The Scottish Government

« Previous | Contents | Next »

Listen

CHAPTER THREE: CURRENT ISSUES IN GAMBLING RESEARCH: MEASUREMENT AND PREVALENCE

3.1 Unfortunately, the evidence base that is needed to address many of the controversies surrounding the social impacts of gambling is, in many areas, lacking. Research into the impacts of gambling is still in its infancy, and is beset with problems.

3.2 Although some data on the social impacts of gambling exists from an international perspective, the evidence base for the U.K., and particularly Scotland, is extremely thin. In addition, much of the available material is methodologically weak and open to interpretation. Although some economic effects are calculable, the social impacts are less easy to quantify, and, to date many studies have produced inconclusive or contradictory results.

3.3 This can exacerbate the controversy that surrounds gambling, with an evidence base that is often not able to resolve the most contentious issues, and which can be interpreted differently by those who are broadly 'for' and those who are 'against' the expansion of gambling. For example, estimates of the prevalence rates of problem gambling tend to be interpreted differently by treatment providers and the gambling industry, with the former highlighting the highest estimates to support their case for funding, and the latter emphasising the lowest ones to deflect attention from the potential harms of their product.

3.4 At the end of a lengthy and exhaustive investigation of the literature, the U.S. National Research Council ( NRC) concluded that the evidence base was so weak in many areas as to make conclusions difficult or impossible to make. It wrote that the study of the costs and benefits of gambling is "still in its infancy", and criticised studies that "utilised methods so inadequate as to invalidate their conclusions", the lack of "systematic data", the substitution of "assumptions for missing data", "haphazard" applications of estimations in many studies, and a general lack of identification of the actual costs and benefits to be studied in the first place ( NRC 1999: 5-18).

3.5 Some of the problems with this research are summarised below.

To begin with, the large sample sizes that are required to produce robust results due to the relative rarity of problem gambling in the population has made survey research expensive, and as a result research that has used adequate sample sizes is rare. Because they are relatively easier to access, many studies have utilised problem gamblers in treatment. The use of such participants can skew results, as individuals in treatment have distinct characteristics, including more severe problems, and are not representative of problem gamblers in the general population. An additional problem with such types of recruitment, and indeed, with many studies of problem gamblers in general, is the lack of control groups, which makes comparisons difficult.

3.6 Many studies provide a static 'snapshot' of the effects of gambling, rather than tracking changes in its development over time. This is particularly important in accounting for the ways that the impacts of gambling can be influenced by other factors, and can vary over time. In particular, the impacts of casinos need to be monitored through longitudinal surveys of a range of social and economic factors, with findings tested against control groups. In all but a minority of studies, this has not happened.

3.7 In addition, the study of gambling and problem gambling has been dominated by medical and psychological perspectives, which tend to focus on gambling as a mental health issue. Here, the focus is on 'pathological' or 'problem' gamblers as individuals with particular mental, personality or even physiological problems. This rather narrow focus tends to draw attention away from the wider effects of gambling on communities and societies as a whole. As such, studies of the impacts of gambling on particular social groups, its effects on communities and its relation to factors that involve wider social frameworks, such as socio-economic deprivation, have been relatively rare.

3.8 In gambling research as a whole, questions of causality are also an issue. Gambling problems often co-exist with other problems, such as substance abuse, mental health problems, and general criminality in relations that are described as 'co-morbid'. In such a situation, disentangling the impacts of one condition or behaviour on the others is extremely difficult. Establishing causation is also fraught with problems, and most studies can do no more than point to possible associations between gambling and a range of adverse effects. The issue of whether gambling causes certain negative impacts, such as crime, divorce and suicide; is a contributory factor, or is merely associated with such conditions, which may well occur anyway, has created much controversy and has not been satisfactorily resolved.

3.9 As the NRC explained, "evaluating studies of conditions that co-occur with pathological gambling requires careful formulation of research questions, such as 'Does gambling precede the onset of other disorders? Do certain disorders exacerbate pathological gambling? Is there a pattern of syndrome clustering?…' Very few studies have addressed even one of those questions" ( NRC 1999: 4-13).

