Stake and chips: Gambling spend and the psychology of interpretation

Although a number of researchers in the field have stated that data about expenditure on gambling is important to collect when doing prevalence surveys, getting accurate and reliable data is not easy to do. The question ‘How much do you spend on gambling?’ appears simple to answer but can be interpreted in many different ways. For instance, consider the following scenario used by Professor Alex Blaszczynski and colleagues at the University of Sydney:

“You recently decided to gamble $120 on your favourite form of gambling. You initially won $60 but then following a bad run of luck, lost $100. Feeling tired, you decided to leave and return home”

When participants in the study were given this scenario above, and asked “How much did you spend on gambling?” they made a number of different interpretations. There are four basic interpretations that ‘spend’ could relate to:

• Stake: This refers to the amount staked (i.e. the amount bet on an individual event, such as a football match, a fixed odds betting terminal or a lottery ticket).
• Outlay: This refers to the sum of multiple bets risked during a whole gambling session.
• Turnover: This refers to the total amount gambled, including any re-invested winnings.
• Net expenditure: This refers to the amount gambled minus any winnings.

In this particular study, approximately two-thirds of the participants (64%), answered $40 (i.e., net expenditure) in the scenario above [i.e., $120-($120+$60-$100)]. Around one-sixth of the participants (17%) answered $120 (i.e., stake). A small number of participants answered $160. Here the participants reasoned the spend was equal to $120+$100-$60. Alternatively some answered $100 that equated to the amount lost. Finally, a very small number of participants (n=5) answered $180 (i.e., turnover), where the participants reasoned that spend was equal to investment plus winnings.

There are also issues surrounding what constitutes an individual session (especially if the person gambling goes to the toilet or has a snack or drink between or during a gambling episode). What this simple study shows is that questions relating to expenditure need to be very precise. Blaszczynski and his colleagues argued that the most relevant estimate of gambling expenditure is net expenditure, as it reflects the actual amount of money the gambler has gambled, and also represents the true cost of gambling to the individual. In the 2007 British Gambling Prevalence Survey (BGPS), participants who had spent money on gambling in the past seven days were first asked for each activity that they had gambled on. “Overall, in the last seven days did you win or lose money?” To this particular question the gamblers could either answer that they lost, won, broke even, or were still awaiting the result. If gamblers had lost money they were asked how much, and were asked to tick one of six boxes indicating the total amount lost. Similarly, if gamblers had won money they were asked how much, and could tick one of six boxes indicating the total amount won. They were also asked to what extent the previous week’s gambling activity had been typical.

The results relating to net expenditure were interesting and perhaps somewhat predictable based on what has been reported in previous literature. Gamblers appeared to over-estimate how much they had won in the previous week, meaning that net expenditure was ‘positive’ on many of the gambling activities (i.e. on these activities, gamblers claimed to have won more than they had lost). A similar finding was also reported in the previous [1999] BGPS. Given that all sectors of the gaming industry make ‘considerable profits’, the results in the BGPS study clearly show that many gamblers do not appear to be making a realistic assessment of their previous week’s spending.

However, this does not necessarily mean that they are ‘lying’, as there is a lot of evidence that gamblers over-estimate winnings and under-estimate losses, due to cognitive biases and heuristics like the ‘fixation on absolute frequency bias’ (using absolute rather than relative frequency as measure of success), concrete information bias (when concrete information such as that based on vivid memories or conspicuous incidents dominates abstract information such as computations or statistical data), and/or flexible attributions (the tendency to attribute successes to one’s own skill and failures to other influences). In short, winning experiences tend to be recalled far more easily than losses (unless the losses are very substantial and have a major detrimental effect on the day-to-day functioning of the individual).

Remembering wins and discounting losses is a consistent finding in the gambling literature. This is more likely to occur on those gambling activities that are played several days a week, rather than those activities that are engaged in once a week such as the National Lottery Draw and the football pools. It is in these latter activities that participants are more likely to have accurate recall of wins and losses, as the weekly outlay is usually identical every week (e.g. buying two lottery tickets every week or being part of a lottery syndicate). The results in the 2007 BGPS do indeed seem to indicate this is the case, with activities such as the National Lottery Draw, and the football pools, reporting weekly net losses.

Furthermore, there are other more general effects (like social desirability) that may be skewing the results in a more socially positive direction. There is also the general observation that people tend to overestimate positive outcomes and underestimate negative ones that has been applied to the psychology of gambling. Most of the positive net expenditures were fairly modest, but on those gambling activities where skill has the potential to be used, the net expenditures were much greater (e.g. online poker as part of online gambling, blackjack as part of casino table games). The results showing that the smaller the number of participants gambling on the particular activity, the greater the overall net win claimed, highlights the fact that individual variability was likely to be more pronounced among lower numbers of participants. It is also likely that some of the activities do indeed include gamblers who genuinely win more than they lose (online poker being a good example). However, the number of people doing this regularly is likely to be relatively small, as there are always more losers than winners in such activities.

Dr Mark Griffiths, Professor of Gambling Studies, International Gaming Research Unit, Nottingham Trent University, Nottingham, UK

Further reading

Blaszczynski, A., Dumlao, V. & Lange, M. (1997). How much do you spend gambling? Ambiguities in survey question items. Journal of Gambling Studies, 13, 237-252.

Gilovich, T. (1983). Biased evaluation and persistence in gambling. Journal of Personality and Social Psychology, 44, 1110-1126.

Griffiths, M.D. (1994). The role of cognitive bias and skill in fruit machine gambling. British Journal of Psychology, 85, 351-369.

Griffiths, M.D. & Wood, R.T.A. (2001). The psychology of lottery gambling. International Gambling Studies, 1, 27-44.

Wagenaar, W. (1988). Paradoxes of Gambling Behaviour. Hove: Lawrence Erlbaum Associates.

Wardle, H., Sproston, K., Orford, J., Erens, B., Griffiths, M.D., Constantine, R. & Pigott, S. (2007). The British Gambling Prevalence Survey 2007. London: The Stationery Office.