Workload Capacity: A Response Time-Based Measure of Automation Dependence

Area of Science:

• Human-Computer Interaction
• Cognitive Psychology
• Decision Science

Background:

• Traditional performance metrics like accuracy and mean response times (RTs) inadequately capture human-automation interaction dynamics in speeded tasks.
• RTs can obscure an operator's strategic adjustments and the distinct contributions of human and automation components.
• A novel measure, workload capacity, derived from RT distributions, offers a more nuanced assessment of team processing efficiency.

Purpose of the Study:

• To introduce and validate workload capacity as a measure of human-automation team processing efficiency.
• To identify operators' automation usage strategies in a speeded decision-making context.
• To assess automation dependence and performance changes under aided conditions.

Main Methods:

• Participants engaged in a speeded probabilistic decision task, both with and without an automated decision aid.
• Workload capacity measures, specifically COR(t) and CAND(t), were calculated using empirical response time distributions.
• Analysis focused on quantifying processing efficiency and detecting strategic operator adjustments.

Main Results:

• Workload capacity measures indicated that the automated aid significantly accelerated human participants' response times.
• Evidence suggests participants did not adjust their decision times in anticipation of the automated aid's input.
• The findings highlight the impact of automation on task completion speed.

Conclusions:

• Workload capacity is a sensitive and informative metric for evaluating human-automation performance in speeded tasks.
• This measure effectively quantifies operator dependence on automation.
• The study underscores the utility of workload capacity for understanding team dynamics in time-constrained decision-making.