Trevo Quantitative Research Report
Introduction
This study explores how varying levels of task automation affect user performance in a lightweight, mobile-compatible environment using Trevo. Participants completed tasks under three conditions: fully manual execution, randomized automation, and self-selected automation. Results suggest that automation type and user agency significantly influence task outcomes. Trevo’s model emphasizes the user’s role in guiding automation to enhance both performance and satisfaction.
Methodology
A custom HTML grid-based game was used to simulate task performance across three modes: manual, randomized automation, and self-selected automation. The experiment used a within-subjects design, and each participant engaged with all three modes. Task instructions, speed, and accuracy were automatically recorded. Scores ranged from 1–16 in manual tasks and 2–32 in automated conditions. Participants were selected via non-random sampling and used mobile devices running Android Lollipop.
Results
- Manual: Scores ranged from 1–16, with an average of 9.2 (SD = 2.3).
- Randomized Automation: Scores ranged from 2–32, average 21.5 (SD = 4.7).
- Self-Selected Automation: Scores ranged from 2–32, average 26.1 (SD = 3.8).
ANOVA showed significant differences in performance across all three modes (F(2,28) = 54.62, p < .001), with self-selected automation outperforming both others significantly.
Analysis
Automation improved performance across all conditions, with self-selected automation providing the greatest benefit. Results support the hypothesis that user agency enhances efficiency and satisfaction. Randomized automation yielded performance gains but lower satisfaction due to unpredictability. Self-selected automation resulted in more consistent and higher scores, indicating its value in user-driven task flow.
Conclusion
This study confirms that while all forms of automation outperform manual task completion, self-selected automation offers the most significant improvements. Trevo's modular design should prioritize adaptive, user-driven automation with override options to balance autonomy and machine assistance—especially on low-end mobile devices.
Research Instrument (Demo)
Instrument Used in our Research.