accessible, in that they are widely dispersed in a range of locations which people may pass by in the course of their everyday activities. Casinos are least accessible, in that many are situated in areas which individuals must travel to using some form of transport, whether in out of town locations (as is common internationally) or in urban centres (as is most often the case in the UK).
4.19 The more time and effort is required to enable an individual to gamble - e.g. in terms of planning and actually getting to a venue - the less likely they are to gamble on impulse. Conversely, the less time and effort is required, the more likely they are to gamble impulsively - and to continue to do so. Convenience is therefore a crucial regulatory issue (particularly when combined with the absence of public awareness activities), and this has implications for the location of casinos.
Availability
4.20 Given the association between high levels of involvement with particular forms of gambling and problem gambling it could be expected that the substantial increases in gambling availability and expenditure that occurred in many parts of the world since the mid-1980s will have generated significant increases in problem gambling prevalence. This has been examined through surveys of the incidence and prevalence of problem gambling in the general population, most of which have been conducted in the United States, Canada and New Zealand.
4.21 Many studies show a positive relationship between increased availability of gambling - especially casino and electronic machine gambling - and problem behaviour. However, closer analysis of some of these results shows that the relationship is not straightforward, nor is it linear. Correlation does not always establish causation, and more sophisticated research is needed in order to tease out relationships and allow stronger causal inferences to be made.
4.22 Studies that have examined the incidence and prevalence of gambling problems, particularly in respect to casino development, are considered next, from paragraph 4.23. Where the material is available, the content of some of the studies considered in this section is covered in some detail. Although it could be presented in more 'summary' fashion, it is felt that the complexity of the issues involved, and the controversy that exists over some of the fine points of interpretation, merit such attention. However, a summary paragraph is also provided at 4.96 for readers who wish to skip this level of analysis.
4.23 A meta-analysis of a number of North American state and provincial surveys of gambling prevalence that had been carried out between 1975 and 1996, found a significant increase in the prevalence of problem gambling over time in the general population. While past year prevalence rates for surveys conducted prior to 1993 averaged 0.8%, rates for post 1993 surveys averaged 1.3%. During this period gambling availability had increased substantially (Shaffer et al 1997) Although statistically significant, the authors noted that the total variability in prevalence explained by time was relatively modest and that it has yet to be determined what other factors explain changing rates over time (Shaffer et al 1997, 1999).
4.24 Similar trends have been reported in other research. For example, various studies have found that Nevada has problem gambling rates between two and three times the national average (Volberg, 2002). More recently, Shaffer et al (2004) examined prevalence rates from counties within Nevada and found that the four counties with the greatest access to casinos had the highest problem gambling rates and the four with the least availability had the lowest rates.
4.25 A study of prevalence rates among five U.S.. states with varying gambling histories and gambling accessibility showed that those with a longer history of legally available gambling had higher rates (Volberg 1994). In states where legal gambling had been available less than ten years, less than 0.5% of the adult population were classified as pathological gamblers; in states where it had been available for over twenty years however, the figure was 1.5%.
4.26 A number of studies have found correlations between gambling availability and branches of Gamblers Anonymous. Lester (1994) found significant relationship between the numbers of legally available forms of gambling and branches of GA, from states such as New York and Nevada, which have many forms of gambling and branches of GA, to Arizona and Utah, which have few of either. In Australia, the Productivity Commission found ten times the number of GA branches in New South Wales, which has high levels of gambling expenditure, than in Western Australia, which has low ones.
4.27 Although this correlation between service provision and gambling availability is likely to be an indicator of problem gambling prevalence, it should also be noted that it is also a policy intervention to treat problem gambling, as well as an indicator of increased awareness of problem gambling in the population.
