Introduction

Problem gambling has been identified as a significant public health concern (Gainsbury et al. 2014). It is characterized by “difficulties in limiting money and/or time spent on gambling which leads to adverse consequences for the gambler, others, or for the community” (Delfabbro 2008, p. 2). In Australia, poker-machines currently generate the most gaming revenue and are evidently the most damaging form of gambling (e.g., Dickerson 2004; Petry 2003). Moreover, three quarters of people who present a serious gambling problem are poker-machine gamblers (Australian Government 2014). Thus, given the disproportionate contribution of poker-machines to problem gambling, it is necessary to investigate the psychological mechanisms that contribute to the maintenance of problematic poker-machine gambling.

Problem Gambling and Cue-Reactive Urge

The cue-reactivity model suggests that gambling cues trigger a particularly strong urge to gamble in problem gamblers (Baudinet and Blaszczynski 2013). Urge may be defined as a state whereby individuals are motivated to seek out particular objects or feelings (Rock and Kambouropoulos 2008; Rohsenow and Monti 1999). In most cue-reactivity studies (e.g., Tricker et al. 2015) urge has been quantified using a unidimensional visual analogue scale (VAS) with participants asked to rate their response on a scale from 0 (no urge) to 10 (extreme urge) to questions such as “How much do you feel like gambling in this present moment?”.

The notion of cue-reactivity is grounded within behavioral conceptualizations of classical and operant conditioning. In terms of classical conditioning, previously neutral stimuli in the gambler’s environment (e.g., poker-machine sounds or lights) elicit behavioral responses (e.g., urge or excitement) following repeated pairings with a gambling-related consequence such as winning money (Franco 2008). The repeated pairing of the previously neutral stimulus with the gambling consequence leads to the continuation of the gambling behavior, even in the absence of money. Moreover, it is proposed that operant conditioning further facilitates this process. Through the intermittent reinforcement of wins delivered on a variable ratio, both heightened arousal states and/or urge is triggered (Baudinet and Blaszczynski 2013). It is proposed that gambling behavior is further reinforced by problem gamblers’ cognitive distortions (e.g., irrational or superstitious beliefs; Joukhador et al. 2003).

It is, therefore, not surprising that some gambling therapies have adopted a cognitive-behavioral cue-exposure framework (e.g., Oakes et al. 2008). Exposure therapy aims to extinguish the urge to gamble through graded cue-exposure and response prevention (Riley et al. 2011). However, despite the existence of exposure therapies, only three published studies (Ashrafioun et al. 2012; Sodano and Wulfert 2010; Wulfert et al. 2009) have examined the relationship between problem-gambling severity and cue-reactive urge to gamble, while only one study (Tricker et al. 2015) has focused specifically on poker-machine problem gambling.

Wulfert et al. (2009) examined a sample of 94 experienced gamblers. Gamblers were presented with four 2-min video clips of their preferred and non-preferred modes of gambling and two control stimuli. Urge to gamble (measured from 1 to 100) was assessed following each cue presentation. Results indicated that urge to gamble was highest following exposure to cues of gamblers’ preferred gambling activity. Additionally, problem gamblers reported greater urge to gamble than social gamblers. Sodano and Wulfert (2010) implemented a similar design in their study of 63 gamblers and found similar results. Importantly, neither study tested for an increase in urge from a neutral cue to a gambling cue, while controlling for baseline urge. In the two aforementioned studies, it is, of course, possible that the higher urge to gamble reported by problem gamblers was a result of, at least in part, individual differences in baseline urge.

Ashrafioun et al. (2012) examined a sample of 48 university student gamblers. Participants were exposed to either an imagery script or a gambling-related photos condition. Gambling urge was assessed at baseline (i.e., prior to gambling cue presentation) and post-test (i.e., following gambling cue presentation). Results indicated that urge to gamble increased for participants in both conditions. Moreover, a significant positive association was found between problem-gambling severity and gambling urge. However, the absence of a control condition (i.e., a neutral cue) precludes the effects of the treatment from being isolated.

