Addiction (2001) 96, 1629–1638

RESEARCH REPORT

Induced mood and persistence at gaming
ADELMA M. HILLS, SANDI HILL, NATALIE MAMONE & MARK DICKERSON
University of Western Sydney, New South Wales, Australia
Abstract Aims. Using the affect infusion model (AIM), this study aimed to assess the impact of prior mood on gaming persistence in regular and non-regular gamblers (assumed to be motivated and heuristic decision-makers, respectively). Design, measures and setting. A 2 3 3 experimental design in a laboratory setting employed factors of gambler type (regular, non-regular) and prior mood (happy, neutral, depressed). Measures were number of trials played on a gambling game, and mood ratings (post-mood induction, during and after play). Participants and intervention. Sixty regular and 60 non-regular gamblers (all male students) were assigned randomly to watch a video inducing one of the three mood conditions. They were then given AUD$10.00 to gamble in a computerized card-cutting game. Findings. AIM predictions were supported: prior mood did not affect gaming persistence for regular gamblers, but non-regulars showed reduced persistence across happy, neutral and depressed moods. After-play mood ratings were related to winnings for regulars, and losing regulars were signi cantly more unhappy after-play compared to during-play, and compared to both their winning counterparts and non-regulars. Mood and winnings were unrelated for non-regulars, with little during- to after-play mood change. Conclusions. The ndings speak to depression as a causal factor in addictive gambling. Depressed mood did not enhance persistence in regular gamblers, but rather failed to have the inhibitory effect observed in non-regular gamblers. Evidence of mood changes during gambling was obtained, although multiple measures (self-report, physiological indicators, anticipated states) of both affective valence and arousal are advocated for future studies.

Induced mood and persistence at gaming Emotional factors in gambling behaviour, particularly problem gambling, have received little attention, with even less attempt to draw upon relevant emotion theory. Such literature as does exist tends to focus on arousal and depression (see Dickerson & Adcock, 1987; Dickerson, 1993; Grif ths, 1995; Coventry & Norman, 1997, 1998; Coventry & Constable, 1999). Equating arousal with excitement, Grif ths (1995) has suggested these two factors might constitute the positive and negative reinforcers

that serve to maintain gambling behaviour. Evidence suggests that regular gamblers gamble to attain a positive state of high arousal and excitement, sometimes facilitated by a concomitant escape from depression.

Arousal and gambling An assumption in the gambling literature is that heightened arousal, typically measured by heart rate increases, is an indicator of excitement. The seminal study of Anderson & Brown (1984)

Correspondence to: Adelma M. Hills, School of Psychology, University of Western Sydney, Locked Bag 1797, South Penrith Distribution Centre, South Penrith NSW 1797, Australia. E-mail: am.hills@uws.edu.au Submitted 28th February 2001; initial review completed 17th April 2001; nal version accepted 16th May 2001.
ISSN 0965–2140 print/ISSN 1360–0443 online/01/111629–10 Ó Society for the Study of Addiction to Alcohol and Other Drugs Carfax Publishing, Taylor & Francis Limited DOI: 10.1080/09652140120080750

1630

Adelma M. Hills et al. Pathological gamblers and to a lesser extent other regular gamblers were more inclined to also report feeling fed up or depressed during play, and this pattern persisted after play. After play more regular than non-regular gamblers reported feeling in a bad mood or angry. One complication was that people often reported feeling more than one emotion, and feeling in a good or bad mood tended to be associated with winning and losing during a gambling session. Grif ths argued that depressed mood prior to gambling may be alleviated by gambling, and excitement during the game may reinforce regular and eventually pathological gambling. The ndings on depressed mood were claimed to strengthen the evidence for a causal link between depression and gambling, despite the abiding question of whether depression causes gambling, or uncontrollable gambling results in depression.

showed that during blackjack play in a casino, regular players showed large changes in heart rate (increases of up to 53 beats per minute (bpm)) and subjective reports of excitement. While such heart rate increases during many forms of gambling have been found, particularly for high frequency gamblers, the results across studies have not been entirely consistent (see reviews in Grif ths, 1995; Coventry & Norman, 1998). Generally, scant attention has been paid to the meaning of heightened arousal, and in some studies (e.g. Dickerson & Adcock, 1987; Coventry & Constable, 1999) a measure of state anxiety, not excitement, has been used as an accompanying subjective measure of arousal. However, on the all-important pleasantness dimension of emotion, anxiety and excitement lie at opposite ends. Coventry & Constable (1999) noted the desynchrony between subjective and physiological measures of arousal, and have recommended that measures be able to distinguish between positive and negative aspects of arousal.