4.28 Some studies have not found these relationships between availability and increased problem gambling. For example, Canadian studies from Alberta, Manitoba, New Brunswick and Nova Scotia have not reported increased rates of problem gambling, despite increases in gambling availability (Alberta Alcohol and Drug Abuse Commission 1998; Criterion Research 1995, Baseline Research 1996; 1996a). Studies from South Dakota and Minnesota also found no significant changes (Volberg and Stufen 1994; Emerson et al 1994).
4.29 In general, despite variability in findings, and a general paucity of research, the National Research Council concluded overall that: "the changes observed in [these] studies was consistent with the view that increased opportunity to gamble results in more pathological and problem gambling" (1999: 84). Most researchers worldwide concur with this opinion.
Casino Availability
4.30 This section will review studies of the availability of casinos on rates of problem gambling. First however, it should be noted that casinos are not homogenous, but are rather distinguished by a range of factors, including, crucially size and geographical location, which can result in very different impacts.
4.31 To begin with, casinos can vary greatly in size, from small enterprises with limited facilities, to enormous destination casinos, offering an entire leisure experience. In addition, they are built in a range of locations, and this 'locational structure' is important in determining overall social impacts. Bill Eadington, Professor of Economics and Director of the Institute for the Study of Gambling and Commercial Gaming at the University of Nevada, argues that casinos can be divided into the following categories:
- Resort or destination casinos: located away from population centres (such as the Gold Coast, Queensland, and Biloxi, Mississippi)
- Rural casinos: located away from population centres (such as Foxwoods in Connecticut; Sun City in South Africa; and many tribal casinos in the U.S.)
- Urban or suburban casinos: located in or near major metropolitan areas (such as those in Detroit; in New Orleans; most U.K. casinos and the Crown Casino in Melbourne) (Eadington 2003).
4.32 Another category can be added to this list which could be described as 'city casinos': urban centres whose primary industry is casino gambling, such as Las Vegas and Atlantic City.
Apart from their locational structure and size, casinos are also characterised by residents with quite different demographic profiles, from affluent white players, to those from impoverished ethnic minority communities.
4.33 The new British category of regional casinos does not have a direct equivalent internationally. In terms of facilities offered, these venues are closest to resort or destination casinos. However, in terms of location, unlike many casinos in the U.S., Australia and elsewhere, regional casinos in Britain are likely to be situated relatively close to major population centres.
4.34 The social costs of casinos tend to be exported to the areas where the gamblers who play them live, so that casinos that draw on large non-local markets tend to have low social costs relative to economic benefits, and those that rely on locals as their primary customers have high ones. Rural and resort casinos are therefore associated with the lowest social costs, since customers for those generally tend to come from beyond the casino's own jurisdiction. However, urban or suburban casinos tend to draw large numbers of local residents to them, and so social costs remain within the community where the gambling facilities themselves are located.
4.35 Eadington notes, however, that there is also an extent to which these casinos may simply retain locals who would have travelled outwith the region to gamble anyway. In this scenario, social costs are no higher, since those individuals would have still 'brought their problems back' to the local community, as it were, as well as leaving the profits in a neighbouring jurisdiction (Eadington 1993).
4.36 All of this means that casinos can have very different types of impacts, depending on size, location and local market. Unfortunately, few studies take these factors into account either when reporting their results, or when attempting comparisons with other research findings. The following discussion is therefore of the impacts of casinos on rates of problem gambling in a general sense, although details of the relevant factors as listed above are included when available.
4.37 In this section, the most noteworthy finding is that studies of casino proximity and rates of problem gambling have found positive, although often non-linear, relationships.
Examples of research
4.38 One of the most striking demonstrations of this came from the survey conducted by the National Opinion Research Centre ( NORC) in the U.S. for the NGISC (Gerstein et al 1999). NORC analysed social and economic changes in 100 communities that lay within a 50 mile radius of one or more casinos between 1980-1997. They found that the location of a casino within 50 miles (versus 50 to 250 miles) of an individual's home was associated with approximately double the prevalence of pathological gambling (2.1% compared to 0.9%).
