Introduction

In most jurisdictions governments regulate access to legalised gambling by: 1) licensing operators and; 2) restricting access to gambling services by controlling numbers and distribution of venues, table games and electronic gaming machines (Delfabbro 2008; Eadington 2004). However the converging capabilities of technological devices such as personal computers (PCs) and devices (PDAs), mobile phones, interactive television, set top boxes and games platforms (PlayStation, X box, PSP) enable unlimited access to the Internet with the potential to gamble 24 h a day 7 days a week (see Griffiths 2003; Griffiths et al. 2006). This effectively means that every television, mobile phone and computer terminal represents an opportunity for any individual, regardless of age, to be in the personal possession of a portable electronic gaming device (King, Delfabbro, and Griffiths 2010; McBride and Derevensky 2008).

Debate continues as to whether ease of access and availability of gambling contributes to the incidence of problem gambling. Some authors have observed relationships between access to gambling and incidence of problems (e.g. Abbott and Volberg 1991, 1996; Volberg 1994), but the relationship may not be inevitable. It has been suggested that individuals and communities may adapt to increased exposure to gambling (Abbott 2006; Shaffer et al. 2004a). Indeed, access to online gambling may not automatically lead to the development of problems as the majority of online gamblers have been observed to gamble in a relatively controlled fashion (LaBrie et al. 2007; 2008; LaPlante et al. 2008). Even if increased access to gambling has relatively minor effects upon the majority of the community, it may appreciably increase the degree of problem for specific individuals (i.e. at risk individuals) (Lund 2008). The present paper considers factors (e.g. interest in technologies) that may increase risk of problems. In addition, as there is considerable potential to monitor online behaviours (Shaffer et al. 2010), and identify overt behaviours that are linked to problem gambling such as duration of play or seeking funds (e.g. Delfabbro et al. 2007; Schellinck and Schrans 2004, 2006), it is important to identify behavioural markers (e.g. surfing behaviours) that could serve as potential indicators of gambling problems (LaBrie et al. 2007, 2008; Shaffer et al. 2010).

Concern over excessive gambling has emerged in response to the increasing availability of computing devices connecting users to financial institutions and products via the Internet (Griffiths et al. 2009). The ability of computing devices to support financial transactions has created new avenues for companies to target consumers in workplace (via personal computers), domestic (via interactive television) or open-air (via wireless or mobile phone) locations. Access and promotion of, gambling opportunities through the technology is now considered to represent a significant risk factor for impulse control (Griffiths 2003, 2007a), and the exacerbation of existing conditions (Widyanto and Griffiths, 2006). In the following sections, the role of technologies (i.e., internet, mobile phone, television) in gambling will be discussed.

Internet

The Internet has allowed the gaming industry to set up websites in unregulated or poorly regulated jurisdictions offering gambling products to consumers within and external to their jurisdiction (Parke and Griffiths 2004), circumventing any restrictions and controls (Eadington 1988, 2004). For instance, Australian expenditure on legal forms of gambling is estimated to be AUD$19 billion dollars (Productivity Commission 2010), with a further AUD$790 million dollars estimated to be spent on illegal online gambling (Productivity Commission 2010). Indeed, there is evidence indicating that higher proportions of Internet gamblers meet the criteria for problem gambling than non-internet gamblers (Griffiths and Barnes 2008; Griffiths et al. 2008; Wood and Williams 2007, 2009). Anonymity, privacy, accessibility and ease of electronic transfer of funds contribute to the popularity of online gambling (Griffiths et al. 2006), and some studies have considered whether the Internet promotes disinhibitory behaviour and decisions (Griffiths et al. 2006; Suler 2004), or impulsive tendencies (Mottram and Fleming 2009).

Mobile Phone

The high market penetration of mobile phones has the potential to place mGaming in the hands of almost everybody, adolescent to adult, within the community (Griffiths 2003, 2007b). Increased personal access 24 h a day, 7 days a week could pose potential problems for those with poor impulse control (Billieux et al. 2008; Phillips et al. 2006). Accordingly, it is reasonable to hypothesise a higher rate of problem gambling among individuals gambling using both Internet and mobile phones.

Interactive Television

With the development of mobile phone and digital television technology, it is becoming possible for consumers to respond to broadcast programming whether via SMS, set top boxes or remote controls. This two-way interaction means that it is possible to vote, enter competitions, place bets on horse races and purchase merchandise from a broadcaster, with this capability improving as more efficient return paths are established. It is predicted that forms of interactive programming will be particularly appealing to vulnerable sections of the community, specifically those that are impulsive and who tend to view television to a regular and excessive degree (Griffiths 2007a; Widyanto and Griffiths, 2006).

