To Trust or Not to Trust? A Simulation-Based Experimental Paradigm

Knodler Jr., Michael A.; Christofa, Eleni; Hajiseyedjavadi, Foroogh; Tainter, Francis; Campbell, Nicholas · 2018 · ROSA P / Safety Research Using Simulation (SAFER-SIM) University Transportation Center

archive: archived pipeline: cataloged verified

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

Summary

This study investigates how system performance and individual differences influence driver trust in automated vehicles, addressing the critical issue that inappropriate trust levels (over-trust or under-trust) can negate the safety benefits of automation. The research aims to understand the dynamics of initial dispositional trust and history-based trust, specifically examining how system failures affect user reliance and behavioral responses. The researchers conducted a simulation-based experiment with 80 participants aged 20–30, recruited from the University of Massachusetts Amherst. Using a fixed-base driving simulator with a 330-degree field of view, subjects drove eight scenarios in both manual and Level 2 automated modes. Participants were assigned to one of five groups based on system reliability: 100% performance, or 88% and 75% reliability with either pedestrian interaction failures or stop-sign intersection failures. Data collection included physiological measures (heart rate and heart rate variability via chest sensors), eye-tracking, video-recorded hand and foot movements, and psychological questionnaires assessing trust and workload. The experimental design utilized a between-subjects approach with counterbalanced scenario sequences to isolate the effects of failure type and frequency. Results indicated that any system failure significantly increased the probability of unnecessary disengagement in subsequent scenarios where the system performed correctly, particularly during intersection interactions compared to pedestrian scenarios. Physiological analysis revealed significant changes in heart rate variability following failure events, correlating with increased anxiety. Specifically, ANOVA tests showed significant HRV differences for scenarios immediately preceding and following failures in groups experiencing one pedestrian failure, one vehicle failure, and two vehicle failures. Behavioral metrics showed an expected increase in hand and foot movements during failure scenarios, with foot movements remaining elevated longer in groups experiencing two pedestrian failures compared to two vehicle failures. The study concludes that system failures degrade trust calibration, leading to distrust manifested through unnecessary disengagement and heightened physiological stress, even when the automation is functioning correctly. These findings highlight that user perception of performance, rather than actual performance alone, drives trust dynamics. The results provide empirical evidence for designing human-automation interaction models that account for the lasting impact of failures on driver behavior, suggesting that trust recovery mechanisms are essential for the safe deployment of automated driving systems.

Key finding

Any type or level of system failure significantly increases the probability of unnecessary disengagement during intersection interactions when the system's response is appropriate.

Methodology

simulator

Sample size: 80

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 bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).

StageOutcomeToolModelPromptAttemptsCompleted
discover success rosap 2 2026-05-23
archive success 1 2026-05-23
extract success cached 2 2026-06-10
clean success 1 2026-06-01
chunk success 1 2026-06-01
embed success 1 2026-06-02
enrich success 1 2026-05-23
promote success 1 2026-05-23
summarize success llm qwen3.6-27b-prismaquant summ-v5 3 2026-06-10
tag success vector_similarity 19 2026-06-11
verify partial 2 2026-06-10

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

Topics

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

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).