Depression and gambling A number of studies have indicated that clinical depression, or transient dysphoric or depressed mood often accompanies frequent gambling (Corless & Dickerson, 1989; Dickerson et al., 1991, 1996; Raviv, 1993; Grif ths, 1995; Castellani et al., 1996). Based on self-reports of the association between diagnosable mood disorders and pathological gambling, McCormick et al. (1984) found that the onset of major depressive episodes appeared to precede the onset of gambling. They reported that gambling seemed to act as an antidepressant for many participants who found the activity energizing and mood-altering. Additionally, a national survey of the extent of gambling related problems in Australia by Dickerson et al. (1996) found that 9% of the population gambled as an escape from feeling depressed. Using retrospective self-reports, Grif ths (1995) asked 60 fruit machine players (30 regular gamblers, 11 of whom were described as pathological; and 30 non-regular gamblers) about the moods they experienced before, during and after gambling. Around 43% of regular players compared to only 7% of non-regular players reported feeling fed up or depressed before play. Around 83% of the regulars were excited during play compared to only 33% of the non-regulars.

Valence and arousal dimensions of emotion From the perspective of emotion theory most of this research confounds the two primary dimensions of emotion, namely, affective valence (from pleasant to unpleasant) and arousal (see the Special Section on the structure of emotion in Journal of Personality and Social Psychology, especially the papers by Diener, 1999; Russell & Barrett, 1999; Watson et al., 1999). Arousal per se is not necessarily pleasurable or a source of reward. It is rewarding when positively valenced in the form of excitement, but punishing when negatively valenced in the form of anxiety. Other things being equal, we are motivated to approach exciting stimuli but to avoid those that induce anxiety. Autonomic measures of arousal such as heart rate and skin conductance are notorious for their inability to distinguish between excitement and anxiety. Similarly, feeling depressed is a state of negatively valenced low arousal, while relaxation is a state of positively valenced low arousal. From an emotional and also motivational perspective the critical dimension is affective valence not arousal, and it is imperative that research speci cally addresses the affective valence associated with any emotional state or level of arousal. Recent Canadian research by Schrans, Schellinck & Walsh (2000) has lent some weight to the possibility that gambling can also be rewarding because it is relaxing as opposed to exciting. They studied players of video lottery

Mood and gaming persistence terminals, a form of electronic gaming machine (egm). The reasons given by 98-problem and 83-non-problem players for why they began to play more regularly included for fun, entertainment and excitement (30%), to win money (30%) and to avoid boredom (20%). However, among 46 unresolved problem-players 37% reported fun and entertainment as a motivation for play, while 24% cited relaxation and the chance to calm down and escape their problems. Twenty per cent cited the chance to win money. Most gambling studies involving emotional factors have relied on self-reports and/or physiological measures of arousal; and most have been essentially correlational in nature, rendering causal explanation problematic. The question of whether prior mood states have differential effects on the gambling behaviour of regular and non-regular gamblers has not yet been answered, nor has it been de nitely established how affective valence changes across gambling sessions for regular and non-regular gamblers.

1631

High affect infusion strategies, on the other hand, occur under conditions of heuristic processing and substantive processing. The former involves simple judgemental targets, of little personal relevance, in the absence of strong motivational objectives, when there are no situational demands for accuracy or detailed processing. Concurrent mood may well in uence the nature and extent of processing undertaken. Substantive processing is the most cognitively demanding strategy, where situational demands require accurate judgments of complex or novel targets, in the absence of any preexisting motivational bias. Under such conditions concurrent moods are hypothesized to in uence the kinds of information accessed from memory to arrive at a decision.

The affect infusion model Emotion theory and research can inform these questions. There is a considerable literature on the effects of positively and negatively valenced mood on decision-making (see Forgas, 1995a, for a review). At its most basic, there is evidence that happy people use simpli ed, heuristic strategies for deciding what to do, whereas sad people use more careful and substantive processing strategies. However, Forgas (1995a) has developed a more comprehensive theory of how affect infuses the decision-making process. His affect infusion model (AIM) postulates four different decision-making contexts, two of which are relatively impervious to mood in uences and two of which are susceptible. Whenever the judgemental target is very familiar or typical (e.g. which familiar consumer product to buy), past judgements stored in memory can be accessed directly with little processing of information and little opportunity for affect infusion. Similarly, there is little affect infusion in motivated processing, where there is strong motivation to achieve a particular outcome. Here, information processing is highly selective and “a speci c, preexisting preference [guides] information search and judgment” (Forgas, 1995a, p. 47). Mood maintenance or mood repair can itself be a goal of motivated processing.