3.10 Such methodological weaknesses are often used to justify the claim that it is not gambling itself that creates problems, but rather pre-existing problems in individuals that results in excessive behaviour. This argument, often proposed by industry, is that it is not so much the gambling product that creates problems, but rather individuals with other difficulties who are drawn to it, and who would manifest problematic behaviour whether or not gambling was available. Such a stance is summed up by one industry official, who said: "Do problem gamblers exist? I am yet to be convinced of this, however I fully acknowledge that there are people with problems who gamble" (Mr Windross, Managing Director, TAB; in PC 1999: 7.4).

3.11 This argument has implications for policy. If problem gambling is regarded as a rare behavioural and/or mental disorder, then only relatively small numbers of vulnerable individuals can be expected to be affected. The best approach could arguably be to simply target treatment towards this small group. If, on the other hand, problems are considered to lie within the product itself, then far greater numbers of people are potentially at risk and policy might be better directed towards limiting the expansion of gambling facilities and/or educating the general public about the risks of harm.

3.12 The vast majority of gambling research demonstrates that the former scenario is not the case. Problem gambling is not generally regarded as a condition affecting a minority of disturbed individuals who would, in the absence of gambling, display other forms of problematic behaviour. Rather, it finds that gambling problems are created by a range of complex associations between individuals, their environment and the gambling opportunities that they are exposed to.

3.13 As a result of these polarised debates, both the negative as well as the positive effects of gambling, particularly casino gambling, tend to be overstated. Claims that casinos create significant wealth, new jobs and regenerate local economies are often exaggerated ( PC 1999; NGISC 1999; Goodman 1995; Grinols 1995). On the other hand, claims that they create massive social problems through increases in problem gambling and crime tend to be equally exaggerated. In fact, it is probably the case that a diligent researcher could select a range of studies to present quite a convincing case either for or against casino expansion. The reality, as usual, is more mundane: overall, the evidence base shows that casinos are neither as economically beneficial as supporters claim, nor as socially damaging as opponents fear.

3.14 A recent symposium on 'The socio-economic impact of gambling' held in Canada in 2003, summed up these problems, and is worth quoting at length:

"Both proponents and critics of state-sanctioned gambling are adamantly committed to their perspectives and each revels in citing the most recent research that appears to support their position. On the basis of this nominal evidence, each group seeks to lobby governments to make gambling-related public policy decisions that are in keeping with their world views…. However, while some of this gambling cost-benefit research is scientifically rigorous, much of it is of dubious quality; in many instances because it has been conducted by proponents or opponents seeking to promote their particular agenda. When the end justifies the means, scientific research suffers…. This circumstance creates a classic policy dilemma for governments; that is, on the basis of purported evidence, largely promoted by those special interest groups, is gambling (a) net beneficial to society, and therefore worthy of continued government policy support, including expansion, or (b) net-costly to society, suggesting that policies to curtail and even significantly decrease legal gambling opportunities are more appropriate?" (Wynne and Shaffer 2003: 112-3).

3.15 However, despite the problems noted here, a core of rigorous research does exist, and has provided some evidence of the social impacts of gambling. It has been utilised throughout this report. In particular, two landmark, large-scale surveys from America and Australia have reviewed a huge amount of published research, as well as conducting new research of their own into the causes, impacts and policy implications of problem gambling. In America, the National Gambling Impact Study Commission ( NGISC) was established to review the state of gambling in that country. It commissioned an analysis of professional literature by the National Research Council ( NRC), and initiated new research by the National Opinion Research Centre ( NORC) at the University of Chicago.

3.16 In Australia, a Productivity Commission ( PC) was established to investigate the state of gambling in that nation. It conducted new research into all aspects of gambling and problem gambling, conducted an exhaustive review of international studies, and produced the report, Australia's Gambling Industries (1999).

3.17 Both reports were aware of the methodological, and other, problems relating to gambling research, and so, informed by this, their conclusions and recommendations were cautious, balanced and rigorous. It is generally considered that they represent the most authoritative statements on the impacts of gambling and problem gambling to date.

