Central and Peripheral Cues to Difficulty in a Dynamic Task

Abstract

Background Subjective workload assessments (e.g., NASA-TLX, SWAT, WP) rely upon judgments of difficulty (JODs) and workload to determine how challenging it is for an operator to complete a task. While it is likely that subjective workload assessments are an aggregate product of many cues to difficulty, including internal feelings of effort, that become available during task completion (Cordes, 1993; Kahneman, 2000), moment-to-moment judgments may evolve as additional external cues become available. That is, JODs are likely based to some degree upon internal (privileged) cues and integral task characteristics (central cues), but are also influenced by peripheral cues that serve as a proxy for task difficulty. These dynamic and multifaceted cues affect the choice to initiate tasks (e.g., Mitchell, 2017), comply with safety procedures (e.g., Sigurdsson, Taylor, & Wirth, 2013), and allocate resources (e.g, Kurzban, Duckworth, Kable, & Meyers, 2013), among other things. Identifying common cues that inform JODs contributes to workload theory and defines avenues for workload intervention. Method We developed a 40-minute videogame task that could be used as a laboratory to test people’s sensitivity to two peripheral cues that are present during many tasks: performance-based feedback and time-on-task. These cues became available as participants completed videogame sessions that were identical in instructions and objective, but differed in the source of their difficulty. Specifically, participants were assigned to one of five conditions in which enemy behavior changed along a continuum with respect to: how quickly enemies moved, how much damage enemies dealt to players, how many shots it took to eliminate an enemy character, how quickly enemies appeared in the level, and how close players could be to enemy characters before being chased. These random changes in task speed, feedback, physical effort, timing, and coordination occurred at the beginning of every level (i.e., block of trials). Participants’ JODs were collected at the beginning of each level and every subsequent three minutes by asking them to indicate whether the task was easier or harder than before. Performance-based feedback Participants began each level with 100 hit points of health and received visual and auditory feedback when they came in contact with an enemy character. Crucially, enemy attacks caused participants to lose hit points; when health dropped below 0, the character died and participants had to wait 30 s before resuming the game from a random location within the game space. Self-reported JODs Participants’ JODs were collected at the beginning of each level and every subsequent three minutes by asking them to indicate whether the task was easier or harder than before (Böckenholt, 2004). The task was paused 3s before and after the pop-up window appeared to reduce the performance costs associated with task interruption (Altmann & Trafton, 2007).