The affect infusion model applied to gambling We applied the predictions of the AIM to the gambling context in an experimental design to examine the impact for regular and non-regular gamblers of prior induced mood (depressed, neutral, happy) on gaming persistence, operationalized as the number of trials played on a computerized card-cutting game under conditions of increasing nancial loss. The importance and relevance of gaming persistence as the dependent variable follows limited evidence that problem gamblers nd it somewhat easier to maintain control over the number of sessions started per week, but that control over session length is much more dif cult (Dickerson, Hinchey & Legg England, 1990; Schellink & Schrans, 1998). Affective valence on a happy– sad mood continuum was tracked during the course of the experiment; that is, after the mood induction, during play and after play. It was hypothesized from the AIM that high frequency gamblers would be more motivated processors, with a strong pre-existing motivation to gamble. The desired outcome could be either or all of winning money, achieving a positive mood state (mood repair) or maintaining a positive mood state (mood maintenance). According to the AIM there is little affect infusion in motivated processing, therefore, gambling persistence was hypothesized to be no different for regular gamblers who had either depressed, neutral or happy prior mood inductions. This hypothesis contrasts with one that can be derived from the link between depression and problem gambling.

1632

Adelma M. Hills et al. Materials Computerized card-cutting game. A computerbased analogue to card machine gambling created by Breen & Zuckerman (1999) served as the gambling game. The game, which was presented on an IBM Pentium computer, involves the player betting on either a low card (cards 1 to 7) or a high card (9 to Ace). Provided there is suf cient available credit, the player can bet up to $10 on each trial. After each trial the computer displays the message “Do you want to continue (Y/N)?”, and the player chooses to continue or discontinue the gambling session. If the player presses the “Y” key, the program responds with the question “How much do you want to bet?” Once the player has typed in the value of the bet, the program automatically prompts “Do you want to bet high or low (H/L)?”, and the player presses the appropriate key. The computer program then cuts the deck, displays the selected card, then the message “You have won $x.xx” or “You have lost $x.xx” to indicate the outcome of the trial, after which it prompts to begin a new trial. As trials progress the percentage of winning games slowly reduces from 50%. Film-based mood induction procedure. Reviews of the validity and effectiveness of mood induction procedures by Westermann et al. (1996) and Gerrards-Hesses, Spies & Hesse (1994) have indicated that lm/story procedures are the most effective for inducing both elated and depressed moods. Thus, three 10-minute edited videos of existing lms, found to be effective in previous research by Forgas (1990, 1995b), were used as the mood induction procedures. These were from the lm Who Will Love my Children? that depicts a tragic story of a mother dying of cancer for the depressed mood induction, a documentary on The Science of Psychology for the neutral induction, and an excerpt from an episode of the British comedy series Fawlty Towers for the happy mood induction. The latter was also used as the debrie ng procedure for participants previously exposed to the sad lm.

If depression plays a causal role, gaming persistence should be greater in the depressed mood condition. In contrast, non-regular gamblers are unlikely to have a strong motivation to gamble, gambling is unlikely to be so familiar to them as to invoke direct access processing, nor is it likely to embody the conditions that generate substantive processing. Instead, non-regular gamblers could be expected to employ a heuristic processing strategy in relation to their gambling-related decisions, particularly for the more simple forms of gambling analogous to electronic gaming machine play. As such, prior mood was hypothesized to impact on gambling persistence for non-regular players. Speci cally, depressed nonregular gamblers were hypothesized to play fewer trials than happy non-regular gamblers, because the former would be inclined to process relevant information more carefully and systematically, soon realizing that skill was not a factor and losing money was the most likely long-term outcome. Happy players, on the other hand, could be expected to engage in less substantive processing and to pursue the game for its novelty value with less concern for the eventual outcome.