4.39 Similar findings were reported in research by Welte et al (2004). This study analysed census data and geographic information, and reported that the location of a casino within ten miles of an individual's home was independently associated with a 90% increase in the odds of being a problem or pathological gambler. In addition, individuals living within disadvantaged neighbourhoods were found to have higher rates of gambling problems.
4.40 A number of prevalence surveys from North America have also suggested that the prevalence of problem gambling increases as a result of the introduction of casino-style gambling. In Minnesota and Iowa, for example, the prevalence of problematic behaviour increased significantly over a four year and six year period respectively (Emerson and Laundegran 1996; Volberg 1995). In the case of Minnesota, the increase followed a dramatic expansion in the availability of lotteries, high stakes bingo and more than twenty casinos in the state. In Iowa, the increase followed the introduction of riverboat casinos, Native American casinos and slot machines at racetracks.
4.41 In Indiana, a recent report examined the proximity of casinos to rates of voluntary exclusion from casinos. Voluntary Exclusion Programs ( VEPs) exist to allow individuals to voluntarily have themselves banned from casinos, for one year, five years or for life. Analysis of trends in VEP rates showed a strong relationship between proximity to a casino and higher enrolment rates, with rates rising as distance from a casino decreased. In addition, the researches also found that calls to a problem gambling helpline that mentioned casino gambling also increased with proximity to a casino. Although the study cautioned that such rates should not be taken as indicators of problem gambling prevalence, they noted that the use of such programmes at least demonstrate that an individual feels that they have a problem with casino gambling (Policy Analytics 2006).
4.42 Unfortunately, there are relatively few studies of the impacts of casinos in urban areas. Detroit, for example, opened three in 1999. Detroit has a large ethnic minority population, a downtown area characterised by high unemployment and crime, and has been experiencing an economic downturn since much of its manufacturing base left in the 1970s. Although opposition to the casino - especially from the media - exists, with constant reports of the bankruptcies and suicide related to it, little is known about its actual social impacts.
4.43 However, the number of voluntary exclusions (or 'dissociated persons' as they are known) has grown from 56 in 2001 to 508 in 2003, indicating a significant increase in individuals who felt they had problems in controlling their gambling (Associated Press 2003). A small scale study of problem gambling in the city noted that the casinos' main market was mostly local, and estimated that over 5% of those living in the city and around 3% of the suburban population, had gambling problems (Wong et al 2000).
4.44 America's other large urban casinos are in and around New Orleans. Between 1993 and 2000, more than a dozen riverboat casinos were introduced throughout the state, with a large land-based enterprise opening in New Orleans in 2000. Not only did this casino fail to regenerate the surrounding it area, it was forced to twice declare bankruptcy. Again little research exists on its social impacts.
4.45 However the Louisiana Office for Addictive Disorders carried out surveys of the region over a seven year period, beginning in 1995, two years after the first casinos opened. Although it is not possible to estimate the impact of the introduction of casinos on rates of problem gambling, and although the surveys had small sample sizes, the rate of problem gambling for the state as a whole appears high, at 3%.
4.46 Three surveys found this rate to be 'remarkably stable' over time, although it fell dramatically in 1998. Significantly, this was also around the time the casino declared bankruptcy. In the most recent survey, over three quarters of calls to the state's gambling helpline were primarily related to casino gambling, with the remainder related to machines. The most regular gamblers were the unemployed, although it is not known what percentage were also problem gamblers, or how much they spent playing (Vogel and Ardoion 2002).
4.47 In Australia, the impacts of casinos in cities like Melbourne, Sydney and Perth have received surprisingly little rigorous analysis by researchers. This may in part be due to that country's focus on its largest form of problem gambling - EGMs - towards which the majority of research is directed. However, it has been noted that the opening of the Burswood casino, a large-scale 'resort' casino in Perth (which houses the state's only EGMs) in 1985, was associated with a dramatic increase in expenditure which subsequently declined and levelled out in the 1990s.
