Policies in sequential decision-making

Humans often need to make decisions that take future situations into account. This kind of sequential choice entails a trade-off between the present and the future, where decision-makers have to decide to accept or reject the currently available option. The approach-avoidance conflict, on the other hand, is a frequently used framework in psychology to study the behavioral trade-off regarding appetitive and aversive stimuli. This study investigates the influence of approach and avoidance conditions on sequential choice. Therefore, we developed experimental conditions in which participants should either approach or avoid environmental dangers in order to optimize their behavior. The main objective is to evaluate meta-decision making strategies under ecologically valid conditions. To accomplish this, we developed a game-like foraging paradigm with probabilistic outcomes where people could use different types of decision rules (or policies). While some policies would only account for one variable of our game environment, others would account for different variables depending on the time-point and the situation. Model comparisons allowed us to test whether participants would rely on a rather myopic/heuristic or a forward-looking policy. We then tested whether policies differed depending on our task’s approach and avoidance condition. Our findings suggest that people generally use a myopic strategy in sequential decision-making. However, we observe increased optimal behavior in situations where the environment bears higher dangers that should be avoided. We suspect the underlying mechanisms to be linked to an internal energetic trade-off between efficient and accurate meta-decision making depending on the risks imposed by the environment.