An Influence Model of the Human-Automation Team: Effects of Workload and Automation Reliability, Transparency and Degree

Wickens, Christopher D.; Sargent, R.; Walters, B. · 2023 · Ergonomics International Journal

DOI: 10.23880/eoij-16000312

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the need for a quantitative framework to predict human-automation team (HAT) performance, moving beyond traditional statistical significance to estimate the specific benefits or costs of automation features. The authors argue that current literature lacks models capable of quantifying how variables like automation reliability, transparency, and degree interact with task workload to influence performance metrics such as accuracy and speed. To address this, the paper presents an "Influence Model" that maps causal relationships between automation properties, intervening cognitive constructs (trust, dependence, situation awareness, and workload), and final performance outcomes. The study combines theoretical modeling with empirical validation. First, the authors define a model comprising 16 causal links connecting task difficulty, automation reliability, transparency, and degree of automation to HAT performance. Second, they conduct a meta-analysis of 50 studies to quantify the effect sizes of automation transparency on these intervening variables and performance metrics. Third, they reference an unpublished experiment on nautical collision avoidance to test the model’s predictions regarding the opposing effects of discrimination task difficulty on automation dependence. The meta-analysis reveals that automation transparency has a large positive effect on situation awareness (Cohen’s d = 1.06) and a large negative effect on error rates (d = -0.96), indicating improved accuracy. Transparency also moderately increases trust (d = 0.79) and dependence (d = 0.45) while reducing response times (d = -0.53), with negligible impact on workload. Crucially, the benefits of transparency differ by context: in routine tasks, transparency primarily improves accuracy, whereas during automation failures, it significantly reduces recovery time. The nautical experiment demonstrated that as task discrimination difficulty increased, automation reliability declined, leading to a proportional decrease in human dependence on automation. This result supported the model’s prediction that the link from task difficulty to reduced reliability and trust outweighs the direct link from difficulty to increased dependence. The significance of this work lies in providing a structured, quantifiable model for evaluating HAT performance. By distinguishing between routine and failure performance and assigning relative weights to causal links, the model allows designers to predict trade-offs in automation features. For instance, it highlights that while transparency improves routine accuracy, its most critical benefit may be reducing takeover time during failures in time-critical environments. The authors conclude that while the polarity of influences is well-established, further empirical validation is needed to refine the strength of these links, particularly regarding the impact of task difficulty.

Key finding

Automation transparency has a large positive effect on situation awareness and performance accuracy, while increasing task difficulty reduces human dependence on automation despite the increased need for assistance.

Methodology

meta_analysis

Sample size: 50

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via openalex_abstract on 2026-05-08.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• automation
• automation complacency bias
• trust calibration