Online Gambling

The provision of gambling services online is a comparatively new industry that in many jurisdictions may occur illegally (Cotte and Latour 2009; Griffiths 2010; Scoolidge 2006) or legally through government approved sites. Currently, China and the USA ban online forms of gambling, while in Australia, sports and wagering bets are permitted with licenced operators but the provision of other forms of online casino type games to Australian citizens is prohibited. In the UK and Gibraltar, legitimate sites operate within their jurisdictions. Setting these aside, there are multiple online providers offering unregulated products across international jurisdictions and this might manifest as consumer exploitation, dissatisfaction and lack of complaint resolution (see Blaszczynski et al. 2008).

There are concerns as to the probity of online casinos (Scoolidge 2006). For example Sévigny et al. (2005) found online gambling sites provided inflated odds of winning on play-for-free demonstration sites fostering the illusion that players were able to win. This procedure, comparable to ‘bait and switch’ techniques used in retail stores, encourages transition to pay-to-play sites that have poorer return to player rates resulting in losses to players.

Other sites ignore responsible gaming guidelines. For example, Smeaton and Griffiths (2004) examined 30 UK gambling websites and found that more than half did not verify the age of consumers, and almost all sites (29/30) did not have a self-limiting option, or references to gamblers help services (26/30), but as timelines for access were not provided, more research on gambling site regulation is needed.

Given the dubious status of some online gambling providers, it is likely that unregulated operators offer scant consumer protection systems (Scoolidge 2006; Smeaton and Griffiths 2004). Some studies have found that online gaming sites continue to offer or promote opportunities to gamble even as a person is attempting to leave the site and discontinue gambling (Griffiths and Parke 2002; Griffiths 2003).

Another concern with online gambling sites is that some sites are not particularly user friendly for players seeking assistance or access to treatment services or advice (Monaghan 2009; Smeaton and Griffiths 2004). Indeed, as the host organisation may be based overseas, the ability to resolve dispute across international borders is minimised if not precluded entirely (Griffiths and Parke 2002; McMullan and Rege 2007; Scoolidge 2006). Furthermore, it may be difficult for online gamblers to detect ‘bogus’ Internet gambling sites given that such sites may appear legitimate, claim to be government or 3rd party regulated (misuse kitemarks), or steal website layouts and designs of reputable companies (Griffiths 2010).

Unfortunately online assistance may be perceived as inadequate where available. Griffiths et al. (2009) administered online surveys to 2,348 online gamblers accessing PlayScan, a system that allows for individuals to set spending limits, view their gambling profile or take a self-diagnostic gambling test. Although specific features of PlayScan were rated as useful, only 26% had actually used the system; hence those deemed to be most ‘at risk’ may choose to avoid systems which may help to prevent online harm.

The use of dubious online gambling sites can potentially lead to problems. There are already a number of websites currently documenting negative online experiences of gamblers (e.g. http://www.casinomeister.com/rogue/index.php). Negative experiences may manifest as increased exposure to Malware such as Spam (http://www.wilmerhale.com/files/upload/FTC_SPAM.pdf), a component of online casino marketing strategies (http://www.casinomeister.com/online_casino_spam.php) that may also represent attempts by cyberextortionists seeking to recruit users’ computers for distributed denial of service attacks (McMullan and Rege 2007). Attempts to initiate complaints may compound problems. There are unfortunately fake-watch-dog sites that remove a consumer’s opportunity to complain and seek assistance. For instance Spammers have created fake unsubscription sites (http://www.spamhaus.org/consumer/removelists.html) with bogus online gambling consumer protection sites also in existence (http://www.casinomeister.com/online_casino_spam.php).

Given new technologies may increase availability and risk of developing a gambling problem (Griffiths 2003, 2007a, b), the present paper considered online risk factors and behaviours that might be linked to gambling problems. This study explored factors (e.g. interest, demographics) promoting access to online gambling services and the development of problem gambling (see Griffiths and Barnes 2008). In addition, we were also interested in investigating Internet surfing behaviour to develop behavioural indices of the likelihood that people would click on to gambling-related, regulatory or counselling sites on the Internet. This would offer insights into the extent to which consumers surf for electronic ‘action’ or ‘assistance’. Further, the consequence of accessing dubious online gambling sites may reflect higher levels of consumer dissatisfaction, manifesting as increased problems with spam and telemarketers and act as an additional index of involvement and impaired control over online behaviours.

