Int J Ment Health Addiction DOI 10.

1007/s11469-010-9281-6

Social Facilitation in Online and Offline Gambling: A Pilot Study

# Springer Science+Business Media, LLC 2010

Abstract To date, there has been relatively little research on Internet gambling. Furthermore, there have been few studies comparing the behaviour of Internet gamblers versus non-Internet gamblers. Using the game of roulette, this study experimentally examined (a) the differences in gambling behaviour between online and offline gamblers, and (b) the role social facilitation in gambling behaviour between online and offline gamblers. A total of 38 participants played online and offline roulette either alone or alongside another gambling participant, and the players chip placement and amount bet was recorded. The study found that those who gambled in online roulette placed more chips per bet and made riskier bets than those who gambled on roulette offline. The study also found that those who gambled alongside another gambler placed more chips and made riskier bets than those who gambled alone. Those who gambled online and in the presence of others, placed the highest number of chips per bet and made the riskiest bets. Keywords Gambling . Online gambling . Internet gambling . Social facilitation . Problem gambling

Social Facilitation in Online and Offline Gambling: A Pilot Study To date, there has been relatively little empirical research on Internet gambling. In general, early studies of Internet gambling prevalence show low prevalence rates of online gambling. For instance, the 1999 British Gambling Prevalence Study (BGPS) found that only 0.2% of the population had gambled online (Sproston et al. 2000). Griffiths (2001) reported that 1% of Internet users had gambled online. Later studies have found increases in Internet gambling prevalence rates. For instance, the latest BGPS showed that (excluding
T. Cole : M. D. Griffiths (*) International Gaming Research Unit, Psychology Division, School of Social Sciences, Nottingham Trent University, Burton Street, Nottingham NG1 4BU, UK e-mail: mark.griffiths@ntu.ac.uk D. J. K. Barrett Psychology Department, University of Leicester, Leicester, UK

Int J Ment Health Addiction

National Lottery gambling), 6% of the British population had gambled online (Griffiths et al. 2009). Reasons for gambling online are multiple. Griffiths and Barnes (2008) found the main reasons for gambling online were: ease of access (84%), flexibility of use (75%), 24-hour availability (66%), because friends do (67%), large variety of gambling choice (57%), successful advertising (40%), anonymity (25%), the opportunity to play demo (free play demonstration) games (21%), and because family members did (14%). Studies have suggested that the main reason for gambling online is convenience (Williams and Wood 2007). Furthermore, a number of studies have reported that Internet gamblers are more likely than non-Internet gamblers to be problem gamblers (Ladd and Petry 2002; Griffiths and Barnes 2008; Griffiths et al. 2009). Such findings suggest the importance of research into Internet gambling behaviour. One possible reason for the difference between online and offline gamblers may be due to the fact that online gamblers are playing with virtual representations of money (i.e., e-cash). Griffiths (2003) has argued that electronic cash does not have the same psychological value as gambling with real money. Gambling with e-cash may lead to a suspension of judgement that disrupts the player s financial value system. In essence, Griffiths argues that compared to real money, e-cash has a lower psychological value. The same principle applies when people gamble with other virtual representations of money (e.g., chips, tokens, credits, etc.). These representations of money may be re-gambled without too much thought compared to real money. Although there is lots of anecdotal evidence that players gamble more with e-cash, there is little empirical evidence. Although there are a number of theoretical papers showing how online and offline gambling differ and possible psychosocial impacts of these differences (Griffiths and Parke 2002; Griffiths 2003; Griffiths et al. 2006), there has been little empirical research investigating these differences apart from a few surveys (Griffiths et al. 2009). Using a case study approach, Griffiths and Parke (2007) examined the differences between online and offline gamblers. Using thematic analysis, the results showed that the traditional gamblers expressed a strong desire to gamble on the Internet for reasons such as convenience (e.g., hours, proximity), improved facilities (e.g., access to their account), and tax-free betting. However, they also reported that there were barriers to Internet gambling including the inability to obtain valid credit or debit cards, and the lack of the physical transaction of collecting winnings that can be highly rewarding. Their findings indicated that there were a number of subtle differences between the two types of gambler on a number of dimensions (i.e., financial stability, motivation, physiological effects, competition, need for acknowledgement, and social facilitation). Although Griffiths and Parke (2003, 2007) have identified social facilitation as a potentially important factor in the maintenance of gambling behaviour, it has been the focus of very few empirical studies. However, Rockloff and Dyer (2007) carried out an experimental study suggesting that players may increasingly engage in risky gambling as a way of impressing other players. In their experiment, participants (n =116) gambled on a slot machine simulation featuring a pre-programmed winning sequence that was then followed by an indefinite losing one. During the experiment, various data were collected from each player including the gambling speed, their average size of bet, the number of games played, and the final loss amount. During the experimental trials, some players were given computer-generated false feedback suggesting that other players nearby were playing the same game and winning. The studys findings indicated that players bet more and lost more money when receiving information about other players winning compared to those players in similar experimental trials but receiving much less information. Therefore, study

