The residual costs of multitasking: Causing trouble down the road
archive: archived pipeline: cataloged verified
Summary
Turrill, Coleman, Hopman, Cooper, & Strayer (2016, HFES Proceedings) measured residual switch costs after intermittent voice interactions with three smartphone digital assistants (Apple Siri, Google Now, Microsoft Cortana) in an instrumented vehicle (2015 Chevy Malibu, Chrysler 200c) on a 4.3-km residential loop with 7 all-way stops. N=65 drivers (30 female, age 21-68, mean 41.3) wore a head-mounted DRT and pressed a button-cued SDA via earbud at marked intersections, completing five conditions: baseline, three SDA tasks, and OSPAN. On-task DRT RT matched OSPAN levels; post-hoc analysis of the on-to-off transitions revealed residual workload that decayed gradually back to baseline over ~18 seconds after the SDA interaction ended.
Key finding
Cognitive workload from voice-based smartphone digital assistants does not return to baseline immediately when the interaction ends - reaction-time costs persist for ~18 seconds after the conversation, so the safety footprint of an in-vehicle voice command extends well past button release.
Methodology
on_road
Sample size: 65
Quality score: 5 / 5