Constructing the decision environment from memory: how overweighting extreme events biases choice

点击翻译按钮获取中文摘要

Decision-making involves choosing between alternatives by considering the probability of each outcome and how rewarding it will be. In some situations, such as reading the side effects of a medication, these probabilities and outcomes are listed and in others, such discovering a new food allergy, we need to learn them from experience. What we remember influences the decisions we make. Some items, such as extremely rewarding events, are more likely to be remembered than others and therefore more likely to influence our decisions. How might memory for particular events influence decision-making? Imagine that you are on holiday and you go to buy milk. You can either turn left and walk 3 minutes to the newsagent or turn right and walk 3 minutes to the supermarket. Let's assume that the last four times you have been to a newsagent you paid the following for milk [£1.50, £0.95, 1.40, £1] and at the supermarket you paid [£1.23, £1.26, £1.25, £1.22]. So, on average you paid the same at both shops. To decide which way to turn, you might sample a few experiences from memory. Which items do you sample and how do you then compare these items? You may think of the time you found a bargain and paid £0.95 and happily turn right. However, you may just as easily end up over paying so you would do well to think of the time you paid £1.50. What you decide depends on the processes which have yet to be examined in the same context: 1) which events you originally encoded; 2) how many experiences you sample from your memory and their "value"; and 3) how you compare the items in your sample. Research has indicated that people are more likely to rely on extreme information when making decisions and this can increase risk-seeking behaviours. The precise memory mechanism underlying this phenomenon are not yet understood. This project investigates how our memory for rewarding events, including extreme events, contributes to decision-making and risky choices. The proposed research will address this question using a suite of theory driven experiments supported by computational models. In a series of experiments, we will assess how healthy individuals encode, store and retrieve rewards in memory and use these memories to make decisions. We will develop computer models to understand how memory guides risky-choice. The project will increase our understanding of the role of memory in risky decision-making and help us to identify novel approaches for therapeutic intervention. Many of people's everyday decisions and choices relating to health-related lifestyle, financial savings, purchasing behaviours and environmental choices are heavily influenced by memory for past experiences. This research will support the development of better, more effective choice architecture interventions. If we can develop a better understanding of how people retrieve information from memory, and how that retrieved information supports choice, we will be able to develop interventions that prompt or nudge memory to improve choice.