Definitions and Measurement of Problem Gambling

3.18 Gambling is a broad concept that includes a range of different activities, including betting on races and gambling machines, gambling in bingo halls, amusement arcades and casinos, as well as playing the lottery and on the Internet. For the vast majority of players, gambling is an enjoyable form of leisure, pursued for a range of reasons - to relax, to socialise, to experience some excitement, and perhaps to win money. For these individuals, gambling losses are simply the price of the entertainment, in much the same way as cinema or football tickets are the cost of a good time.

3.19 However, for a minority gambling losses are uncontrolled and often represent a desperate attempt to win back money rather than to enjoy a game. For these individuals, problem gambling consists of behaviour that is out of control and that disrupts personal, family and employment relations. It can also initiate a spiral of related problems from debt and bankruptcy, to criminal activity, domestic violence, familial breakdown and suicide.

3.20 A substantial amount of research into the causes of such behaviour has been conducted internationally. Although there is considerable disagreement over the exact nature of the relationships and causes of the disorder, in general experts agree that it involves some degree of failure to control impulses to act on the spur of the moment, and is often tied in with a misunderstanding of how games work and with faulty perceptions that they can be controlled. Failure to keep track of the amount of time and money spent gambling and efforts to 'win back' losses, are also predictors of problems. These characteristics have implications for the formation of responsible gambling policies, discussed in Chapter Six.

3.21 The terms used to describe problem gambling are derived from largely psychiatric attempts to classify and measure the behaviour, meaning that discussions of such behaviour tend to be based on the assumption that problem gambling is a clinical psychological disorder.

3.22 The term 'pathological gambling' was first included in the third edition of the Diagnostic and Statistical Manual ( DSM-III) of the American Psychiatric Association (1980), where it was described as an impulse control disorder - a compulsion characterized by an inability to resist overwhelming and irrational drives. Other impulse control disorders include, for example, kleptomania and pyromania. Focus soon shifted to its supposedly addictive characteristics however, and subsequent editions of the manual saw it reclassified in terms similar to those for psychoactive substance dependency with the term 'pathological gambling' consistently used to reflect its chronic, progressive character ( APA 1987, 1994).

3.23 The criteria used to define it derive from three broad conceptual themes often associated with 'addictions' to substances such as drugs and alcohol, namely, compulsion or craving, loss of control, and continuing the behaviour despite adverse consequences. Although, technically, pathological gambling is defined here as an impulse control disorder, many clinicians and commentators consider it an addiction, akin to disorders associated with drug and alcohol dependency.

3.24 Around the same time as these debates were going on, another screen - the South Oaks Gambling Screen ( SOGS) - was developed and became widely used for the measurement of gambling problems (Lesieur and Blume 1987). Although it is widely used, many criticisms have been levelled at the SOGS, the most significant being that it was designed to test individuals for psychological problems in clinical settings, not for general populations, and also that it tends to produce many 'false positives': i.e. that it tends to over-estimate the numbers of problem gamblers in surveys of the general population.

3.25 In both the SOGS and the DSM-IV, a positive response to five or more questions results in a classification of 'probable pathological gambler'; less than five classifies a 'problem gambler'.

3.26 Since the introduction of these two screens, over twenty others have been developed for a range of purposes, including screening, diagnosis, population monitoring and treatment planning. In general, they define pathological and problem gambling as behaviour that is out of control, and that has come to disrupt personal, family, social and vocational life, with the former regarded as a more severe condition than the latter. In reality however, the terms are often used interchangeably and, moreover, it can be difficult to distinguish between them since pathological gamblers will undoubtedly have been problem players at some point, and both types of players can experience fluctuations in the severity of their condition.