4.48 Similarly, the opening of the Crown Casino in the riverside area of Melbourne in 1994 was also associated with a dramatic increase in expenditure, which has subsequently declined ( SA Centre for Economic Studies 2005). A survey of gambling activity in Melbourne for the Casino and Gaming Authority reported a steady decline in casino participation since the casino opened, from a peak of 25% in 1997, to 18% in 1998, and 16% in 1999. However, the study noted that those who do participate tend to play more often and spend more when they do. Unfortunately, it did not provide reliable estimates of problem gambling trends (Roy Morgan Research 2000).
4.49 In Canada, studies of the impact of casino openings have reported mixed results. For example, Room et al (1999) studied social impacts of the introduction of a casino in Niagara Falls, Canada, in 1996. Niagara Falls was in the position of being a small sized city (population 75,000) which had experienced an economic downturn as a result of the closure of several large plants, and was in a good position to attract tourists, including many from across the border in the U.S., to a destination or 'resort' casino.
4.50 One year after the casino opened, local residents' casino gambling had increased four fold, with expenditure increasing by 25%. Participation was particularly marked among middle and lower income groups, and young and single people. Participation increased 4.6 times for low income individuals, but only 2.5 times for high income ones. The authors note that prior to casino opening, casino gambling was largely limited to higher income groups in Niagara Falls.
4.51 Problem gambling also increased by a statistically significant amount, from 2.5% to 4.4% 5. There were also significantly more reports of problems among family members and friends. The authors reported that these problems were confined to the city of Niagara, and did not spread throughout the rest of the province.
4.52 They also found that support for the casino among local residents actually rose, and that three quarters of the populations supported it one year after opening. This was despite the fact that both the harms and the benefits that many had expected from the casino failed to materialise. Many residents expected increased crime, problem gambling, marital and familial breakdown, traffic congestion, noise and drinking disturbances. With the exception of problem gambling, none of these materialised. On the other hand, fewer people reported benefits such as increased employment and incomes, more and better stores, services and entertainment, and increased property values, that they had expected a year before.
4.53 The authors' findings concurred with these perceptions, reporting only slight economic benefits in contrast to the significant gains predicted, and stated that, 'at least in the short term, problems from the increased availability of gambling manifested themselves not in the public arena but rather in the arena of private life' (Room et al 1999: 1449).
4.54 A positive relationship between casino proximity and gambling problems was also found in the 1999 New Zealand national survey (Abbott & Volberg 2000). In that study, higher prevalence rates were found in the two New Zealand cities, Auckland and Christchurch, with casinos than in other cities without them. In some analyses this relationship remained when other factors associated with problem gambling were controlled for statistically, suggesting the differences were a consequence of the presence of casinos.
4.55 If the numbers of individuals presenting for treatment after gambling availability has increased this is often taken as evidence of a relationship between the two. Evidence from the New Zealand gambling hotline indicated a sharp rise in calls from people reporting problems with casino gambling following the opening of a new casino (Sullivan et al 1997). However, such relations should be treated with caution, as there is no way of knowing whether the increase in calls was due to new problems arising from participation in casino gambling, a result of increased public awareness and advertising of the hotline number in the casino itself, or other factors.
4.56 There has been little research in Britain on the effects of availability on problem gambling, although the launch of the National Lottery in the U.K. in 1994 provided an opportunity to examine the impact of the introduction of a major form of gambling on gambling problems. One survey by a market research group found a 17% increase in calls to Gamblers Anonymous in the year following the introduction of the Lottery (Mintel 1995). Another study examined the Family Expenditure Survey data a year before and a year after the introduction of the Lottery for evidence of increased problem gambling, and found significant increases in expenditure on gambling after the introduction of the Lottery (Grun & McKeigue 2000) 6.
