Australasian Science: Australia's authority on science since 1938

The Crocodile and the One-Armed Bandit

Credit: auremar

People who bet the maximum line and bet size on commercial poker machines in Australia lose $1400 per hour on average. Credit: auremar

By Matthew Rockloff & Nancy Greer

Researchers have introduced crocodiles to test how excitement influences gambling behaviour on poker machines.

Have you ever wondered how crocodiles affect your willingness to gamble? It seems like an unlikely question and an equally unlikely proposition. However, most people could easily accept that gambling is an exciting experience. Indeed, the experience of excitement is one of the principal attractions of gambling. Equally, contact with an ancient and dangerous predator could invoke excitement.

Researchers have spent some time documenting the excitement of gambling, which includes measurable aspects of physiological arousal such as increases in skin conductance. When people are physiologically aroused there are small changes in perspiration on the surface of their skin. Using a finger sensor, these changes can be measured since perspiration has salts that make it a good conductor of electricity.

While is seems somewhat obvious that gambling causes excitement in the form of physiological arousal, it is less obvious how such arousal might, in turn, affect people’s gambling choices. The source of arousal could be positive through the anticipation of winning, or negative through a concern about losing.

There is also a more fundamental problem of measurement. People find gambling exciting, and this excitement influences subsequent gambling. However, we are faced with a chicken-and-egg problem: how can we tell what we are measuring? People are already gambling when we try to understand how their excitement affects their gambling. Are we measuring how gambling causes excitement or how excitement influences gambling?

A Bridge Too Far

One way to solve this problem is to generate excitement, or physiological arousal, prior to gambling. Our experiment, which received the 2017 Ig Nobel Prize in Economics, was inspired by Dutton and Aron’s famous 1974 bridge study. In this study, an attractive female interviewer stood at the end of both a low bridge and, alternately, a high -bridge spanning the Calippo Canyon in Vancouver. This latter bridge is suspended 70 metres above a river and has a length of 140 metres, causing it to sway in the breeze and with foot traffic. In the study, male subjects crossed each of the bridges to be interviewed by an attractive young woman standing at the end.

There were several interesting findings, but perhaps the most remarkable came after the attractive interviewer gave each participant a card with her number “in case they had any questions about the study”. As Dutton and Aron predicted, more men called the interviewer who had crossed the high bridge than who had crossed the low bridge, presumably because they misattributed their physiological arousal (fear) from crossing the bridge as sexual attraction. Readers might wonder if they were just “different” types of men who chose to cross the high bridge, but Dutton and Aron noted that most people cross the suspension bridge since it is the main attraction at the park.

Enter the Crocodile

Gambling has a deep history in Australia, and has embedded itself into our culture. A common saying is that Australians will bet on two flies running up a wall. However, as an ancient wind-swept continent, there are few high suspension bridges in Australia, and none in Central Queensland where we were working when we chose to investigate this problem. We needed an alternative.

Saltwater crocodiles prefer warm climates, and have a typical range that ends at Gladstone’s Boyne River, just south of Rockhampton where we were based. A major tourist attraction near Rockhampton is the Koorana Crocodile Farm. This farm raises saltwater crocodiles for their hides and meat, but also provides educational tours. This became our substitute for the suspension bridge: a way to produce mild physiological arousal without participants being aware of the source of their excitement.

The Experiment

We asked people to gamble on a laptop-simulated poker machine with an initial $20 bank of real money that we provided to them. People were told that they could quit playing the poker machine at any time. They would keep the amount shown on the machine, which would be paid to them immediately in cash.

Tourists would arrive at the crocodile farm for a 1-hour tour. At random, some tours were designated for the “test” condition and others for the “control”.

In the test condition, a tourist chosen at random would be asked to gamble after the tour. At the end of the tour, people were given the opportunity to hold a live 1-metre-long juvenile crocodile. Its mouth was taped shut to avoid injury, but its sharp claws and protruding teeth were exposed. Tourists were advised to take care when handling the animal.