The experimental design implemented by Tricker et al. (2015) allows for more conclusive findings. Thirty-seven poker-machine gamblers completed an online repeated measures survey. Urge to gamble (measured by a VAS) was assessed at baseline, following a neutral stimulus (i.e., 3-min video of women making tea) and following a gambling stimulus (i.e., 3-min video of the same women playing poker-machines). After controlling for baseline urge to gamble, problem-gambling severity (measured by the Problem Gambling Severity Index, PGSI; Ferris and Wynne 2001) significantly predicted an increase in urge from neutral cue to gambling cue.

Problem Gambling, Altered State of Awareness and Cue-Reactive Urge

A variable that may contribute to an understanding of gambling urge is altered state of awareness (ASA). We note that some problem gamblers have described feelings of losing volitional control over their gambling behaviors during play (McCormick and Delfabbro 2011). Moreover, the South Australian Department of Families and Communities (2007) found that 39 % of problem gamblers often, or very often, lost track of reality while gambling. This subjective experience may be conceptualized as an ASA, which refers to a subjective feeling of an unusual state of consciousness relative to one’s ordinary waking state (Pekala 1991). The ASA construct has been investigated in a number of alcohol cue-reactivity studies (Kambouropoulos and Rock 2009, 2010; Rock and Kambouropoulos 2009, 2012) and in two poker-machine gambling studies (McCormick and Delfabbro 2011; Tricker et al. 2015).

In a pilot study, McCormick and Delfabbro (2011) investigated the relationship between problem-gambling severity and ASA during poker-machine play. Participants completed two conditions: one baseline (i.e., sitting quietly for 3 min) and one gambling condition (i.e., playing poker-machines for 20 min). Following each condition, participants completed a survey that included the ASA dimension of the Phenomenology of Consciousness Inventory (PCI-ASA; Pekala 1991). However, there was no neutral condition and due to the small sample size (N = 14), the authors did not undertake null hypothesis significance testing.

In a subsequent study, Tricker et al. (2015) tested experimentally whether ASA (measured by the PCI-ASA; Pekala 1991) mediated the relationship between problem-gambling severity (measured by the PGSI; Ferris and Wynne 2001) and cue-reactive urge to gamble (measured by a VAS). Measurements of cue-reactive ASA and cue-reactive urge to gamble were taken at three temporal points: baseline, neutral cue (i.e., a 3-min video of women making tea) and gambling cue (i.e., a 3-min video of the same women playing poker-machines). Extrapolating from Rock and Kambouropoulos’ (2009) alcohol cue-reactivity study, Tricker et al. hypothesized that, relative to neutral cue-exposure, gambling cue-exposure may cause the gambler to re-experience, rather than merely remember, the ASAs associated with previous poker-machine gambling and this may, in turn, promote urge to gamble. Results supported this hypothesis and a significant indirect effect was found from problem-gambling severity to cue-reactive urge through cue-reactive ASA.

Problem Gambling, Trait Mindfulness and Cue-Reactive Urge

Trait mindfulness is a construct that may contribute to individual differences in problem-gambling behaviors and the urge to gamble experienced in response to gambling cues. Mindfulness may be defined as “paying attention in a particular way: on purpose, in the present moment, and non judgmentally” (Kabat-Zinn 1994, p. 4). Trait mindfulness is conceptualized as an adaptive dispositional attribute reflected by present-focused attention and self-regulation that fluctuates naturally in individuals (Weinstein et al. 2009). High levels of trait mindfulness involve a propensity to attend to present moment experience as a means of becoming aware of one’s automatic reactions and, thus, remain nonreactive when confronted by distressing thoughts or emotions (Garland 2011). Indeed, higher levels of trait mindfulness have been found to be associated with individuals engaging in adaptive emotion regulation strategies and lower reported symptoms of psychological distress (e.g., Bullis et al. 2014). It might be argued that individuals high on trait mindfulness have the propensity to more effectively regulate their reactions (e.g., cognitions, emotions, behaviors) to gambling cues (i.e., ‘surfing’ the urge to gamble; Shonin et al. 2013).