Int J Ment Health Addiction

appeared to suggest that even just the (implied) presence of other slot machine players increased gambling intensity (i.e., they played and gambled more than those players gambling alone). Hardoon and Derevensky (2001) conducted a study on children gambling in groups. They found that girls had increased mean wagers when they gambled in groups. No increases in mean wagers were found among the boys. Their results also indicated that children were more susceptible to peer influences than adults. Other factors, such as the medium of playing, may affect social facilitation. For instance, Griffiths and Parke (2007) have speculated that the one positive aspect of Internet gambling is that it may reduce the social facilitation risk. They also support the idea that gambling with friends may facilitate higher risk betting behaviour, although no empirical support was provided for such an assertion. Given the paucity of empirical evidence in the area, the following study experimentally examined (a) the differences in gambling behaviour between online and offline gamblers playing roulette, and (b) the role of social facilitation in gambling behaviour between online and offline gamblers playing roulette. Social facilitation was manipulated using a factorial design with two conditions: (1) gambling alone (online and offline), and (2) gambling alongside another player (online and offline). On the basis of previous (admittedly limited) empirical research, it was predicted that players would place higher bets in the online condition than offline, and stake higher bets when gambling alongside someone else.

Method Participants A total of 38 participants were recruited using an opportunity sample from an East Midlands university in the UK. The sample included 19 men and 19 women all of whom had gambled previously but were not problem gamblers. The mean age of the participants was 20.3 years (S.D.=1.8 years) and none of the participants was a problem gambler. Problem gambling was assessed using the South Oaks Gambling Screen (SOGS; see Procedure section for more details). Design The study followed a 22 design (i.e., gambling online alone, gambling offline alone, gambling online with others, and gambling offline with others). All players participated in both the online and offline conditions but participated in only one of the alone or with other conditions. Furthermore, the study was laboratory-based, participants did not use their own money, and the outcome (win/loss) was not paid in real money by or to the participants. Materials The study used an online roulette wheel and an offline roulette wheel. The offline roulette table and wheel was a standard American roulette design. The online roulette table and wheel utilised software downloaded from the UK gaming operator Ladbrokes (http://casino.ladbrokes.com/en/roulette/type). Participants were also provided with $200 chips. The denominations of chips were matched as closely as possible to the online version of the game. Participants were allocated one chip worth $50, three chips worth $20, five chips worth $10, six chips worth $5 and ten chips worth $1. Rationale for Using Roulette The reason for choosing roulette for this study was because roulette is widely played both in online and offline formats and requires no skill to play. It is also a game that can be played both alone or with others. Using roulette also provided