3.27 More recently, a public health approach has attempted to simplify these various categories into a more streamlined diagnostic framework. Rather than defining problem or pathological gamblers as a discrete group, with qualitatively different characteristics from 'normal' or recreational gamblers, this approach proposes a system that measures gambling behaviour along a continuum with less severe problems at one end, 'pathological gambling' at the other, and a variety of states of more or less severity in between. (Shaffer and Hall 2001). From this perspective, problem gamblers, as well as individuals who score even lower on problem gambling screens ('at-risk gamblers') are of concern because they represent much larger proportions of the population than pathological gamblers, and because their difficulties may become more severe over time.

3.28 Somewhat ironically, the definition and measurement of the extent of problem gambling in a population can itself be problematic. The terms used to describe problem gambling are derived from the different screening instruments outlined above, and the lack of a single, standard screen can make comparisons between studies and over time difficult. Also, prevalence rates can be measured in one of two ways: by counting those who meet the criteria for problem or pathological gambling only in the past twelve months ("past year gamblers"), or by including all those who have ever fulfilled the criteria ("lifetime gamblers"). There are problems with both approaches, as "lifetime" measures may overestimate the problem, while "past year" measures may underestimate it.

3.29 It is worth pointing out here that, despite the emphasis on psychiatric characteristics, it could well be more useful to adopt a more pragmatic definition of problem gambling. Quite simply, for many gamblers, 'problem' behaviour can be defined in more straightforward economic terms as playing that they can no longer afford. When considering the correlations of problem gambling with social deprivation, which, it should be noted, most international studies fail to do, this simple but crucial feature needs to be borne in mind.

The Prevalence of Problem Gambling

3.30 The lack of a standard, universal criterion leads to difficulties in estimating the extent of problem gambling. However, despite this, many jurisdictions have made efforts to measure levels of problem gambling in their populations, and these surveys have produced relatively similar rates. Most western nations with developed gambling industries report rates of problem gambling between 0.6 and 3%.

3.30 A summary of international rates, based on past year prevalence, is provided in Table 3.1. It can be seen that Britain currently has relatively low rates of problem gambling, compared with those of other jurisdictions.

3.31 Although it is clear from this that numbers of problem gamblers are relatively small, they nevertheless account for a large proportion of gambling expenditure (i.e. losses). Some studies have shown that between a third and a half of casino profits come from this group (Lesieur 1998, Volverg 2001) while others have demonstrated they account for a third of the entire gambling industry's market ( PC 1999).

Table 3.1. International Past Year Gambling Prevalence Estimates

Source: Shaffer et al. 2004

Problem Gambling Prevalence in Britain and Scotland

3.32 In 1999, the National Centre for Social Research (NatCen) conducted the first prevalence survey for Great Britain. The British Gambling Prevalence Survey utilized both screens ¹ to determine the prevalence of those who had experienced problems in the past year. The DSM-IV put the figure at 0.6%, or 275,000 people, while the SOGS estimated it to be 0.8%, or 370,000 people. Rates for Scotland were 0.7% and 1%, although numbers were not large enough to be statistically significant. As a point of comparison, the estimated number of problem drug users in this country is 270,000 (Frisher et al 2001).

3.33 Consistent with the demographic profile of problem gamblers in other countries, the British survey provided a profile of a problem player who was male, aged 35 or under, had parents with gambling problems and was on a low income. Problem behaviour was also associated with starting gambling at an early age and being separated or divorced. Numbers of casino gamblers in the survey were too small to determine the profile of problem players, although other studies have done this and this is discussed in the next chapter.

3.34 The survey found a significant correlation of problem gambling with household income, with those in the lowest income categories nearly three times as likely to be defined as a problem gambler. Prevalence amongst men was found to be two to three times higher than amongst women. Problem gambling was also inversely correlated with age, decreasing as age increased, although this relationship was clearer for men than for women. Young men, defined as those less than 35 years in the case of the SOGS and less than 25 years in the case of DSM-IV were, therefore, the group with the highest prevalence of problem gambling. Single people were more likely to be problem gamblers than married people, although as the authors point out, this is likely to be a function of age (single people tend to be younger) than marital status per se.

3.35 Problem behaviour was also associated with having parents who gambled and starting gambling at an early age.

« Previous | Contents | Next »