Although limited research exists examining trait mindfulness in a cue-reactivity context, Garland (2011) examined the relationship between trait mindfulness and cue-reactive urge to drink alcohol. Fifty-eight alcohol-dependent adults were exposed to a number of alcohol-related photos. The cue-reactivity protocol consisted of four phases (i.e., resting baseline, stress cue-exposure, alcohol cue-exposure, recovery period). Results indicated that higher trait mindfulness (measured by the Five Facet Mindfulness Questionnaire, FFMQ; Baer et al. 2006) was associated with less subjective difficulty resisting alcohol urges (measured by a VAS).

Interestingly, whilst mindfulness-based therapies are beginning to be introduced as an adjunct to traditional cognitive-behavioral therapy to address problem gambling (e.g., Toneatto et al. 2014), very little experimental evidence exists regarding the links between problem-gambling severity, trait mindfulness and urge to gamble. To the best of our knowledge, only three published studies (de Lisleet al. 2014; Lakey et al. 2007; Riley 2012) have examined these associations.

In a non-randomized study, Lakey et al. (2007) examined a convenience sample of 185 university student gamblers. A significant negative association was found between trait mindfulness (measured by the Mindfulness Attention Awareness Scale, MAAS; Brown and Ryan 2003) and problem-gambling severity (measured by the Diagnostic Interview for Gambling Severity, DIGS; Winters et al. 2002). Subsequently, Riley (2012) examined a convenience sample of 103 treatment seeking problem gamblers and found a consistent association. Taken together, these results suggest that higher trait mindfulness generates a level of present centered awareness of gambling thoughts and feelings allowing for more adaptive behavioral choices (Lakey et al. 2007).

In one other non-randomized study, de Lisle et al. (2014) investigated a sample of 78 self-identified problem gamblers. A significant (negative) association was found between trait mindfulness (measured by the MAAS) and problem-gambling severity (measured by the PGSI). Additionally, trait mindfulness was significantly (negatively) associated with gambling urge (measured by the Gambling Urges Scale, GUS; Raylu and Oei 2004).

Importantly, the aforementioned studies (e.g., de Lisle et al. 2014) highlight that trait mindfulness is negatively associated with problem-gambling severity and urge to gamble. However, no studies have examined these relationships using an experimental design or specifically in relation to poker-machine gambling. In addition, perhaps trait mindfulness moderates the effect of problem-gambling severity with regards to cue-reactive urge to gamble such that the coupling of high trait mindfulness/low problem-gambling severity would be associated with lower cue-reactive urge to gamble (from neutral cue to gambling cue). In other words, perhaps individuals with high levels of trait mindfulness and low levels of problem-gambling severity will experience less cue-reactive urge to gamble relative to their low trait mindfulness and high problem-gambling severity counterparts.

The Current Study

The current study aimed to extend the results of Tricker et al. (2015) who found that cue-reactive ASA mediated the relationship between problem-gambling severity and cue-reactive urge to gamble (from neutral cue to gambling cue). Additionally, this study aimed to investigate the associations between trait mindfulness, problem-gambling severity and cue-reactive urge to gamble, while controlling for baseline urge to gamble. Finally, this study investigated whether trait mindfulness moderated the effect of problem-gambling severity on urge to gamble (from neutral cue to gambling cue).

First, in order to test for a cue-reactivity effect the following hypothesis was formulated:

H1

Urge to gamble will increase from neutral cue to gambling cue, while controlling for baseline urge to gamble.

Second, in order to test Tricker et al.’s (2015) mediation model the following hypothesis was formulated:

H2

Cue-reactive ASA will mediate the relationship between problem-gambling severity and cue-reactive urge to gamble.

Third, as previously stated, three studies (de Lisle et al. 2014; Lakey et al. 2007; Riley 2012) found significant negative associations between trait mindfulness and problem-gambling severity, and one study (de Lisle et al. 2014) found a significant negative association between trait mindfulness and urge to gamble. Interestingly, no published studies have examined this link in the context of poker-machine gambling or tested this relationship using an experimental design. Therefore, we formulated the following hypotheses:

H3

Trait mindfulness will negatively correlate with problem-gambling severity; and

H4

Trait mindfulness will negatively correlate with cue-reactive urge to gambling, while controlling for baseline urge to gamble.