Method

Participants

The total sample consisted of 1,141 respondents (490 males and 646 females) ranging in age from 16 to 75 years of age (M = 37.7, SD = 12.79). The majority had some form of university qualification, with 354 (31.2%) having attained an undergraduate degree and 272 (24.0%), a postgraduate degree. Of the remainder, 201 (20.2%) had completed a technical or business college course, and 279 (24.6%) reported secondary school as the highest education attained. There were no statistically significant gender differences in age t(1121) = 1.214, p > .05 or education t(1129) = 0.657, p > .05.

Of the total sample, 654 participants (72.5%) were classified as non-problem or non-gamblers; 133 participants (14.7%) were low-risk gamblers; 87 participants (9.6%) were moderate-risk gamblers; and 28 participants (3.1%) were problem gamblers using the Problem Gambling Severity Index of the Canadian Problem Gambling Index (Ferris and Wynne 2001).

Materials

A battery of questionnaires assessed problem gambling status, degree of interest in online gambling, exposure to spam, dissatisfaction and other aspects of online gambling and Internet use.

Problem Gambling Severity Index

The Problem Gambling Severity Index (PGSI) of the Canadian Problem Gambling Index (CPGI) (Ferris and Wynne 2001) is a nine-point self-report scale used to assess problem gambling status. Scores on the PGSI range from ‘0’ to ‘27’, and are used to classify people into: 0=non-problem or non-gambler, 1–2=low risk gambler; 3–7=moderate risk gambler; and 8+=problem gambler. It is a reliable scale with a Cronbach’s alpha of .84, and a test-retest reliability of .78 (Wynne 2002).

Digital Access Scale

We developed a scale examining interest in access to digital technologies. There were five items addressing availability of digital services (e.g. cable TV, internet access). Items were: 1) I like to subscribe to cable television networks (e.g. SkyChannel, Foxtel etc.); 2) I prefer to visit places (e.g. hotels, motels) that subscribe to cable television networks (e.g. SkyChannel, Foxtel, etc.); 3) I will only choose to stay in hotels with free Internet connections when I travel; 4) I frequently use Internet cafes to go online on the Internet; and 5) We should all take up digital television as soon as possible. These items produced a “Digital Access” scale addressing interest in access to digital technology. The scale had a Cronbach’s alpha of 0.668. Although this is not ideal psychometrically, such a value is acceptable for research purposes (DeVellis 1991). Construct validity was supported by significant correlations between the Digital Access scale and the Internet Problem Use Scale (r = 0.301, n = 997, p < .001). There were also small but significant positive correlations between an interest in digital access and self-reported work-related Internet use (log transformed) (r = 0.146, n = 1003, p < .001) or non-work-related Internet use (log transformed) (r = 0.151, n = 1006, p < .001). A person desiring digital access was more likely to report problems controlling their Internet usage and their levels of Internet use tend to be somewhat higher.

Problem Technology Use Scales

Scales were also administered to measure problem usage patterns associated with a variety of Internet Communication Technologies (i.e. Internet, Mobile Phones, television). The scales were developed from a 20-item questionnaire designed to measure problem Internet usage (Armstrong et al. 2000) and a 27-item questionnaire aimed at measuring problem mobile phone usage (Bianchi and Phillips 2005). An item analysis (Phillips et al. 2008) suggested that five specific items predicted problems controlling the use of these technologies. The general form of the items were: 1) diminished control—I find myself using the technology for longer periods of time than I intended; 2) importance—I find it difficult to keep up to date with current affairs and sports, without using the technology; 3) problems—I find myself using the technology when I should be doing other things and it causes problems; 4) mood enhancement—The technology makes me feel better when I am feeling down; 5) escapism—I find myself using the technology as a way of escaping from daily stresses.

The scales all had good reliabilities. The Internet Problem Use Scale (IPUS) obtained a Cronbach’s alpha of 0.860 and the problem use scale for mobile phones (MPUS) had a Cronbach’s alpha of 0.900. A Television Problem Use Scale (TVPUS) obtained a Cronbach’s alpha of 0.804.

As a test of the validity each Problem Use Scale was correlated with self-reported usage. There were significant correlations between a specific Problem Use Scale and self-reported use of that specific technology. For instance the IPUS correlated with self-reported Internet use (log transformed) (r = 0.485, n = 1041, p < .001), the MPUS correlated with self-reported use of mobiles (log transformed) (r = 0.442, n = 1006, p < .001). In addition, the TVPUS correlated with self-reported television use (log transformed) (r = 0.286, n = 1074, p < .01).

