Small Grants 2021-22

The AGRI Small Research Grant ‘The role of response options on human decision-making quality’ has enabled our laboratory to establish and validate a new n-response 2-outcome game (Dice Dual), which advances our understanding of competitive decision-making.

When assessing how the quality of decision-making changes as a function of environmental complexity, there can be confounds between the number of responses and the number of outcomes available. For example, Matching Pennies (e.g., Dyson, Musgrave, Rowe & Sandhur, 2020) requires individuals to pick from heads or tails, resulting in wins or losses. In contrast, Rock, Paper, Scissors (RPS; Sundvall & Dyson, 2022) requires individuals to pick from three responses, resulting in win, losses and draws. Thus, it is not clear whether the change in the number of response options, the number of outcomes, or, an interaction between the two, gives rise to differences in performance.

Our game Dice Dual represents a binary outcome game (wins, losses) with the flexible delivery of response options (currently 2, 4 or 6), using the selection of a die side. We have been able to leverage this new paradigm in the exploration of four research question:

a) Decision-making quality as a function of response options. Two experiments were carried out in which two response (Matching Pennies), three response (RPS), five response (Elemental; Brockbank & Vul, 2020) games were compared to a novel six response version of Dice Dual (Na & Dyson, in preparation). Specifically, we found that shift behaviour becomes more likely as the number of response options increases, and following losses relative to wins. In examining correlations in individual performance, there was some evidence that the five response game did not have similar properties to the other games, calling into question the generalizability of that game. In contrast, Dice Dual strongly correlated with two and three response games, validating the integration of this paradigm into the literature. Importantly, Dice Dual provided a unique signature of shift behaviour, allowing us to separate two classes of exploratory behaviuor (see below).

b) Taxonomy of exploration behaviour. In our game Dice Dual, both participants and computerized opponent pick one side of a die. According to one version of the game, if the sum of the dice are odd the participant wins but if the sum of the dice are even the computer wins. In this respect, we able to uniquely define two classes of exploratory behaviour: within-category and between-category shifting. In the case of between-category shifting, participants move from odd to even die side selection, thus representing exploratory behaviour relevant to the goal of the game. In contrast, within-category shifting, participants move within certain odd or even sets (e.g., 5 to 1 to 3). This latter form of shift behaviour is purely exploratory in nature as it is irrelevant to the goal of the game. Moreover, this form of goal-irrelevant exploratory behaviour appears somewhat independent of the surface complexity of the game, in contrast to the goal-relevant exploratory behaviour that increased as the number of responses also increased. This strongly suggests that different classes of exploration are under more or less cognitive control, highlighting aspects of performance that may or may not be the target of change in gambling behaviour.

c) Testing the generality of findings. In Zhang, Huynh & Dyson (in revision) we explored the novel goal-directed behaviour of performing as badly as possible, in an attempt to mitigate the detrimental impact of negative outcomes. We showed a double dissociation where actions following wins were more consistent during win maximization, but actions following losses were more consistent during loss maximization. Importantly, we were able to use Dice Dual as a fourth experiment in this series to show that our data was not limited to a single experimental paradigm. This broader distinction between goal-congruent and goal-incongruent feedback suggests that individuals are able to flexibly redefine their definition of ‘success’, and provide a novel reconsideration of the way we think about ‘losing’ in the context of gambling.

d) Integrated measurement of subjective risk. In understanding decision-making quality in competitive environments, it is equally important to assess how subjective perceptions of the environment impact on performance. For example, an individual may experience a high win rate and falsely attribute success to skill rather than luck (e.g., Roulette). A major problem in collecting data on subjective impressions is that there are independent from actual game responses. This naturally dissociates subjective impression from behaviour. In our Dice Dual game, we successfully leverage variation in die side selection as an integrated measurement of subjective risk that maintains game flow. Specifically, in Dyson (in revision) I developed a version of Dice Dual where the die side also contributed to the amount of points won or lost. The data clearly showed that participants were more risky directly after wins relative to losses, and, as their win rate increased. Thus, in a single response, participants indicate 1) objective knowledge of the game, and, 2) subjective confidence about their knowledge of the game.

The development of a novel new n-response 2-outcome game (Dice Dual) may offer unique insights into the measurement of gambling behaviour with respect to the integrated measurement of subjective risk (cf, priority area, public policy). Data regarding individual perceptions of gambling and gambling environments may be collected before- or after-the-fact, and independently from gambling behaviour itself. As described above in d) a single response in this game can represent both subjective and objective measures of game knowledge. Furthermore, the use of a flexible number of responses can be used to understand how the surface complexity of a gambling environment affects performance at the individual level, providing further information about when individuals may naturally be at their most cognitively vulnerable during gambling.

Our work was presented at provisional, national and international levels: the Alberta Gambling Research Institute Conference (AGRI 2022), the Canadian Society for Brain, Behaviour, and Cognitive Science (CSBBCS 2022), and the Meeting of the Society for Applied Research in Memory and Cognition (SARMAC 2023). Furthermore, the work generated by this grant has contributed to three empirical papers- two currently in revision, with a third currently in preparation.

The Dice Dual paradigm will now be central to a new, innovative touch-screen table technology currently installed in the Re:Cognition Lab. Via the careful mounting of a large touch-screen monitor to a mechanical arm, we have created a horizontally oriented touch-table rather than vertically oriented touch-screen thereby simulating social gambling environments where players physically stand across from one another. Thus, we now advance our research by having naturalistic competitive game play between individuals, while at the same time as maintaining the precise level of experimental control required in empirical work.

Three empirical papers - two currently in revision, with a third currently in preparation.

Dyson, B. J., Baik, L. & Na, E. (2023). Novel insights in competitive decision-making via a new n-response 2-outcome game. Poster presentation at the 14th Biennial Meeting of the Society for Applied Research in Memory and Cognition, 9th – 12th August, Nagoya, Japan.

Na, E. Baik, L. & Dyson, B. J. (2023). Goal-relevant and goal-irrelevant exploratory (shift) behaviour as a function of response options. Poster presentation at the Alberta Gambling Research Institute Conference, March 30th – April 1st, Banff, Canada.

Zhang, Y. & Dyson, B. J. (2023). Improving behavioural consistency as a function of expectation: Failure is not an inherently negative state. Poster presentation at the Alberta Gambling Research Institute Conference, March 30th – April 1st, Banff, Canada.

Na, E. & Dyson, B. J. (2022). Evaluating a novel 6-response zero-sum game against binary and non-binary spaces: Subjective complexity and the expression of stay behaviours. Poster presentation at the Annual Meeting of the Canadian Society for Brain, Behaviour, and Cognitive Science, 18th- 20th July, Halifax, Canada.

Zhang, Y. & Dyson, B. J. (2022). Improving behavioural consistency as a function of expectation: Failure is not an inherently negative state. Poster presentation at the Annual Meeting of the Canadian Society for Brain, Behaviour, and Cognitive Science, 18th- 20th July, Halifax, Canada.