Finally, given our previously stated contention that perhaps individuals with high levels of trait mindfulness coupled with low levels of problem-gambling severity will experience less cue-reactive urge to gamble compared to individuals with low trait mindfulness coupled with high problem-gambling severity, we formulated the following hypothesis:

H5

Trait mindfulness will moderate the effect of problem-gambling severity on urge to gamble.

Method

Participants

Thirty-eight poker-machine gamblers (29 male and nine female) completed the present study. Following the precedent of McCormick et al. (2012) and MacClaren et al. (2012) the selection criteria were: (a) having played poker-machines at least twice a month in the last 12 months; and (b) poker-machines were the person’s preferred mode of gambling. Participants were recruited via advertisements placed on Facebook, Reddit, online gambling forums and two Sydney newspapers (i.e., Fairfield City Champion and Liverpool City Champion). A total of 188 participants commenced the present study, reading at least the study information sheet. Consistent with the findings of Tricker et al. (2015), a significant number of those individuals did not complete the study and their data was not included in the analysis. The sample of 38, therefore, reflects a completion rate of 20 %. Various characteristics of the sample are displayed in Table 1.

Table 1 Participant characteristics
Full size table

Materials and Measures

The online study was created using Qualtrics™ software (version 9340538 of the Qualtrics Research Suite, copyright 2015; Qualtrics, Provo, UT, USA). The online platform consisted of an information page, consent form, demographics questionnaire, debriefing page, two 3-min video cues (described below) and the assessment measures (described below).

Video Cues

Both video cues were 3-min in length and identical to those used by Tricker et al. (2015) in their poker-machine gambling study. The 3-min cue presentation time is consistent with numerous alcohol cue-reactivity studies (e.g., Kambouropoulos and Rock 2010; Rock and Kambouropoulos 2009, 2012). The neutral cue consisted of video and audio of two women preparing a cup of tea and the gambling cue consisted of video and audio of the same two women playing poker-machines in the gaming room of a hotel (i.e., pub). As manipulation checks, page timers were placed on each video-cue page to ensure clips were watched in their entirety. Participants could not proceed until 3-min (i.e., the length of the video cue) had expired.

Problem-Gambling Severity

Problem-gambling severity was measured using the Problem Gambling Severity Index (PGSI; Ferris and Wynne 2001). The PGSI was constructed to measure problem gambling in the general population (Holtgraves 2009). The PGSI consists of nine items, four of which assess problem-gambling behaviors (e.g., “How often have you bet more than you could afford to lose”) and five that assess adverse consequences of gambling (e.g., “How often has your gambling caused any financial problems for you or your household?”). Items are rated on a four-point Likert scale (0 = never, 1 = sometimes, 2 = most of the time, 3 = almost always). Scores are determined by summing all item scores. In terms of total scores on the PGSI, 0 = no problems; 1–2 = low level of problems; 3–7 = moderate level of problems; and 8–27 = problem gambling (Ferris and Wynne 2001). The PGSI has demonstrated good internal consistency (Cronbach’s alpha = .84) and test–retest reliability (r = .78) (Ferris and Wynne 2001). Additionally, the PGSI has displayed good criterion validity (r = .83) with the South Oaks Gambling Screen and Diagnostic and Statistical Manual of Mental Disorders-IV classifications of problem gambling (Ferris and Wynne 2001). The Cronbach’s alpha for this sample was .95.

Trait Mindfulness

Trait mindfulness was measured using the MASS (Brown and Ryan 2003). The MAAS measures the tendency to be mindful of moment-to-moment experience (Brown and Ryan 2003). The MAAS measures 15 items on a six-point Likert scale (1 = almost always; 6 = almost never). Scores are determined by calculating the mean of all items where a higher averaged total score indicates greater mindfulness. Respondents indicate how frequently they have the experience described in each statement (e.g., “I find it difficult to stay focused on what’s happening in the present”). The MAAS has demonstrated construct validity both within community (MacKillop and Anderson 2007) and clinical samples (Carlson and Brown 2005). Additionally, the MAAS has been found to have strong internal consistency (Cronbach’s α = .87) and test–retest reliability (r = .81) in adult samples (Brown and Ryan 2003). In the present study, the MAAS yielded a Cronbach’s alpha of .93.