Problems with Digital Services

We developed a series of items to address participants’ dissatisfaction with current protective measures, particularly the concern for electronic privacy and attempts to protect themselves from scams and spam. These items were: 1) I believe that complaints about telemarketers have no effect upon the numbers of unwanted phone calls I receive; 2) I believe that unsubscribing to an electronic mailing list has no effect upon the amount of SPAM I receive; and 3) I believe that unsubscribing to an SMS mailing list is of no use at all. Each question used 6-point Likert scales to assess the extent to which participants agreed with each statement. These three questions were sufficiently highly correlated that they could generate a scale with a Cronbach’s alpha of 0.715. The scale addresses dissatisfaction with current measures that are in place to protect electronic privacy. Construct validity was demonstrated as the scale had low but significant correlations with the amounts of Spam received (log transformed) (r = 0.195, n = 1003, p < .001) the number of times people reported being contacted by telemarketers (log transformed) (r = 0.288, n = 1009, p < .001) and interest in digital services (r = 0.121, n = 998, p < .001).

Procedure

The Monash University ethics committee approved the conduct of the study. The questionnaire was advertised in Australian national newspapers and posted on a website until completion in December 2008. Participants completed the electronic survey to be entered in a draw to win one of 10 iPods.

Results

Scores above the midpoint on problem technology use scales indicate the extent to which respondents considered they had a problem with their use of specific digital services (see Table 1). High scores on the problem Internet use scale indicate that people reported problems limiting their Internet use. In contrast, lower scores on the problem mobile phone use scale, and problem TV use scale indicate less problems with these technologies. However higher scores on the problems with digital services scale indicate a degree of dissatisfaction with protective measures associated with digital services.

Table 1 Self reported levels of problems with digital services
Full size table

As there are indications that some people had problems controlling their use of specific digital services, multiple regression analysis was conducted to determine which variables could predict problem gambling (see Table 2). A significant proportion of the variance (7.7%) could be accounted for F(7, 835) = 11.059, p < .001 by the study variables. An interest in digital services t(835) = 4.000, p < .001, problems with television t(835) = 2.534, p < .05 and mobile phone use t(835) = 4.192, p < .001 predicted problem gambling status. Male gender t(835) = −2.671, p < .01, and lower education levels t(835) = −2.719, p < .01 were also significant predictors of problem gambling status. Problems controlling online impulses predicted gambling problems, and indices of online access and use also predicted gambling problems.

Table 2 Electronic interests and demographic factors predicting risk of developing a gambling problem
Full size table

Upon completion of the questionnaire participants were asked what type of information they would like to see at the next link. 24.3% of participants selected a regulatory site, 6.2% a counselling site, and 3.4% a gambling site. These surfing preferences along with other online behaviours such as complaints were entered into a multiple regression to predict the risk of developing a gambling problem (see Table 3). A significant proportion of the variance (12.4%) could be accounted for F(6, 880) = 21.871, p < .001 by the study variables. Problem gambling status could be predicted from a tendency to click on gambling related t(880) = 7.589, p < .001 and counselling related links t(880) = 3.998, p < .001. Complaints about spam t(880) = 3.743, p < .001 and attempts to unsubscribe from mailing lists t(880) = 3.276, p < .001 were also predictors of problem gambling status.

Table 3 Electronic behaviours predicting risk of gambling problems (Internet surfing, complaints)
Full size table

Discussion

Although jurisdictions seek to control gambling by restricting access, digital services provide a number of technological loopholes that may sidestep existing controls (Delfabbro 2008; Eadington 2004). As people may have problems controlling their online behaviours, the present paper considered whether electronic interests and behaviours could predict the risk of developing a gambling problem. The Internet elicited the higher levels of concern related to self control. Although the majority of participants expressed concerns over their Internet use, and were dissatisfied with current protective measures, these concerns and dissatisfactions were not specifically predictive of gambling problems. Instead, a greater risk of a gambling problem was associated with an interest in electronic services, and problems limiting one’s television and mobile phone use. Clicking on gambling or counselling related links, and an increased level of problems with spam also predicted gambling problems.

Problem Technology Use

The majority of the sample reported problems with the internet, but reports of problems controlling one’s television and mobile phone use were associated with gambling problems. The mobile phone potentially offers a greater (and potentially unsupervised) access of the individual to gambling related services 24 h a day, 7 days a week (Griffiths 2007b; King et al. 2010). It is less clear why the television might be a problem, but a number of interactive services utilising the TV have been developed to target viewers with poor impulse control who call or SMS in at premium rates to win prizes (Griffiths 2007a), and the present data could be viewed as an indicator of the potential harm this is causing.

There have of course been suggestions that there is an addictive personality (Jacobs 1988) that is prone to a variety of impulsive, self-stimulatory or escapist behaviours. However, gamblers are a heterogeneous population, and a predisposition will by no means account for the diverse ways in which a gambling problem actually manifests (Shaffer et al. 2004b). Rather than demonstrating that interest in technology in general was associated with a greater risk of a gambling problem, the present study demonstrated that participants with an interest in specific technologies (i.e. TV and mobile phones) were more likely to be at risk.