Method Participants The participants were 120 male students over the age of 18 years from the university of Western Sydney who volunteered to take part in the study. Their mean age was 25.95 years (SD 5 8.28), and all had indicated that electronic gaming machines were their preferred form of gambling. Sixty participants who gambled at least once a week were classi ed as regular gamblers, and 60 who gambled less than once a month comprised the non-regular gamblers. Scores of 14 or above on the Beck Depression Inventory II (Beck, Steer & Brown, 1996) were used to screen out any potential participants who were clinically depressed. Participants from each gambler classi cation were then assigned randomly to either a depressed, neutral or happy mood induction group. All participants were treated in accordance with the ethical standards of the Australian Psychological Society, and signed an informed consent form before participation.

Procedure Participants completed the study individually. Depending on the mood induction condition to which he had been assigned, each participant began the study by watching the relevant video. To counter demand characteristics it was explained

Mood and gaming persistence that the video may or may not evoke feelings, and that it was safe to go along with any feelings if they occurred (Westermann et al., 1996). Immediately after watching the video, the participant indicated his current mood by placing a mark on a 10-cm bipolar visual analogue scale with the words sad and happy at either end (the post-induction mood measure). The participant was then thanked for taking part in the study and given AUD$10.00 as reimbursement for travelling expenses, before being asked if he would like to continue with the study and play the computerized card game using the $10.00 as 10 credit points for the game. He was told he could play the game for as long as he liked and stop whenever he wished, and that any money won on the game was his to keep. All participants agreed to continue. They were then given details of how to play the game, and after playing ve practice trials they indicated their current mood on a second visual analogue scale (the during-play mood measure). When the participant decided to nish the game he indicated his current mood on a third visual analogue scale (the after-play mood measure). Before leaving, each participant was informed about the nature of the study, thanked for his participation, and paid any money he had won on the game. This debrie ng for participants exposed to the depressed mood induction included watching the happy mood induction video to alleviate any residual effects, as Frost & Green (1982) had found induced depressed mood could linger longer than 10 minutes in the absence of such debrie ng. These participants then completed a nal mood scale to con rm they left in a happy mood. Results Mood across the gambling session as a function of mood induction and type of gambler The bipolar visual analogue scales were scored

1633

from 0 to 10, with 0 indicative of maximum sadness and 10 maximum happiness. A score of 5 occurred at the midway point; scores below 5 were in the sad end of the range; those above 5 in the happy end. Mood scores at three times (post-induction, during-play, and after-play) were analysed using a mixed repeated-measures analysis of variance (ANOVA), with type of gambler (regular, nonregular) and mood induction group (depressed, neutral, happy) as between-subjects factors. The ANOVA test assumptions were satisfactory, except that homogeneity of covariance was not achieved for time. Using an alpha level of 0.05, there was a statistically signi cant main effect of mood induction, F(2, 114) 5 12.56, p , 0.001, g 2 5 181. With a Greenhouse–Geisser adjustment to the degrees of freedom, there was also a signi cant main effect of time F(1.889, 215.294) 5 18.55, p , 0.001, g 2 5 0.14; and signi cant interactions of time 3 mood induction, F(3.777, 215.294) 5 39.38, p , 0.001, g 2 5 0.409; and time 3 type of gambler, F(1.889, 215.294) 5 7.24, p 5 0.001, g 2 5 0.06. None of the other effects achieved signi cance (p . 0.20). Descriptive statistics for the time by mood induction interaction are shown in Table 1. Of primary interest were the post-induction means, which provided manipulation checks for the experimental induction of mood. Tukey post hoc pairwise comparisons showed that the mood induction was successful; all the means were signi cantly different from one another, with the saddest score in the depressed mood group and the happiest score in the happy mood group. The neutral mood group was midway between these two extremes with a somewhat happy mean. During play and after play this pattern disappeared, there being no signi cant differences among the three mood induction groups, all having means in the mid- to happy range.

Table 1. Mean mood scores at three times as a function of mood induction group Post-induction Mood induction Depressed Neutral Happy Total M 3.45 6.16 7.78 5.80 SD 1.93 1.33 1.45 2.34 During-play M 6.49 6.91 6.90 6.77 SD 1.63 1.68 1.56 1.62 After-play M 6.52 5.94 6.63 6.36 SD 2.21 2.10 1.83 2.06 Total M 5.49 6.34 7.10

1634

Adelma M. Hills et al.
Table 2. Mean mood scores at three times as a function of type of gambler Post-induction Gambler type Regular Non-regular Total M 5.68 5.92 5.80 SD 2.42 2.37 2.34 During-play M 6.90 6.64 6.77 SD 1.50 1.74 1.62 After-play M 5.89 6.84 6.36 SD 2.07 1.94 2.06 Total M 6.15 6.47