Altered State of Awareness

Altered state of awareness was measured using the ASA dimension of the PCI (Pekala 1991). The PCI is a retrospective 53-item measure designed to assess 12 major dimensions of phenomenology (e.g., volitional control, arousal, positive affect, ASA) and 14 minor dimensions (e.g., fear, joy, time sense). The ASA dimension consists of three pairs of questions (e.g., “My state of awareness was very different from what I usually experience,” versus “My state of awareness was no different than usual”) rated on a seven-point Likert scale where 0 indicates no or little increased intensity and 6 indicates much or complete intensity (Pekala and Wenger 1983). Scores are computed by calculating the mean of the three PCI-ASA items. The PCI has demonstrated good internal consistency and criterion validity with Cronbach’s alphas ranging from .7 to .9 (Pekala et al. 1986). Furthermore, it has demonstrated criterion validity by reliably discriminating between qualitatively different states of phenomenology (Pekala et al. 1986).

Urge to Gamble

Gambling urge was measured using a VAS. The VAS presented as a horizontal line. Participants were asked, “In the present moment, how strong is your urge to play poker machines?” anchored on the left by no urge (scored as 0) and extreme urge (scored as 10). VASs have been found to be an effective and reliable method to assess urge in cue-reactivity studies (Kambouropoulos and Rock 2010).

Design

The current study employed the repeated measures experimental design of Tricker et al. (2015) and, thus, consisted of three phases: (1) baseline; (2) neutral cue presentation; and (3) gambling cue presentation (see also Rock and Kambouropoulos 2007). Video cues were presented in the same order for all participants. Importantly, presenting the experimental cue (i.e., gambling cue) after the neutral cue provides the most conservative estimation of cue-reactivity and is a standard aspect of cue-reactivity protocols (see Monti et al. 1987). That is, it is asserted that if the gambling cue was presented first, then any cue-reactivity could carry-over to the neutral cue (Rohsenow and Niaura 1999).

Procedure

After ethics approval was granted for the present study, advertisements were distributed via Facebook, Reddit, online gambling forums and two Sydney newspapers. Advertisements contained information about the study and a link to the online platform. The online platform presented participants with the welcome page, the information page, the consent page, the demographics questionnaire, the PGSI and the MAAS. Prior to cue presentation, participants were instructed to turn their audio on. To assess baseline urge to gamble, participants were presented with the VAS followed by the PCI-ASA items to assess baseline ASA. Subsequently, the neutral-cue video was presented followed by the VAS (to assess current urge to gamble) and the PCI-ASA (to assess retrospectively ASA during the video). Next, the gambling-cue video was presented followed by the VAS (to assess current urge to gamble) and the PCI-ASA (to retrospectively assess ASA during the video). Finally, participants were presented with the debriefing page.

Results

Hypothesis 1

To establish whether urge to gamble increased from neutral cue to gambling cue exposure (while controlling for baseline urge), a repeated measures analysis of covariance (ANCOVA) was performed. Baseline urge to gamble (M = 2.42, SD = 2.84) was entered as a covariate in the analysis. The analysis consisted of one within-subjects variable, cue, with two levels (i.e., neutral and gambling). Changes across cue condition on urge to gamble were examined. The ANCOVA indicated that, after accounting for baseline urge, there was a statistically significant main effect of cue, F(1, 36) = 10.16, p = .003, partial η2 = .220. Specifically, urge to gamble increased statistically significantly from neutral cue (M = 1.76, SD = 2.27) to gambling cue (M = 3.76, SD = 3.36). The hypothesis was supported.

Hypothesis 2

To establish whether cue-reactive ASA (from neutral cue to gambling cue) mediated the relationship between problem-gambling severity and cue-reactive urge (from neutral cue to gambling cue), a mediation analysis was undertaken. The analysis used Preacher and Hayes’ (2004) bootstrapping method and was performed using model 4 of Hayes’ (2013) PROCESS macro for SPSS (version 2.041). The current study used 5000 bootstrap iterations in accordance with Hayes’ (2013) recommendation. Problem-gambling severity (PGSI score) was entered as the predictor variable, urge-change (gambling cue VAS score minus neutral cue VAS score) as the criterion variable and ASA (gambling cue PCI-ASA score minus neutral cue PCI-ASA score) as the proposed mediator. The results of the mediation analysis are presented in Fig. 1.