4.57 The authors argued that these results were consistent with the 'single distribution theory' which has also been applied to other areas of public health, such as alcohol consumption. The single distribution theory states that the extent of an activity like gambling is distributed amongst the population in a curve, characterised in the main by moderate consumption, but by a minority of excessive behaviour in the tail. The curve responds as a single entity to changes in overall distribution so, for example, when general gambling activity increases, then the proportion of excessive gambling will increase too.
4.58 Although there is some divergence in the findings reported here, much of the research suggests that factors that lead to increased availability and opportunities to gamble throughout the general population will also contribute to a corresponding increase in the prevalence of problem gambling. However, these relationships are complex. Recent longitudinal surveys and other research from North America, Australia and New Zealand has begun to cast doubt on this relationship, and suggests that it may apply in the initial stages of gambling expansion but, in some situations, subsequently break down.
Longitudinal Studies
4.59 While some of these studies of prevalence and incidence provide interesting results, they do not track changes over time, and so firm conclusions cannot be drawn about the relationship between availability and problem gambling. The question of whether gambling availability actually causes increased problem gambling is one that requires longitudinal research to answer. Several longitudinal studies have examined these relationships, and these have tended to produce some unexpected results.
4.60 The surveys in the first part of this section discuss 'gambling' in general, including, of course, casino gambling, even though this is not always discussed separately. The surveys in the second part do consider casino gambling separately.
Longitudinal general gambling studies
4.61 In New Zealand, longitudinal surveys have tracked patterns of gambling over seven years, and have found reduced rates of problem gambling, despite increases in availability 7. The first of these surveys was conducted in 1991 and found a lifetime pathological gambling prevalence rate of 2.7%; and a current one of 1.2%. A second survey conducted in 1996 estimated current pathological gambling to be 0.4%, a third of the previous study.
4.62 However, although comparable in terms of methodological quality to most gambling surveys the report's authors did not consider it to be sufficiently robust to generate reliable findings. One concern was the low response rate for Maori and Pacific Islanders, groups at particularly high risk for problem gambling in the previous survey (North Health 1996).
4.63 A third survey was completed in 1999 (Abbott & Volberg 2000). During the interval, gambling availability had dramatically increased: casinos had been introduced to Christchurch in 1996 and Auckland in 1998, and expenditure on gambling machines had increased three-fold. Given the increased availability and expenditure on high-risk forms of gambling, rates of problem gambling would have been expected to rise. However, they were actually significantly lower than in 1991: the lifetime estimate was one percent; the current estimate 0.5% 8.
4.64 Studies from Australia also appear to show decreases in problem gambling, despite increased in availability. One survey of four Australian cities conducted in 1991 estimated 'probable pathological gambling' to be 6.6% (Dickerson et al 1996). After this study had been carried out, Australia underwent a period of increased gambling availability and expenditure, particularly on high intensity electronic gaming machines, which greatly expanded in clubs, pubs and casinos. However, subsequent surveys actually measured lower rates, and the latest estimate stands at 2.1% ( PC 1999).
4.65 As with New Zealand, these findings do not support the expectation of increased problems with increased availability and expenditure. However, the authors of these studies were careful to point to the possibility of methodological shortcomings that they were, as yet, unaware of, and urged caution when interpreting results.
4.66 In addition to these New Zealand and Australian national studies, a number of repeat surveys have been conducted at state and provincial levels in Canada and the United States. Not all report similar trends to those of the Australian and New Zealand data. For example, studies in Iowa, New York, Minnesota and Québec found substantial and statistically significant increases in prevalence ranging between six to ten years (Volberg, 1995, 1996; Jacques et al, 1997).
4.67 In Minnesota and Iowa problem gambling increased significantly over a four year and six year period respectively (Emerson and Laundegran 1996; Volberg 1995). In the case of Minnesota, the increase followed a dramatic expansion in the availability of lotteries, high stakes bingo and more than twenty casinos in the state. In Iowa, the increase followed the introduction of riverboat casinos, Native American casinos and slot machines at racetracks.