Descriptive statistics for the time by type of gambler interaction are shown in Table 2 and illustrated in Fig. 1. Simple effects analyses (collapsed across mood induction) with Sidak contrasts showed a signi cant effect of time for both regular gamblers, F(1.705, 100.619) 5 9.05, p 5 0.001 and non-regular gamblers, F(1.481, 87.378) 5 6.28, p 5 0.003. The contrasts showed a signi cant increase in happiness during play compared to post-mood induction for both groups. While there was no signi cant difference during and after play for non-regular gamblers, regular gamblers exhibited a signi cant decrease in happiness after play compared to during play. The difference between regular and non-regular gamblers was not signi cant post-mood induction, t(118) 5 0.58, p 5 0.58 nor during play, t(118) 5 0.87, p 5 0.39; but regular gamblers had signi cantly lower happiness scores than nonregular gamblers after play, t(118) 5 2.60, p 5 0.01.

One possible explanation for this after-play difference arose from the fact that the regular gamblers won less (M 5 $3.90, SD 5 $8.66) than the non-regular gamblers (M 5 $7.23, SD 5 $7.50), t(118) 5 2.25, p 5 0.026. Forty- ve regular gamblers left with nothing, compared to 25 non-regular gamblers. The winning regular gamblers won between $5 and $50, while the winning non-regular gamblers won between $5 and $30. Furthermore, there was a signi cant positive correlation between amount won and mood for regular gamblers, r(58) 5 0.41, p 5 0.001, but not for non-regular gamblers, r(58) 5 0.20, p 5 0.135. Among regular gamblers, the 15 winners were signi cantly happier (M 5 7.51, SD 5 1.67) than the 45 losers (M 5 5.34, SD 5 1.92), t(58) 5 3.91, p , 0.001; furthermore, the winners played signi cantly fewer trials (M 5 16.73, SD 5 7.70) than the losers (M 5 30.76, SD 5 14.79), unequal variances t(46.970) 5 4.72, p , 0.001.

Figure 1. Mean mood scores at each of three times for regular and non-regular gamblers.

Mood and gaming persistence
Table 3. Mean number of trials played as a function of type of gambler and mood induction group Regular gamblers Mood induction Depressed Neutral Happy Total M 30.55 23.25 27.95 27.25 SD 12.90 14.51 16.15 14.65 Non-regular gamblers M 11.70 16.80 21.70 16.73 SD 4.82 11.66 13.78 11.38 M 21.12 20.02 24.83 21.99 Total SD 13.55 13.40 15.16 14.09

1635

Trials played as a function of mood induction and type of gambler A 2 3 3 type of gambler (regular, non-regular) by mood induction group (depressed, neutral, happy) between-subjects ANOVA was performed on the number of trials played on the computerized gambling game in one gambling session. The ANOVA test assumptions were satisfactory, except that homogeneity of variance was violated. Using an alpha level of 0.05, the main effect of mood induction was not signi cant, F(2,114) 5 1.54, p 5 0.219, g 2 5 0.026. However, there was a signi cant main effect of type of gambler, F(1,114) 5 20.17, p , 0.001, g 2 5 15 (regular gamblers played more trials than non-regular gamblers); and a signi cant type of gambler by mood induction interaction, F(2,114) 5 3.17, p 5 0.046, g 2 5 0.053. The descriptive statistics are shown in Table 3, and illustrated in Fig. 2.

Post hoc pairwise comparisons were performed using the Games–Howell procedure for unequal variances. These showed no signi cant differences in the number of trials played by regular gamblers across any of the mood induction groups. For non-regular gamblers, however, there was a clear trend of increasing trials played from depressed, through neutral, to happy mood induction groups. Signi cantly more trials were played by the happy mood group than by the depressed mood group. However, the differences between the neutral and depressed mood groups, and the neutral and happy mood groups were not signi cant. The standard deviation was also signi cantly lower in the depressed mood group. While none of these participants played more than 20 trials, 45% of the neutral non-regular group played more than 20 trials, as did 50% of the happy non-regular group.

Figure 2. Mean number of trials played as a function of type of gambler and mood induction groups.

