Citation

Frazier S, Pitts BJ, McComb S. Cogn. Technol. Work 2022; 24(4): 557-587.

Copyright

(Copyright © 2022, Holtzbrinck Springer Nature Publishing Group)

DOI

10.1007/s10111-022-00708-0

PMID

unavailable

Abstract

Traditionally, automation was introduced to alleviate workload associated with tedious and repetitive tasks. Recently, automation is being used to augment knowledge work, which includes high-level cognitive activities. As automated systems are being extended to perform skill-based tasks, the work required of humans may be altered, potentially affecting their cognitive workloads. Researchers have investigated the influences of automation on cognitive workload across different domains and tasks by assessing changes in task performance, perceived (subjective) workload, and physiological states. A major challenge in comparing results and drawing inferences across studies is that a profusion of measures is often used to assess cognitive workload. The experimental tasks employed across many domains further complicates synthesizing findings. Thus, the aim of this review is to examine how cognitive workload is assessed when at least two different measures of cognitive workload are used in research focused on human-automated knowledge work. To accomplish this aim, the various approaches employed to measure cognitive workload were first summarized. Then, automated and cognitive experimental tasks were classified, utilizing existing frameworks, to identify associations, dissociations, and insensitivities across task types. Finally, recommendations were provided for aligning task types, study designs, and measurement selections, along with expanding the types of tasks and measures used when studying automation applications supporting knowledge work.

Language: en

Keywords

Automation; Cognitive workload; Human-automated systems; Knowledge work; Measurement