4.68 A comprehensive review of fifteen longitudinal North American studies reported that seven showed an increase in prevalence estimates over time, and eight showed lower ones (Abbott 2001). In most cases, particularly where the interval between surveys was three years or less, changes were small and generally not statistically significant. In four of the six studies where the gap was more than three years, increases were apparent, and for the remaining two studies, slight decreases were found. A similar variation in findings is evident for the small number of Australian state-level 'replications', although again methodological problems compromise straightforward interpretation of these studies ( PC 1999).
4.69 In all of the studies considered above, gambling availability and expenditure has increased between surveys. However, what is notable is that, although the amount of individuals who gamble occasionally has increased, the proportion who report much heavier participation has reduced significantly (Abbott 2001; Volberg 2001). If the percentage of the overall population who gambles heavily on high-risk types of games - i.e. the group most likely to be problem players - actually decreases while overall expenditure rises, then it could be expected that problem gambling prevalence would level out or even reduce.
4.70 In this context, a recent study of four states is particularly interesting (Volberg 2006). In all the four states - Montana, North Dakota, Oregon and Washington State - substantial amounts of legal gambling existed at the time of the first (baseline) survey and increased further by the second (replication).
4.71 All four states introduced new casinos - two each in Montana and Oregon, five in North Dakota and ten in Washington State. In addition to opening this large number of new casinos, Washington State also allowed commercial 'card rooms' to greatly increase their maximum number of tables. Two states - Montana and Oregon - permitted electronic gaming machines to operate, although the density of machines was much greater in Montana than in Oregon. North Dakota was the only state without a lottery but had over 300 small charitable gambling operations in bars and restaurants.
4.72 Findings were unexpected. Casino gambling increased substantially, as would be expected from the dramatic increase in availability. Overall participation dropped however - probably as a consequence of casino gambling displacing other types. However, trends in changes in problem gambling were uneven: they increased in Montana and North Dakota and decreased in Oregon and Washington State. More severe cases of problem gambling changed most dramatically, with substantial and statistically significant decreases in Oregon and Washington State and substantial and statistically significant increases in Montana and North Dakota.
4.73 This is particularly surprising since Washington State, in particular, introduced large numbers of new casinos. The author concluded that 'these data suggest that something more than gambling availability and participation can affect the prevalence of problem and pathological gambling' (Volberg 2006). Looked at more closely, that factor appears to be the presence or absence of services to treat problem gambling. While Oregon spent around $2 million per year on services and Washington State around £150,000, Montana and North Dakota spent virtually nothing.
4.74 Overall, the results clearly demonstrated that problem gambling prevalence declined in the states with services and increased in the states without them. This indicates that availability is not the only factor at play in determining changes in rates of problem gambling, and that appropriate service provision can also be a crucial factor here.
4.75 Reports of declines in problem gambling, despite the introduction of new forms of gambling such as casinos, in the New Zealand and Australian studies mentioned earlier are consistent with the findings of this U.S. study. In New Zealand, between the 1991 and 1999 surveys, a national gambling problem helpline and extensive network of specialist counselling and treatment services was established. There was also a high level of publicity about the risks associated with excessive gambling and problem gambling.
4.76 In Australia too, as gambling availability and participation increased, many states and territories developed services to deal with problem gambling between the 1991 'four cities' study and 1998 national survey. It is likely that the expansion of such services has had some impact on overall rates of problem gambling in these jurisdictions. However, while this may be the case, as ever, it should be cautioned that further research is required to determine the impact of service provision and other factors on problem gambling prevalence.
4.77 In addition to these general studies of changes in problem gambling prevalence, a few longitudinal studies have examined the impacts of casinos alone.