An interest in digital services may lead people to access online gambling (King et al. 2010). The present study utilised both questionnaire items and also tracked an online behaviour (see LaBrie et al. 2008; Wood and Griffiths 2007) to determine predictors of online gambling. A number of online behaviours were found to be predictive of gambling problems. Participants who clicked on either gambling or counselling related links while surfing, were more likely to report a gambling problem.

Spam

Increased problems with malware such as spam were related to gambling problems. This may be because poorer impulse control heightens vulnerability, or because dubious operators seek this cohort out to recruit their computers for use in cyberextortion bids (McMullan and Rege 2007).

As problem gamblers reported receiving and unsubscribing from spam more often, it is likely that at risk individuals access more dubious propositions online, such that their contact details and e-mail addresses (and possibly those of their peers) are added to more junk mailing lists (Phillips et al. 2011). Indeed, Wood et al. (2007) reported that 11.4% of their sample of participants first started to play poker online after receiving a spam e-mail or pop-up link to do so. This is of concern, as some mailing lists may actually involve funds transfer such that appreciable costs are incurred when a message is received (i.e. premium SMS services), with recipients losing money whenever they receive such messages. Indeed, by accessing online gambling sites (which may be legitimate or illegitimate) people are identifying themselves as having a vulnerability to further dubious online propositions given the comorbidity between impulse control disorders and problem gambling (Grant and Kim 2003).

Limitations

The present study used the Internet to collect data on online risk factors and behaviours that might be associated with gambling problems. Although there may be concerns that an Internet sample may be biased, Gosling et al. (2004) have demonstrated that Internet surveys are more diverse and representative over a range of key demographic variables such as gender, socioeconomic status, location and age, and the findings from such surveys are comparable with those acquired by traditional methods. Otherwise an internet survey is more likely to sample the online population that is actually at risk of online gambling problems.

The present study predominantly used self-reports, and as such can be limited by people’s insight into their behaviour, and the accuracy of memory. However, the providers of gambling services have better information as to gambling behaviour as they can track user behaviour. Online gambling transactions have been analysed (see Davies 2007; LaBrie et al. 2008), but there may be problems gaining access to such information (Griffiths and Whitty 2010). Indeed, consumer tracking may suffer from appreciable amounts of missing data if gamblers flit from provider to provider seeking a better deal (Phillips and Ogeil 2011). In addition, while some behaviours indicative of the DSM-IV criteria for problem gambling can be tracked (e.g., chasing losses, preoccupation with gambling) others (e.g. the experience of withdrawal, escape and relationship problems) are not likely to be detected by such technology (Griffiths and Whitty 2010). Nevertheless, the present study endeavoured to bridge the gap between self report and consumer tracking by including behavioural measures (internet surfing), and these actually linked specific website selections to increased risk of gambling problems. The present study demonstrated that people at risk of a gambling problem were more likely to click onto gambling or counselling related links. Indeed, these specific behaviours were amongst the better predictors of gambling problems in our online sample.

The present paper demonstrated an association between interest in technology and gambling problems. However correlations between interest in online services and gambling problems are not necessarily causal in nature. Online services may not cause gambling problems, instead gamblers or people with poor impulse control may gravitate to this medium because of the greater accessibility or better perceived value of online gambling (Griffiths 2003).

In the present study the scales addressing problems controlling technology use were quite reliable and correlated well with levels of technology use, but the scales assessing consumer interest and dissatisfaction had lower reliability. Reliability can be a function of scale length (DeVellis 1991), with longer scales being more accurate. Future work should elaborate upon the questions used to provide better instruments.

Although the data was obtained on an Australian sample, the potential for the Internet to put gambling provided by offshore operators within the reach of any consumers with access to electronic cash is a worldwide problem (Scoolidge 2006). There are thus parallels between Australia and other jurisdictions seeking to curb some forms of online gambling.

Conclusion

Given the increasing accessibility of gambling, the present paper considered potential interests and online behaviours that might be associated with an increased risk of gambling problems. Generalised impulse control problems across a variety of technologies did not seem to be linked with gambling problems. Instead it appears that at risk individuals were interested in specific technologies (TV, Mobile phones) for the purposes of gambling and this is of concern given the increased potential for gambling online. This increasing access to technology that supports gambling (e.g., via Internet or mobile phone) may also result in increased exposure to harmful messages and software (e.g. spam, scams). Such schemes are concerning given that technology has outstripped control measures.