1636

Adelma M. Hills et al. motivate problem gamblers to gamble whatever mood they are experiencing at the outset. The second goal of the study was to track changes in affective valence across the gambling session. Regular and non-regular gamblers had comparable mood scores post-mood induction and during play. They both showed a signi cant increase in happiness during play compared to post-mood induction, although this increase was greater for regular players. Non-regular gamblers showed no change in mood after play compared to during play, and their after-play mood was not related to winnings. However, after-play mood was related to winnings for regular gamblers. Those regular gamblers who lost were signi cantly more unhappy after-play compared to during-play, and signi cantly more unhappy than either their counterparts who had won or non-regular gamblers. It is easy to understand why winning gamblers would be motivated to gamble on another occasion, but feeling less happy as a result of losing ordinarily should inhibit subsequent gambling. Unfortunately, the results of this study suggest regular gamblers may have lost any such inhibition—sad mood provides no emotional deterrent to continued gambling. All these ndings provide possible insights into the psychological processes that initiate the development of impaired control that in the longerterm results in harmful impacts to the player. If the relatively greater increase in happiness experienced by the regular gamblers during play compared to post-mood induction is reliable, it begs the question of whether this characterized their earliest gambling experiences. If so, it could be expected to predispose them to becoming regular gamblers. In the process, the loss of an inhibiting effect of depressed mood would remove one more barrier on the path to pathological gambling. Increasingly, gambling could be relied upon to provide depressed mood repair in addition to happy mood maintenance. This, combined with the special features of gambling that give it an addictive character (absent in passions for football, or stamp collecting, for instance), namely, the use of a powerful reinforcer (money), delivered on an intermittent reinforcement schedule, in a socially encouraged and undemanding context, where little effort is required, could be expected to put such individuals at severe risk of progressing to pathological gambling.

There was no signi cant difference in the number of trials played by regular and nonregular gamblers exposed to the neutral mood induction or the happy mood induction. However, depressed regular players played signi cantly more trials than depressed nonregular players.

Discussion The predictions of the AIM, based on the assumption that regular gamblers are motivated processors and non-regular gamblers are heuristic processors, were supported by the results. There were no signi cant differences in gaming persistence for regular gamblers across depressed, neutral and happy prior mood inductions. There was no evidence that prior mood infused their decisions to continue or discontinue gambling. In contrast, non-regular gamblers showed reduced gaming persistence across prior mood conditions of happy through neutral to sad. These results for gaming persistence, to the extent they have ecological validity, suggest that mild, non-clinical levels of dysphoria or sadness prior to a session of gaming do not predispose players to lengthy, potentially problematic sessions of gaming as previously suggested (Dickerson & Adcock 1987; Dickerson, 1993). Regular gamblers appear motivated to gamble regardless of their prior mood and despite increasing losses. Potentially most important was the fact that prior depressed mood did not inhibit their gambling, as it appeared to do for non-regular gamblers. Speculating to the wider context, this could explain why at any one time we could expect to nd relatively more problem gamblers who are depressed compared to non-problem gamblers, and why relatively fewer non-regular gamblers are likely to self-report feeling depressed prior to starting a gambling session. When feeling depressed, non-regular gamblers may be reluctant to gamble at all, as found by Corless & Dickerson (1989). This interpretation does not preclude the possibility that winning money, mood maintenance and mood repair are all part of the motivation for gambling, with mood maintenance being important for happy gamblers and mood repair for depressed gamblers. In this sense, prior depressed mood could well be a causal factor, but because there are multiple mood effects one or other aspect will

Mood and gaming persistence One aspect of our ndings was not entirely satisfying. The self-reports of current mood state indicated that the mood induction procedure worked well, but that the effects had disappeared by the time participants began to gamble proper. Why then should we expect differences if by the time participants were making the decision to continue or not everyone was relatively happy? Even so, the pattern of differences in gaming persistence for non-regular gamblers was clear. We can only assume that there were ongoing residual effects of the prior mood induction of which participants were not aware, or that were recurring intermittently. The time of the duringplay measure was also at a point where the novelty value of the game and the fact they had just been given $10 may have momentarily reversed any depressed mood induction effects, that none the less returned to affect subsequent decision-making. Further research is needed to clarify this issue, perhaps by asking if any thoughts and feelings associated with the sad video arose during the game. Despite participants being told that the mood induction video may or may not evoke feelings, there were clear demand characteristics associated with the experimental procedure that could have accounted for the differences in mood ratings post-mood induction, and after practice trials on the card game. Perhaps all mood groups reported similar relatively happy moods at this point because they assumed the game was intended to have this effect. Mood-related demand characteristics could possibly have in uenced the results in a variety of ways. For this reason, future research is well advised to include distractor items and any other devices to disguise the focus on mood state and true purpose of the experiment. The ecological validity of the present study invokes an issue well recognized in gambling research (Anderson & Brown, 1984; Dickerson, 1993). The computerized card-cutting game developed by Breen & Zuckerman (1999), together with the return of the $10 participation fee, seems to have good face validity to players (Kyngdon & Dickerson, 1999). The greater persistence, regardless of mood induction, shown by regular players suggests that the more frequent and longer egm sessions of such players in reallife venues generalizes to the laboratory gaming situation. Even so, it must be acknowledged that the present research took place in a relatively

