As human beings, we often have to decide between rewards that are smaller, but available sooner, and rewards that are larger, but that we have to wait for longer. We wonder, e.g., whether we should spend our money now or save it for pension. Such so-called delay discounting decisions have usually been studied with a focus on the outcome of the decision process, leading to a variety of discounting models (see Doyle, 2013 for an overview). While these models offer a good description of the average outcome of the decision process -the final choice -they mostly leave open how the decision process itself unfolds in time. Here, we complement the outcome-focused approaches to delay discounting with a process-based approach. We show how a decision process model, namely, an attractor model, can extend outcome-focused discounting curves into the temporal dimension. We derive temporal predictions from the model and validate them in an experiment in which we manipulate the order of presentation of options and combine this with a continuous behavioural measure of the decision process, i.e., mouse tracking.
Delay discountingEarly authors acknowledged that time comes at a cost and assumed that, rationally, 1 the value of delayed options should be discounted following an exponential discounting function (Samuelson, 1937). In reality, however, individuals often show a stronger discounting of rewards, especially for small delays (Green, Fristoe, & Myerson, 1994;Myerson & Green, 1995). A wide range of research aimed to close this explanatory gap between normative models and empirically determined models of delay discounting (Frederick, Loewenstein, & O'Donoghue, 2002).Step by step: Harvesting the dynamics of delay discounting decisions Stefan Scherbaum 1 , Simon Frisch 1 and Maja Dshemuchadse 2 Abstract People show a tendency to devalue rewards when they are delayed in time. This so-called delay discounting often happens to an extent that seems irrational from an economical perspective. Research studying outcomes of delay discounting decisions has successfully derived descriptive models for such choice preferences. However, this outcome-based approach faces limitations in integrating the influence of contextual factors on the decision. Recently, this outcome-centred perspective on delay discounting has been complemented by a focus on the process dynamics leading to delay discounting decisions. Here, we embrace and add to this approach: We show how an attractor model can extend discounting descriptive discounting curves into the temporal dimension. From the model, we derive three predictions and study the predictions in a delay discounting experiment based on mouse tracking. We find differences in discounting depending on the order of option presentation and more direct movements to options presented first. Together with the analysis of specific temporal patterns of information integration, these results show that considering the continuous process dynamics of delay discounting decisions and harvesting them with continuous behaviou...