Int J Ment Health Addiction

another dimension in which to assess gambling behaviour. In roulette, players are able to place chips in a number of different areas. There are areas where the chances of winning are 1 in 2 (e.g., player betting on red or black), and other places where the odds are 1 in 35 (e.g., player betting on an individual number). Players are therefore making riskier bets if they place their chips in positions where there is a lower probability of winning. Players risk-taking behaviour can be measured not only by the value of chip stake they play per roulette spin, but also where they place their chips per roulette spin. Therefore, percentage of inside bets (betting on specific numbers) was recorded as a gauge of the participants risk-taking in addition to the stake amount they bet per roulette spin. Procedure Participants were randomly assigned to either gamble alone or in pairs (i.e., play alongside another participant). Before the experiment was conducted, all participants filled out a short questionnaire asking for basic demographic information alongside the South Oaks Gambling Screen (SOGS). The SOGS is a 20-item screening instrument, designed to identify the presence of problem gambling (Lesieur and Blume 1987). As mentioned above, none of the participants were classed as problem gamblers using this instrument. Once participants had completed this questionnaire, they were provided with information about roulette and how it is played and bet upon. Participants were asked to play both online and offline roulette (either alone or in pairs with someone else) for 15 min each. In the online roulette gambling condition, players gambled by choosing how much they wanted to gamble then placed chips in the positions they wished to play, which then lit up on the screen. Players were permitted to place as many chips as they wanted. Players then clicked the spin button to finish placing their bet. During the experiment in the online (gambling in pairs) condition, the first author provided the participants with feedback to each participant about how the other participant was doing while they were gambling. This was to keep each gambler informed of the progress of the other gambler so that the impact of social facilitation online was maximised. After 15 min (or when the participants ran out of chips), participants were told to stop playing and were asked to play the same game offline. In both conditions, participants were given the same amount of time and chips. To counterbalance the design, approximately half of the trials started with the online gambling task first followed by the offline gambling task, and vice-versa. In the offline condition, participants placed their own chips when betting. Once the chips had been placed by the participants (i.e., they had made their bet), the first author spun the roulette wheel and when the ball stopped, the winning number on was called out. Losing chips were removed from the table and returns were placed for winning bets. Players were permitted to place as many chips as they wanted in conditions identical to the online roulette gambling task. Following each bet, the amount gambled per spin was recorded in addition to the betting risk (i.e., the probability of winning). To ensure experimental control over each of the conditions, the offline roulette table was as similar to the online roulette table, with participants placed on the same side of the table. Both games used the American format of the game and players were given the same

Table 1 Mean Amounts Bet (and standard deviations) per Roulette Spin Condition Alone Paired Online mean amount bet (and SD) $32.87 ($18.03) $50.30 ($33.90) Offline mean amount bet (and SD) $25.87 ($11.80) $44.39 ($21.30)

Int J Ment Health Addiction

Fig. 1 Mean amount bet in dollars by gambling environment and social situation

amount of chips in both online and offline conditions. After completion of both phases of the experiment, participants were debriefed.

Results Two separate 2 (Medium: Online vs. Offline) 2 (Social Facilitation: Gamble in pairs vs. gamble alone) mixed design ANOVAs were carried out with medium (online vs. offline) as the repeated measure factor, and gambling in pairs or alone as the between subjects factor in order to measure the effect the two dependent variables had (amount bet per gamble and percentage of inside bets) (Table 1; Fig. 1). Levenes test of homogeneity of variance was significant for the amount bet (p <0.05), therefore the data were transformed by calculating the square root of the amount bet. Given that a transformation of the amount bet (square root) was performed, some degree of caution is therefore warranted in the interpretation of results. Upon transformation, the data fulfilled the requirements of homogeneity of variance (p >0.05). It was found that there was a main effect of medium on amount bet (F(1,36)=82.4; p <0.01) and also a main effect for social facilitation (F(1,36)=4.1; p <0.05). However, there was no significant interaction between medium and social facilitation in relation to the amount of chips bet (F(1,36)= 3.93; p >0.05). However, the data revealed a trend towards an interaction (p =0.055). The results indicated that participants placed higher bets online than they did offline. The results also indicated that those participants who gambled in pairs also placed more chips than those who gambled alone. However although there was no interaction between the two conditions, the data revealed a trend towards an interaction. This suggests the effect of social facilitation might vary depending upon the gambling medium (Table 2; Fig. 2). There was no main effect of medium on the amount of inside bet chips placed (F(1,36)= 3.5; p >0.05). However, there was a main effect for social facilitation on amount of inside bet chips placed (F(1,36)=98.5; p <0.05). There was also a significant interaction between medium and social facilitation in relation to percentage of inside bets placed (F(1,36)=62.3; p <0.05). The results here suggest that the presence of others increases the number of inside
Table 2 Percentage of inside bets (and standard deviations) Condition Alone Paired Online percentage inside bets (and SD) 44.4% (29.4%) 52.5% (38.5%) Offline percentage inside bets (and SD) 35.9% (25.8%) 44.3% (28.6%)

Int J Ment Health Addiction

Fig. 2 Percentage of inside bets by gambling environment and social situation

bets the participants placed during roulette play and that playing online in the presence of others also increases the amount of inside bets.

