Prospective memory in dynamic environments: The role of uncertainty

Abstract

Prospective memory (ProM) refers to remembering to perform a task in the future. Successfully remembering a ProM task entails remembering not only that something has to be done, but also when it has to be done, and what has to be done. Busy operational environments may require operators to rapidly switch between multiple safety-critical tasks. In dynamic environments, an operator’s inability to predict upcoming tasks may interfere with any or all aspects of ProM (i.e., remembering that something needs to be done, when it needs to be done, and what needs to be done). However, previous research has not manipulated uncertainty about the context of ProM execution influences ProM. The current dissertation uses a task-switching paradigm to investigate whether uncertainty about the context in which ProM tasks are executed would affect ProM success. In the certain condition, participants could predict the order of upcoming tasks. In the uncertain condition, participants experienced tasks in a random, unpredictable order. In Experiment 1, the ProM task consisted of remembering to perform an action after being presented with the ProM cue, but only after completing the current (0 delay), next (1 delay), or next two (2 delay) ongoing tasks. Participants in the certain condition were expected to use knowledge about the order of upcoming tasks to imagine the specific context in which the ProM response had to be executed. In contrast, participants in the uncertain condition were unable to encode a more specific context of ProM execution. Results showed that participants in the certain condition were significantly better at remembering when the ProM task had to executed, as well as what action had to be undertaken to execute the ProM task for the longest delay (2 delay). Encoding the specific context of execution seems to have helped automatize detection of the context for ProM execution, thus boosting ProM performance in the certain condition. Experiment 2 investigated whether specifying a concrete context of execution would similarly help ProM for participants in the uncertain condition. In contrast to Experiment 1 in which participants were told to remember to execute a ProM task after a specific number of ongoing tasks, in the second experiment, participants were told to remember to execute a ProM task, but only after they had completed a specific ongoing task (e.g., dose calculation). As in Experiment 1, Experiment 2 also varied the delay between ProM instruction and ProM execution: the specific ongoing task after which the ProM response had to be made was either the current task that participants were working on, or the next task, or two tasks later. Unlike Experiment 1, however, there was no difference on ProM performance between the certain and uncertain task order groups. Whereas Experiment 1 found that performing tasks in a predictable order helped in remembering ProM tasks that had to be performed after a specific number of tasks, Experiment 2 suggests that task order did not affect ProM execution if details about the context of execution were well-specified at the time of ProM encoding. The results of the dissertation may help inform how uncertainty at the time of encoding influences ProM in operational environments that are characterized by frequent and unpredictable task switching. Specifically, changing the framing of ProM so that the context of execution of future tasks is encoded in terms of details about the retrieval context rather than completing a specific number of tasks helps support ProM in unpredictable environments.