The evidence is not presented as a guarantee or an endorsement. It is a disciplined basis for deciding where practical training and program review should look.
Research and operating record
Credentials and study documentation are available on request.
What did the study ask?
Part 107 remote-pilot certification relies on a written knowledge test and does not include a practical flight check. The thesis asked whether the abilities implicated in reported UAS accidents and incidents align with what that knowledge test assesses.
The study coded 166 U.S. accident and incident reports from NASA ASRS and NTSB, covering 2016 through 2025, to HFACS 8.0 Level 1.
The unsafe acts grouped into three patterns
The percentages are rounded, which is why the approximate values do not sum to exactly 100%.
The strongest signal sat outside the written test
Unsafe-act type was significantly associated with whether Part 107 assesses the related ability.
The errors cluster in what the test doesn’t measure.
Unsafe-act type was significantly associated with what Part 107 assesses (Cramér’s V = .44, a medium-to-large effect; p < .001). Skill-based errors concentrated in competencies the written test does not assess (69.5%); known deviations concentrated in the competencies it does (84.9%). A written exam can show a pilot recognizes the rules. It cannot show whether they can hold control, scan, manage workload, or recover when it goes sideways.
That gap is what the practice is built around. A certificate proves knowledge. Flight competence, decision-making under load, and a culture that follows the rules it already knows are separate things, and they are what the accident record keeps surfacing. No matter the aircraft or sensor, the FAA and the data put the operator in command and responsible for the flight.
What the study does and does not establish
This was a retrospective association study of publicly reported events. It identifies where coded unsafe acts and certification coverage align. It does not establish causation, evaluate any training program’s effectiveness, promise an outcome, or rank operators by experience. Figures are from the thesis dataset, N = 166, HFACS 8.0 Level 1.
Public reporting systems also have selection and reporting limits. The practice uses the study to aim questions and scenarios, not to claim that the dataset represents every Part 107 operation or event.
How the findings are used
The remote-pilot competency taxonomy helps structure the Readiness Call and target scenario-based training. It keeps three program layers visible: what an operator knows, what the operator can demonstrate, and what the operating culture reinforces over time.
Program discovery
Locate whether training, governance, records, or another constraint comes first.
Readiness Call
Discuss the program as it runs and receive a verbal next-step recommendation.
Scenario-based training
Connect practical work and decision points to the patterns in the event record.
Operator toolkit
The templates and reference material that carry the taxonomy into daily operations.