Fig. 1
figure 1

Mediation model showing non-significant mediation. Solid lines represent significant pathways, dotted line represents non-significant pathway, b = unstandardised regression coefficient, β = the completely standardised regression coefficient of the indirect effect (the index of mediation), CI = bias-corrected and accelerated bootstrapped confidence-intervals based on 5000 samples

Full size image

As shown in Fig. 1, path ‘a’ was not statistically significant (b = .06, p = .180) and, thus, cue-reactive ASA was not a statistically significant mediator [indirect effect: b = .004, 99 % CI (−.04, .16)]. Moreover, variations in levels of the independent variable did not statistically significantly account for variations in the presumed mediator (see Baron and Kenny 1986). The hypothesis was not supported.

Hypothesis 3 and 4

The bivariate correlation between trait mindfulness and problem-gambling severity was statistically significant, r(37) = −624, p = .000. Similarly, the partial correlation between trait mindfulness and cue-reactive urge to gamble (while controlling for baseline urge to gamble) was statistically significant, r(37) = −463, p = .004. The hypotheses were supported.

Hypothesis 5

To establish whether trait mindfulness moderated the effect of problem-gambling severity on urge to gamble (from neutral cue to gambling cue) a hierarchical multiple regression was performed. The criterion variable was urge-change (gambling cue VAS score minus neutral cue VAS score). Problem-gambling severity (PGSI score) was included at step 1, followed by the interaction term of problem-gambling severity and trait mindfulness (MAAS score) at step 2. At step 1, problem-gambling severity was a statistically significant predictor of urge to gamble (b = .18, p = .002). At step two, the interaction between problem-gambling severity and trait mindfulness with respect to their effect on urge to gamble (from neutral cue to gambling cue) was not statistically significant (b = −.08, p = .075; see Table 2). The hypothesis was not supported.

Table 2 Moderator analysis displaying problem gambling × trait mindfulness interaction predicting cue-reactive urge (from neutral cue to gambling cue)
Full size table

Discussion

The present study aimed to examine trait mindfulness, cue-reactive altered state of awareness and cue-reactive urge to gamble in poker-machine gamblers. Poker-machine gamblers reported greater subjective urge to gamble in response to exposure to a poker-machine cue than a neutral cue (while controlling for baseline urge), thus, supporting H1. This finding is consistent with Kushner et al.’s (2007) proposal that cue reactive urge to gamble amongst problem gamblers can be understood in terms of classical conditioning theory, whereby stimuli which were once neutral become paired with gambling activity. This pairing means that stimuli associated with gambling (e.g., flashing lights on poker machines), rather than gambling behaviour itself, come to elicit similar cognitive, affective and physiological reactions (including urges) to those experienced during actual gambling behaviour. This result is noteworthy given that, to date, only one previous study (Tricker et al. 2015) has employed the cue reactivity protocol to examine urge to gamble in a sample of poker-machine players.