Discussion Overall, the results are consistent with the limited previous research. The results confirmed the hypothesis that players staked higher bets and made riskier bets online than offline, and that players staked higher bets and made riskier bets in a social situation than when playing alone. The results suggest that gambling online may be facilitating participants to place higher stakes per roulette spin and that gambling alongside others led players to stake more than when playing alone. There are a number of reasons that could perhaps help explain why online players staked higher bets and made riskier bets. For instance, it was clear that in the experiment, participants were able to gamble at much quicker rates in the online condition than the offline condition (i.e., the event frequency was a lot higher). This was because withdrawing and handing out chips won and lost was instantaneous online (whereas offline the process was much slower). Such an observation supports the wealth of literature that tends to show that certain types of gambling with high event frequencies tend to be more problematic to individuals (Meyer et al. 2009). Another noticeable difference between the online and offline gambling in the experiment was in relation to the differing sound effects during play. By winning chips, the online casino used in this experiment played a jovial or congratulatory noise. This could be argued as an effective way in which online casinos can facilitate gamblers playing behaviour by encouraging future play by rewarding past behaviour. Another potential explanation for gambling with higher stakes and making riskier bets could be the psychological value that the chips hold. The online condition obviously did not involve the participants holding the chips, whereas the offline condition required the participants to place chips themselves. Although there is little research into the difference in psychological value of online chips and offline chips, it could perhaps be argued that the value of placing chips in an offline table game is higher than placing virtual chips in an online game. There is little empirical research suggesting this to be the case, but the participants in this experiment often said they felt it was easier to gamble with online chips compared to chips offline. Further research on how players value chips online versus offline may prove useful. The results of this study provided some support for the hypothesis that online gamblers are more likely to make riskier bets (i.e., betting on those positions that have a lower probability of winning). The results showing that those who gambled alongside someone else placed higher

Int J Ment Health Addiction

stakes per roulette spin supported previous research suggesting that social facilitation has an effect on gambling behaviour (Rockloff 2008). Rockloff (2008) suggested there was a possible need to impress fellow players that was achieved by the placing of larger bets. Outside of the gambling field, a meta-analytic study of research evidence by Mullen et al. (1997) concluded that (in general), there is a small but significant tendency for others presence to increase arousal when measured using psychophysiological techniques. However, the authors also noted that this effect was moderated by: (a) the type of situation, and (b) which other people were present. Perhaps when people gamble with others present, it increases their own arousal that in turns facilitates higher and/or riskier bets. Such an observation suggests that future researchers could measure players psychophysiological reactions while gambling with others as opposed to previous research that has measured arousal levels while playing alone (Griffiths 1993). With past research suggesting that both social facilitation and medium (i.e., online or offline) may increase players risk-taking behaviour, a priori it would suggest there should be a significant interaction between the two. Results of the two conditions (amount of money bet and percentage of inside bets) tended to support this hypothesis. The mean scores clearly showed that players gambled the most when they were online and while gambling with someone else. Although the results are interesting, the study is not without a number of acknowledged limitations. One of the main issues concerns the use of chips with no monetary value (rather than real money or chips). For ethical and practical reasons, it was not possible to conduct such a study using participants own money in this experiment. The problem with this highlights Griffiths (2003) claims that the psychological value of virtual representations of money is less than that of real money. This suggests that participants in this experiment would be less likely to take risks with real money than with online money. If this is the case, then the psychological value of gambling with chips that have no monetary value should be considerably less than the value of real money, therefore players would be willing to take larger and more extreme risks with non-monetary chips. However, any use of real money would have other ethical implications that would have made the study unfeasible in the first place. Overall, the fact that participants used non-monetary chips in all conditions to gamble made the psychological value of the chips the same in all conditions. Therefore, although it could be argued that participants would gamble and potentially lose at a much higher rate than if they were using real money, it was the most practical and pragmatic way in which the gambling could be researched. It should also be noted that roulette is typically played alone in online casinos. The data presented here suggest that if online casino operators were to introduce multi-player gaming and/or a chat facility to allow players to play and chat at the same time, it may perhaps create conditions for increased gambling by players. Another potential methodological limitation is that each participant only participated in one conditioneither alone (online and offline) or with another person (online and offline). To compare how a persons betting directly changed due to social facilitation, future research studies could involve gamblers participating in offline only (alone and with others) or online (alone and with others). Here it could be seen how social interaction (rather than the medium) influenced a particular persons playing (i.e., within-subjects design). Such a study might highlight that gamblers may just have been conservative or risky players in the first place. It could perhaps be argued that the results presented in this study may not be due to social facilitation but may be because of the disinhibiting medium and/or the fact that participants gambled with virtual forms of money (which is an interesting observation in itself). Another limitation was that the study used student gamblers rather than members of the general public. However, there is some empirical evidence that when it comes to online gambling, students are a vulnerable group. Studies using self-selected samples of online