1637

arti cial and impoverished gambling context, involved quite mild mood inductions and sampled a small part of the emotional repertoire experienced during a gambling session. Arguably, this has been an important beginning, with a rare grounding in emotion theory, from which future research can develop. It has served to demonstrate the value of applying emotion theory, the AIM in this case, to an understanding of gambling behaviour. At the same time the validity of the AIM has been enhanced through this test of its predictions in an applied context, although further research is needed to determine precisely why depressed and happy prior moods affected non-regular gamblers as they did. Do players in these two conditions process information about gambling in the different ways predicted by affect infusion theory? The longitudinal question is also important; that is, whether regular gamblers’ emotional responses to gambling change over time or are in evidence from the outset. The emotional experiences of female gamblers and of gamblers in the realworld context must also be addressed. Future research might also examine emotional experience during gambling in much more detail, using both subjective self-reports and physiological measures, especially with a view to understanding the interplay of affective valence and arousal in producing depression, relaxation, excitement and anxiety. In this respect, it may be important to revisit reversal theory (Apter, 1989), which sets arousal and affective valence in the context of a comprehensive theory of motivation, emotion and personality. Among other things it predicts personality differences in the motivation to avoid boredom by seeking excitement, or to avoid anxiety by seeking relaxation, all of which have implications for the gambling experience of different individuals (see Anderson & Brown, 1987; Brown, 1988). Those individuals who nd risking money produces high arousal anxiety more than high arousal excitement may be protected from gambling excesses. Others may live lives of such high anxiety that gambling provides a source of relaxation, as reported by some individuals in the Schrans et al. (2000) study. In addition to examining experienced emotions, regret theory (Baron, 1994) and subjective expected emotion theory (Mellers, Schwartz & Ritov, 1999) point to the need for research to also examine the role of anticipated emotions in gambling decisions and behaviour.

1638

Adelma M. Hills et al.
for media-recruited problem gamblers, Journal of Gambling Studies, 6, 87–102. DIENER, E. (1999) Introduction to the Special Section on the structure of emotion, Journal of Personality and Social Psychology, 76, 803–804. FORGAS, J. P. (1990) Affective in uences on individual and group judgments, European Journal of Social Psychology, 20, 441–453. FORGAS, J. P. (1995a) Mood and judgment: The Affect Infusion Model (AIM), Psychological Bulletin, 117, 39–66. FORGAS, J. P. (1995b) Strange couples: mood effects on judgments and memory about protypical and atypical relationships, Personality and Social Psychology Bulletin, 21, 747–765. FROST, R. O. & GREEN, M. L. (1982) Velton mood induction procedure effects: duration and postexperimental removal, Personality and Social Psychology Bulletin, 8, 341–347. GERRARDS-HESSE, A., SPIES, K. & HESSE, F. W. (1994) Experimental inductions of emotional states and their effectiveness: a review, British Journal of Psychology, 85, 55–78. GRIFFITHS, M. (1995) The role of subjective mood states in the maintenance of fruit machine gambling behaviour, Journal of Gambling Studies, 11, 123– 135. KYNGDON, A. & DICKERSON, M. G. (1999) An experimental study of the effect of prior consumption of alcohol on a simulated gaming activity, Addiction, 94, 697–707. MC CORMICK , R. A., RUSSO , A. M., RAMIREZ, L. F. & TABER, J. I. (1984) Affective disorders among pathological gamblers seeking treatment, American Journal of Psychiatry, 141, 215–218. MELLERS, B., SCHWARTZ, A. & RITOV , I. (1999) Emotion-based choice, Journal of Experimental Psychology: General, 128, 332–345. RAVIV, M. (1993) Personality characteristics of sexual addicts and pathological gamblers, Journal of Gambling Studies, 9, 17–30. RUSSELL, J. A. & BARRETT, L. F. (1999) Core affect, prototypical emotional episodes, and other things called emotion: dissecting the elephant, Journal of Personality and Social Psychology, 76, 805–819. SCHELLINCK, T. & SCHRANS, T. (1998) Nova Scotia VL Players’ Survey, Report by Focal Research Consultants Ltd for the Nova Scotia Department of Health, Halifax, Nova Scotia. SCHRANS, T., SCHELLINCK, T. & W ALSH, G. (2000) Department of Health Nova Scotia regular VL players follow-up study: a comparative analysis of problem development and resolution, Report by Focal Research Consultants Ltd for the Nova Scotia Department of Health, Halifax, Nova Scotia. WATSON, D., WIESE, D., VAIDYA, J. & TELLEGEN, A. (1999) The two general activation systems of affect: structural ndings, evolutionary considerations, and psychobiological evidence, Journal of Personality and Social Psychology, 76, 820–838. WESTERMANN, R., SPIES, K., STAHL, G. & HESS, F. W. (1996) Relative effectiveness and validity of mood induction procedures: a meta-analysis, European Journal of Social Psychology, 26, 557–580.