Longitudinal Casino Studies
4.78 In 1994, Windsor, Ontario established a large destination or 'resort' casino which quickly became the highest-grossing casino in North America, taking in over one million dollars per day. The Windsor casino provided a unique opportunity to examine the effects of a large casino on a relatively small city (population 200,000), and a team of researchers examined its impacts one and four years after opening (Govoni et al. 1998; Frisch et al 1999). The study reported that interest in the casino peaked upon opening, with just under half the population visiting at least once. Expenditure among Windsor residents increased, and four years later, there was a large increase in the percentage of local residents who gambled (from 66% to 82%).
4.79 However, the study found no statistically significant increases in problem gambling either one or four years after the casino opened. Rates only moved from 3.6% to 3.7% over the period. Because of population increases, and increased numbers of people gambling, the absolute number of problem gamblers did change, rising from an estimated 4,600 to 6000. The authors noted however that this did not represent an increased risk of developing a gambling problem, and was not a direct effect of the casino.
4.80 In addition, the study found that residents' approval of the casino increased over time: with 63% approving of it four years after opening, compared with 54% before opening. Disapproval also decreased in the same time, from 30% before opening to 24% four years later. Crime also decreased in the downtown casino area, attributed to increased policing (Frisch et al 1999).
4.81 Another recent study of a casino opening in Quebec found no overall statistically significant increase in problem gambling five years after opening (Ladouceur and Jacques 2006). Researchers found that, when the casino opened, local residents visited more regularly and lost more money gambling than they had previously on other activities. However, this 'spike' of activity tapered off after five years. The proportion of pathological gamblers in the area rose from 1.4% prior to the casino to 1.8 % a year later, before dropping to 1% after five years. The Quebec average is 0.8%.
4.82 Yet another Canadian study is following casinos in four relatively rural areas, to assess changes in prevalence rates and local resident's perceptions of the casinos impacts in their communities over time (Nuffield and Hann 2003). Three years after they had opened, problem gambling had increased very slightly, but not by statistically significant amounts among residents in all of the four areas. In addition, spending patterns seemed relatively unchanged. In one area, Brantford, before the casino opened, residents in the two lowest income groups were most likely to have spent nothing on gambling in the last month, while those in the highest groups were most likely to have spent money on it. The differences were statistically significant.
4.83 Three years after the casino opened, these differences were no longer statistically significant, but the patterns remained unchanged: the lowest income groups were still most likely to spend nothing and the highest income groups most (Shercon Associates Inc. 2003). These findings were consistent with those of a large Ontario survey of problem gambling, which concluded that:
"low income persons are not more likely to gamble, but they are more vulnerable to gambling problems if they do. Individuals with incomes less than $30,000 are the most likely to be classed as problem gamblers. Interestingly, they are also the most likely to be non-gamblers" (in Shercon Ass. 2003, 54).
4.84 In addition, in each of the areas, residents were more aware of gambling treatment facilities after the opening of the casino than before. In all areas, and in Brantford in particular, there was a substantial increase in the amount of public advertising of the phone number for the Problem Gambling Helpline, and a local Addiction Services agency had undertaken a wide ranging series of public presentations to various public organisations, the general public and schools, raising awareness of the risks of problem gambling.
4.85 Again, in Brantford, the numbers of individuals seeking counselling for gambling problems increased after the opening of the casino. This increase was much higher, proportionally, than the slight increase in numbers of problem players. It could be speculated that the awareness raising and extension of treatment had helped to keep the numbers of potential problem players down.
4.86 A note of caution should be sounded here, however. These Canadian casinos may be very different, in terms of their local populations and locations, than those proposed for Scotland. They tend to be located in areas that are relatively sparsely populated, with relatively affluent residents, and attract large numbers of tourists which provides their main turnover. It is likely that the applicability of findings from these studies to Scotland would not be straightforward.