Int J Ment Health Addiction

student gamblers have shown high rates of problem gambling (Wood et al. 2007; Griffiths and Barnes 2008). Given these findings, this would appear a good rationale for carrying out research on this particular segment of the population. Despite these acknowledged methodological limitations and flaws, it should also be noted that there has been no previous gambling research examining interactions between gambling medium and social facilitation. Therefore, the study is of value and is something that further research can build on and improve. Overall, this pilot study suggests that gambling medium and social facilitation may be important components of determining stake size in a gaming environment or for the design of an online gambling website. The combination of other aspects of the environment needs to be accounted for through further testing in future research.

References
Griffiths, M. D. (1993). Tolerance in gambling: an objective measure using the psychophysiological analysis of male fruit machine gamblers. Addictive Behaviors, 18, 365372. Griffiths M. D. (2001). Internet gambling: preliminary results of the first UK prevalence study. Journal of Gambling Issues 5: Available at: http://www.camh.net/egambling/issue5/research/griffiths_article.html (Last accessed January 4, 2009). Griffiths, M. D. (2003). Internet gambling: issues, concerns, and recommendations. CyberPsychology & Behavior, 6, 557568. Griffiths, M. D., & Parke, J. (2002). The social impact of internet gambling. Social Science Computer Review, 20, 312320. Griffiths, M. D., & Parke, J. (2003). The environmental psychology of gambling. In G. Reith (Ed.), Gambling: Who wins? Who Loses? (pp. 277292). New York: Prometheus Books. Griffiths, M. D., & Parke, J. (2007). Betting on the couch: a thematic analysis of Internet gambling using case studies. Social Psychological Review, 9(2), 2936. Griffiths, M. D., & Barnes, A. (2008). Internet gambling: an online empirical study among student gamblers. International Journal of Mental Health and Addiction, 6, 194204. Griffiths, M. D., Parke, A., Wood, R. T. A., & Parke, J. (2006). Internet gambling: an overview of psychosocial impacts. Gaming Research and Review Journal, 27(1), 2739. Griffiths, M. D., Wardle, J., Orford, J., Sproston, K., & Erens, B. (2009). Socio-demographic correlates of internet gambling: findings from the 2007 British Gambling Prevalence Survey. CyberPsychology & Behavior, 12, 199202. Hardoon, K. K., & Derevensky, J. L. (2001). Social influences involved in childrens gambling behavior. Journal of Gambling Studies, 17, 191196. Ladd, G. T., & Petry, N. M. (2002). Disordered gambling among university-based medical and dental patients: a focus on internet gambling. Psychology of Addictive Behaviors, 16, 7679. Lesieur, H. R., & Blume, S. B. (1987). South oaks gambling screen (SOGS): A new instrument for the identification of pathological gamblers. The American Journal of Psychiatry, 144, 11841188. Meyer, G., Hayer, T., & Griffiths, M. D. (2009). Problem gambling in Europe: challenges, prevention, and interventions. New York: Springer. Mullen, B., Bryant, B., & Driskell, J. E. (1997). Presence of others and arousal: an integration. Group Dynamics: Theory, Research, and Practice, 1, 5264. Rockloff M. J. (2008). Crowd size impacts on gaming intensity: The moderating effect of social facilitation on EGM betting behaviour. Paper presented at National Association for Gambling Studies Annual Conference, Adelaide. December. Rockloff, M. J., & Dyer, V. (2007). An experiment on the social facilitation of gambling behavior. Journal of Gambling Studies, 23, 112. Sproston, K., Erens, B., & Orford, J. (2000). Gambling behaviour in Britain, results from the British gambling prevalence survey. London: National Centre For Social Research. Williams, R. J., & Wood, R. T. (2007). Internet Gambling: a comprehensive review and synthesis of the literature. Report prepared for the Ontario Problem Gambling Research Centre, Guelph, Ontario, Canada Wood, R. T. A., Griffiths, M. D., & Parke, J. (2007). The acquisition, development, and maintenance of online poker playing in a student sample. CyberPsychology & Behavior, 10, 354361.