The present study did not find that ASA mediated the relationship between problem-gambling severity and cue-reactive urge to gamble (H2). More specifically, problem-gambling severity acted directly on cue-reactive urge; that is, the direct (i.e., ‘c’) pathway was significant. In addition, whereas path ‘b’ (cue-reactive ASA → cue-reactive urge) was significant, we note that path ‘a’ (problem-gambling severity → cue-reactive ASA) was not significant and, thus, an indirect effect was not found and the mediation model was not significant. In addition, variations in levels of problem-gambling severity did not significantly account for variations in the presumed mediator, ASA. The significant ‘b’ pathway supports the findings of Tricker et al. (2015); however, the significant direct effect and non-significant ‘a’ pathway is inconsistent with the results of Tricker et al. Given that the mean baseline, neutral and cue-reactive ASA scores in the present study were 1.35, 1.30, and 1.66, respectively, (theoretical range = 0–6; 0 = “none or little”, whereas a score of 6 = “much or complete”), it seems reasonable to conclude that the gambling cue did not elicit an ASA effect. We note that McCormick and Delfabbro (2011) concluded that, even in their in situ study, 20-min poker-machine playing sessions were not sufficient to allow variations in conscious awareness to emerge. However, we emphasize that the present study used the stimulus conditions developed and administered by Tricker et al. One might speculate that, compared to the present study’s sample, Tricker et al.’s sample consisted of a higher proportion of individuals endowed with trait-like tendencies to experience ASAs, e.g., trait absorption, which may be defined as “a disposition for having episodes of “total” attention that fully engage one’s representational (i.e., perceptual, enactive, imaginative, and ideational) resources” (Tellegen and Atkinson 1974, p. 268). Broadly speaking, the trait absorption construct may be interpreted as situated engrossment in a particular stimulus. Perhaps, not surprising then, absorption ability has been significantly positively correlated with receptivity to ASAs associated with meditation practices (e.g., Pekala et al. 1985; Weinstein and Smith 1992). Consequently, one might hypothesize that the ability to become ‘absorbed’ in a poker-machine stimulus and, thus, eliminate ‘noise’ due to extraneous thought impressions and bodily perceptions may enhance phenomenological responsiveness (e.g., ASAs) to that stimulus condition. It might be the case that only a sub-set of the poker-machine problem gambling population are high ‘absorbers’ and it is that sub-set who are more likely to experience cue-reactive ASAs. Thus, future poker-machine gambling studies might benefit from quantifying participants’ absorption ability, rather than simply quantifying problem-gambling severity, in the context of testing mediation models with cue-reactive ASA as the proposed mediator and cue-reactive urge to gamble as the dependent variable.

The present study found that trait mindfulness was negatively correlated with problem-gambling severity, thus, supporting H3. This finding is consistent with a small but growing body of literature (e.g., de Lisle et al. 2014; Lakey et al. 2007; Riley 2012). More notably, trait mindfulness was negatively correlated with cue-reactive urge to gamble, while controlling for baseline urge, thereby, supporting H4. Whilst previous findings have suggested that urge to gamble is negatively associated with trait mindfulness (de Lisle et al. 2014) this is the first study, to our knowledge, which has demonstrated this relationship with cue-reactive urge to gamble and it is also the first to demonstrate this specifically in a population of poker-machine gamblers. This result suggests that high levels of trait mindfulness may be adaptive when individuals with problem-gambling behavior are confronted with gambling cues (e.g., poker-machines, adverts for gambling events/locations), allowing them to be better equipped to notice and accept the subjective urge to gamble and, in so doing, allow this urge to subside (i.e., ‘surfing’ this urge; Shonin et al. 2013). It has been argued that a higher level of dispositional mindfulness is beneficial in that it allows greater present-centered awareness and affords individuals the opportunity to engage in an adaptive behavioral response (i.e., not consistent with problematic gambling behaviors; de Lisle et al. 2014; Lakey et al. 2007). Consistent with this argument, trait mindfulness has also been found to be negatively associated with impulsivity in the general population (Rajesh et al. 2013). Higher levels of trait mindfulness may, therefore, be hypothesized to attenuate both cue-reactive urge to gamble and impulsive gambling behavior through higher levels of subjective awareness of such processes and lower levels of impulsivity driving gambling behaviors. Given that trait mindfulness is considered to be plastic and modifiable through training (e.g., Garland 2011), the results of the present study further strengthen the rationale for investigating the potential utility of mindfulness-based interventions in addressing problem-gambling behaviors.

Finally, the present study did not find that trait mindfulness moderated the effect of problem-gambling severity on cue-reactive urge (H5). This non-significant result may be explained, at least in part, by a high level of covariance between problem-gambling severity and trait mindfulness (i.e., individuals with high levels of problem-gambling severity will display low trait mindfulness, and vice versa), thus, attenuating any moderating effect.