4.87 Moving to urban casinos, the research base on social impact is much less developed.
4.88 In New Zealand, Abbott, Williams and Volberg (1999; 2004) examined rates of problem gambling from their longitudinal survey. In the seven years between surveys, casinos had been introduced to two of New Zealand's three major metropolitan centres, providing a natural experiment whereby participants living in them could be compared with their counterparts in the city without a casino. It was also possible to control statistically for other factors that differed between the casino and non-casino centres, including those known to be associated with problem gambling. The study did not find any significant impact from casinos on participants' problem gambling.
4.89 However, although the results did not show a relationship between the introduction of casinos and problem gambling, it should be noted that participants were not fully representative of the adult population living in centres studied. All either had gambling problems or gambled frequently without problems prior to the introduction of casinos. By the time they were re-assessed, many had reduced their interest and involvement in gambling and all were in their mid-twenties or older.
4.90 While it cannot be concluded from that study that casino introduction had no effect on gambling prevalence in the general population, the findings do suggest casinos did not create more severe problems for people with existing problems. As mentioned earlier, the 1999 national survey was representative of the adult population, and it did find higher prevalence rates in the two major cities with casinos. It is possible therefore that any increased problems were concentrated in the sectors of the population not included in the longitudinal study (Abbott & Volberg 2000)
4.91 From the studies reviewed so far, it can be seen that the relation of gambling availability to problem gambling, particularly with respect to casinos, and over time, is extremely complex. While the National Research Council concluded that research findings were "consistent with the view that increased opportunity to gamble results in more pathological and problem gambling" ( NRC 1999, 84), the Productivity Commission acknowledged the ambiguity and decided that studies were "inconclusive about the links between access and problems" ( PC 1999, 8.27).
4.92 In attempting to account for the existence of stable or declining rates of problem gambling over time, even when availability increases, some researchers have suggested that societies might undergo a process of adaptation, whereby they adjust to the existence of gambling around them, and develop various informal controls and social sanctions that protect against developing problematic behaviour (Shaffer et al 1997) 9.
4.93 In 2004, the Responsibility in Gambling Trust ( RIGT), in the U.K., commissioned an international review of the research evidence on the risk factors associated with problem gambling (Abbott et al 2004). After an extensive analysis, the report's authors concluded that the relationship between availability of gambling opportunities and increased rates of problem gambling was had not been established and if existed was not linear.
4.94 They identified studies which showed jurisdictions in which rates remained the same or even declined over time, despite the introduction of increased opportunities to gamble. They also highlighted the idea of 'adaptation', whereby communities develop informal coping mechanisms to deal with gambling availability, and also the influence of public health campaigns to raise awareness of the risks of problem gambling, and availability of services to provide treatment for those who develop problems.
4.95 The Report concluded that:
"While acknowledging the various methodological problems and potential for confounding factors to complicate interpretation, in our view the replication [longitudinal] studies do not support the proposal that gambling problems invariably increase with rising levels of gambling exposure. This does not mean there is no relationship. Rather it is probable that other factors, such as greater awareness of problem gambling, availability of problem gambling services, and changing participation patterns, have potential to counteract the problem generating effects of increased availability. This may be a highly dynamic process, with the relative balance between risk and protective factors shifting over time" (Abbott et al 2004: 156)
Conclusions
4.96 Availability of gambling and type of game are features that are strongly associated with problem gambling. These features come together in non-casino EGMs which, worldwide, show the highest association with gambling problems. The relationship with other forms of gambling is less clear cut. Casinos are also associated with high rates of problematic behaviour, although the picture becomes more complex here, with a range of other factors, including the size and location of the casino, coming into play. Simply not enough is known about these complex variables, and the way they interact to influence problem behaviour to allow conclusive statements on the impacts of casinos to be made. Although much of the research seems to suggest that proximity to casinos increases rates of problem gambling in the local population, longitudinal research also suggests that such rates may well decline over periods of around two to seven years as communities adapt to the presence of gambling around them. Furthermore, there is also evidence to suggest that public health initiatives, such as the provision of treatment and information about problem gambling, can militate against such increases.
« Previous | Contents | Next »