Immediately after holding the crocodile we approached the chosen participant to gamble on our simulated poker machine. For the control condition we instead approached the chosen participants immediately before the tour, and prior to them having any exposure to saltwater crocodiles.

Our results were somewhat complex, because we needed to account for two additional factors that might affect people’s gambling under the induced condition of excitement.

First, some of our participants, despite being only tourists, had pre-existing gambling problems. It seemed reasonable to assume that different experiences with gambling, and in particular experience with gambling problems, might affect people’s emotional motivations for gambling.

Second, we assumed that people might interpret their physiological states of excitement in different ways. Some people might interpret their feelings as a positive mood state, whereas others might subjectively feel more negative. Although few people were in a “bad mood” per se, this was a tourist experience so at least some could more easily accept their subjective feelings as a positive feeling of excitement.

As a consequence, we surveyed people’s gambling problems and mood state immediately after playing our game.

Unbeknown to our participants, our game was rigged with an early sequence of wins followed by an indefinite string of losses. People could win up to $62 in our experiment, although most walked away with far less. On average people won around $40 in the first 20 spins, but thereafter gambled away most or all of it. The game ended automatically if people reached a bank of $0, although the players were allowed (and encouraged) to quit at any time.

Our outcomes included several indicators of gambling intensity; that is, player actions that would tend to magnify losses over the long-run in commercial gambling. These measures included bet size, bet speed and persistence while losing.

Positive Excitement Magnifies Bet Size for Problem Gamblers

The findings of our study related mainly to bet size rather than the other indicators of gambling intensity. People with pre-existing gambling problems placed larger bets after holding the crocodile, as long as we also found that they were in a positive mood. In contrast, gamblers with pre-existing problems who were in a negative mood had lower average bet sizes after holding the crocodile compared with the control set of participants who played the game before the tour.

That Winning Feeling

Our experiment shed new light on the emotional determinants of gambling. By producing excitement in people immediately prior to gambling, we were able to observe its effects on their gambling choices.

Our most powerful results addressed people with pre-existing problems. People can interpret their physiological arousal in different ways, and our experiment showed that the interpretation of arousal is the key to understanding how that will affect their choices – and specifically their choice of bet size.

Gamblers who bet more on each spin can have dramatically larger losses than those who bet less. In fact, people who bet the maximum line and bet size on commercial poker machines in Australia can lose an average of $1400 per hour, and (if unlucky) potentially more.

Two-Factor Theory of Emotion

The results in this experiment fit neatly with Schachter and Singer’s two-factor theory of emotion. According to this, people experience arousal from something in the environment, such as a crocodile. The arousal is the source of their emotional experience, but they need to scan the environment to interpret what that emotion means. Is it a fun feeling of excitement, or a fear?

Mild arousal can sometimes be misdirected to another source. In the bridge study, the source of the arousal was the high bridge but the mis­attributed reason for the emotion was the attractive interviewer approaching the subject at the end. Similarly, in our experiment the source of the arousal was the crocodile, but the reason for this excitement was mis­attributed to the simulated poker-machine gambling experience.

Feeling Lucky Can Be Dangerous

Poker machines, unlike some other forms of gambling, are completely random. No skill element is present, and cannot be – at least in Australia – by regulation. Using “feelings” to regulate play is inherently irrational, and can often put people at unnecessary risk for gambling-related harm.

Our quirky experiment showed that emotions can influence our gambling decisions, and these influences sometimes are not in our best interests.

The Ig Nobel Prize

Our experiment was awarded the 2017 Ig Nobel Prize in Economics at Harvard Unversity’s Memorial Hall. This long-running awards ceremony celebrates research that “at first makes you laugh, and then makes you think”. Other celebrated research in 2017 included the Physics Prize for using fluid dynamics to probe the question “Can a cat be both a solid and a liquid?” and the Anatomy Prize for exploring “Why do old men have big ears?”.


Professor Matthew Rockloff is Head of the Experimental Gambling Research Laboratory at Central Queensland University, where Nancy Greer is a PhD student.