The persistence of distraction: The hidden costs of intermittent multitasking
DOI: 10.1037/xap0000388
archive: archived pipeline: cataloged verified
Summary
Two-experiment study (J Exp Psychol Applied, 2022) examining residual costs of intermittent multitasking using the DRT (ISO 17488). Participants performed a pursuit-tracking task (Exp 1, N=32) or drove a high-fidelity simulator (Exp 2, N=47, Easy vs Hard traffic) while periodically counting backwards by 1's or 3's in 20-s on-task epochs followed by 30-s (Exp 1) or 45-s (Exp 2) off-task recovery intervals. DRT RT and Hit Rate, plus pupil diameter and forward-roadway gaze (Exp 2), tracked cognitive load. Multitasking costs persisted well after the counting task ceased, decaying as a negatively accelerated power function and remaining above baseline throughout the 30-s (Exp 1) and 45-s (Exp 2) off-task intervals. Costs were larger for the harder 3's task and for the Hard driving environment. Eye-tracking did not show oversampling of the forward roadway during recovery.
Key finding
Cognitive workload from a brief secondary task (20-s counting backwards) does not switch off when the task ends: DRT-indexed costs persisted for the full 30-45 s off-task interval, decayed as a power function, and scaled with task difficulty and driving demand. Pupillometry mirrored the DRT pattern while eye-tracking showed no recovery-related visual scanning, supporting a proactive interference (lingering working-memory contents) over a situational-awareness-recovery account.
Methodology
Exp 1: 32 participants performed a pursuit-tracking task with a 3-min single-task baseline followed by 10 repetitions of {20-s counting backwards by 1's, 30-s recovery, 20-s counting by 3's, 30-s recovery}, repeated twice. Exp 2: 47 participants drove a DS-600 fixed-base simulator on a 19-mile loop in Easy (~500 veh/lane/hr) and Hard (~1700 veh/lane/hr) scenarios, with the same counting/recovery structure but 45-s recovery and ~26-min drives. DRT (ISO 17488) with vibrotactile stimulus delivered the workload probe (RT, Hit Rate); Exp 2 added Seeing Machines Fovio eye-tracking for pupil diameter and gaze ROIs. Analyses used planned t-test contrasts of 3-s off-task bins vs single-task baseline and vs on-task average, plus power-function fits.
Sample size: Exp 1: N=32 (16 M, 16 F, mean age 25.8). Exp 2: N=47 (22 M, 25 F, mean age 23.0).
Quality score: 5 / 5