Measuring Cognitive Workload in Automated Knowledge Work Environments
Automation, as defined by Parasuraman et al. (2000, p. 287), is a “device or system that either partially or fully, accomplishes a function that was previously, partially, or fully accomplished by a human operator.” Traditionally, automation was introduced to (physical) work environments to alleviate workload associated with tedious and repetitive tasks. Over the past few decades, automation has begun to augment knowledge work, which includes high-level cognitive activities. As automated systems expand to perform skill-based tasks, the work required of humans is inevitably altered, potentially affecting their cognitive workloads. Years of research has shown that automation can reduce cognitive workload, but other work suggests that cognitive workload may increase or remain unchanged when automation is introduced. These conflicting results prompt the need for further investigation to better delineate the relationship between automation and cognitive workload.
A plethora of factors may explain why the relationship between automation and cognitive workload is inconsistent. Therefore, this research takes steps toward addressing knowledge gaps within the human-automation interaction literature related to understanding how automation used in knowledge work environments affects peoples’ task completion. Specifically, this work investigates how two moderators, task complexity and age, influence the automation and cognitive workload relationship. These moderators were of interest for two reasons. First, task complexity, which occurs when the structure of a task imposes demands on a person’s cognitive processes, increases the demands of a task, which can result in the use of more cognitive resources. Second, age is of interest because advanced technologies are increasingly being utilized by a wide user demographic, particularly the rapidly-growing older adult population.
The goals of this dissertation were achieved by employing both qualitative and quantitative methods to examine how (1) automation is assessed in knowledge work environments, (2) automation affects cognitive workload, and (3) task complexity and age moderate the relationship between automation and cognitive workload. These goals were first addressed via the construction of a conceptual framework that describes the effects that task complexity and age have on the relationship between automation and cognitive workload. Next, a systematic review of the human-automation interaction literature in knowledge work environments was performed to examine researchers’ use of cognitive workload measures. Finally, a controlled-laboratory experiment and a scenario-based survey were conducted to collect data from people of different ages about how task complexity and age influence the relationship between automation and cognitive workload.
Findings from the systematic literature review showed that researchers primarily employ subjective and performance measures to assess cognitive workload. Results from the laboratory experiment suggested that automation improved measures of cognitive workload. Also, task complexity negatively affected the relationship between automation and cognitive workload, but age was not found to be a moderator. The scenario-based survey revealed that task performance was similar among younger, middle-aged, and older adults. However, younger adults had a more favorable opinion of automation compared to both middle-aged and older age groups.
Overall, this research (1) enhances our knowledge of the relationship between automation and cognitive workload, (2) informs the design of future human-automation studies with strategically selected task types and measurement choices, based on patterns that emerged from the literature review, and (3) can ultimately guide designers in better developing technologies to support people in performing various activities in their work and leisure environments.
- Doctor of Philosophy
- Industrial Engineering
- West Lafayette