References
ANDERSON, G. & BROWN , R. I. F. (1984) Sensation seeking and arousal in real and laboratory gambling, British Journal of Psychology, 85, 401–410. ANDERSON, G. & BROWN , R. I. F. (1987) Some applications of reversal theory to the explanation of gambling and gambling addictions, Journal of Gambling Behavior, 3, 179–189. APTER, M. J. (1989) Reversal Theory: motivation, emotion and personality (London, Routledge). BARON, J. (1994) Thinking and Deciding (Cambridge, Cambridge University Press). BECK, A. T., STEER, R. A. & BROWN , G. K. (1996) Beck Depression Inventory—11 (San Antonio, TX, The Psychological Corporation). BREEN, B. & ZUCKERMAN, M. (1999) “Chasing” in gambling behaviour: personality and cognitive determinants, Personality and Individual Differences, 27, 1097–1111. BROWN , R. I. F. (1988) Reversal theory and subjective experience in the explanation of addiction and relapse, in: APTER, M. J., KERR, J. H. & COWLES, M. P. (Eds) Progress in Reversal Theory. Advances in psychology, 51, pp. 191–211 (Amsterdam, NorthHolland). CASTELLANI, B., WOOTTON , E., RUGLE, L., WEDGEWORTH, R., P RABUCKI, K. & OLSEN, R. (1996) Homelessness, negative affect, and coping among veterans with gambling problems who misused substances, Psychiatric Services, 47, 298–299. CORLESS, T. & DICKERSON, M. G. (1989) Gamblers’ self perceptions of the determinants of impaired control, British Journal of Addiction, 84, 1527–1537. COVENTRY, K. R. & CONSTABLE, B. (1999) Physiological arousal and sensation seeking in female fruit machine gamblers, Addiction, 94, 425–430. COVENTRY, K. R. & NORMAN, A. C. (1997) Arousal, sensation seeking and frequency of gambling in offcourse racing bettors, British Journal of Psychology, 88, 671–681. COVENTRY, K. R. & NORMAN, A. C. (1998) Arousal, erroneous verbalizations and the illusion of control during a computer-generated gambling task, British Journal of Psychology, 89, 629–645. DICKERSON, M. G. (1993) Internal and external determinants of persistent gambling: problems in generalising from one form of gambling to another, Journal of Gambling Studies, 9, 225–245. DICKERSON, M. G. & ADCOCK , S. G. (1987) Mood, arousal and cognitions in persistent gambling: Preliminary investigation of a theoretical model, Journal of Gambling Behaviour, 3, 3–15. DICKERSON, M. G., BARON, E., HONG , S. M. & COTTRELL, D. (1996) Estimating the extent and degree of gambling related problems in the Australian population: a national survey, Journal of Gambling Studies, 12, 161–178. DICKERSON, M. G., CUNNINGHAM, R., LEGG ENGLAND, S. & HINCHY, J. (1991) On the determinants of persistent gambling behaviour III: personality, prior mood and poker machine play, International Journal of Addictions, 26, 531–548. DICKERSON, M. G., HINCHY J. & LEGG ENGLAND, S. (1990) An evaluation of a minimal intervention