Methodological Limitations and Future Directions

The use of an online cue-reactivity platform provides a convenient and easily accessible means to test for gambling urge and may provide a level of anonymity in participation that suits the often-covert nature of poker-machine playing. The present study’s sample consisted of a group of individuals aged between 19 and 48, whose self-identified preferred mode of gambling was poker-machines. The sample, whilst small, reported a level of gambling severity that is comparable to that reported among Australian poker-machine gamblers who gamble weekly (Productivity Commission 2010) and also with Tricker et al.’s (2015) sample. However, the sample consisted of volunteers who chose to participate as a result of seeing an online study invitation and it cannot be assumed that these participants were representative of the wider poker-machine gambling community. The high rate of drop-out following commencement may also be seen to be problematic but not unexpected given findings from previous online gambling research studies (Tricker et al. 2015) and the high levels of impulsivity associated with this population (MacClaren et al. 2012). It should be noted that the sample included individuals from outside Australia and the methodology did not account for any potential cultural differences in gambling behaviors or cues. Research findings associated with an online platform do not necessarily generalize to the real world and the representativeness of online samples may be difficult to assess (e.g., Gosling et al. 2004). Relatedly, the online nature of the present study necessitated the use of a gambling cue, which may be argued to lack ecological validity. Indeed, there is evidence that simulated gambling scenarios may fail to elicit the same cognitive-affective responses that would be expected in a real-world scenario (e.g., McCormick and Delfabbro 2011) and this may partially account for the present study’s non-significant findings in relation to ASA scores, which must, therefore, be interpreted with caution. Nevertheless, if it is accepted that such an online stimulus would act as a less potent stimulus than those encountered in real-world settings, then the present study’s significant findings regarding trait mindfulness and cue-reactive urge to gamble may be argued to be all the more notable.

The MAAS (Brown and Ryan 2003) was employed in the present study to assess trait mindfulness. This measure was selected due to the fact that it is the most widely utilized measure of trait mindfulness in the problem gambling literature (e.g., de Lisle et al. 2014) and because it is considered a valid and reliable measure of individual attention to, and awareness of, present occurrences (MacKillop and Anderson 2007). Nevertheless, it is important to note that, while this measure conceptualizes trait mindfulness as a unidimensional construct, a number of authors (e.g., Sauer et al. 2013) have argued that trait mindfulness is a multidimensional construct, which incorporates factors such as acceptance and emotion regulation in addition to attentional awareness. Therefore, future studies might seek to incorporate a more multi-faceted trait mindfulness measure such as the FFMQ (Baer et al. 2006).

An additional limitation of the present study may be that the sample had a mean PGSI score of 6.15, a score reflective of “moderate” levels of problem gambling, which is below the higher range of “problem gambling” suggested by Ferris and Wynne (2001). Thus, it may be the case that our sample did not reflect adequately the responses of individuals with higher levels of problem-gambling severity. However, we note that Tricker et al.’s (2015) sample had a mean PGSI score of 4.6.

Finally, given that gambling research has focused generally on gamblers as a homogenous group (e.g., Riley 2012), future research might examine gambling cue-reactivity in relation to specific gambling modalities. For example, we note that in the present study’s sample of poker-machine gamblers, participants reported greater cue-reactive urge to gamble relative to a neutral cue (while controlling for baseline urge) when exposed to a cue adhering to their gambling mode preference. Thus, we recommend that future research administer gambling cues consistent with the gambling modality under investigation. In addition, it may be interesting to examine whether trait mindfulness and cue-reactive altered state of awareness are constructs pertinent to other gambling modalities.

Conclusion

In Australia, three-quarters of those gamblers who present a serious gambling problem are poker-machine gamblers (Australian Government 2014). Interestingly, while we found a significant cue-reactive urge to gamble effect, we did not find that problem-gambling severity predicted cue-reactive altered state of awareness. In addition, we did find that trait mindfulness was significantly negatively associated with both problem-gambling severity and cue-reactive urge to gamble. However, trait mindfulness did not significantly interact with problem-gambling severity with regards to predicting urge to gamble. Replication of the present study using larger samples and extensions of the present study’s design using real-world gambling environments is required to further explore these constructs.

Importantly, the present study was the first to demonstrate a negative association between trait mindfulness and cue-reactive urge to gamble in a population of poker-machine gamblers. Thus, it is hoped that this finding may catalyze a line of trait-mindfulness-focused cue-reactivity research inquiry that may, ultimately, contribute to the improvement of both exposure therapies and public policy in what is a highly contentious and potentially destructive industry.