A useful evaluation begins with a narrow claim and ends with a conclusion proportional to the record. It does not begin by choosing between belief and dismissal. The worked example below uses a public case that the All-domain Anomaly Resolution Office has already assessed, so the exercise demonstrates a method without announcing a new contested finding.
Worked example: the Western United States objects
Claim to test: A military infrared video recorded several objects whose appearance or motion required an anomalous explanation.
Authoritative record: AARO’s Western United States case-resolution report says military personnel reported five equidistant lights in restricted military airspace in 2021. AARO assessed that the objects were almost certainly commercial aircraft on established air corridors, at great distance from the sensor. Its public imagery page says commercial flight data and radar tracks aligned with the objects.
| Layer | Statement | Label |
|---|---|---|
| Claim | The recorded objects required an anomalous explanation. | Statement being tested |
| Evidence | A military platform recorded small lights or dots; personnel reported a possible restricted-airspace incursion. | Source fact reported by AARO |
| Evidence | AARO compared the video with commercial flight and radar information. | Source fact reported by AARO |
| Institutional finding | AARO assessed the objects were almost certainly distant commercial aircraft. | AARO conclusion |
| Inference | The public record does not require extraordinary performance to explain the imagery. | Calibrated writer inference |
| Unknown | A public reader does not possess every underlying classified or platform-specific datum used in the assessment. | Visible limitation |
1. Locate the earliest available source
Start with the official video entry and case-resolution report, not a cropped repost, reaction video, or headline. Record the issuer, title, case date, publication location, and stable URL. In this example the closest public analytical record is AARO’s report. A later article may help explain it, but it cannot silently replace the original. Use the Primary Sources guide to preserve enough citation detail to retrieve the item if a link changes.
2. Separate observation from interpretation
The observation is that personnel and an infrared sensor recorded multiple lights or dots that appeared regularly spaced and raised an airspace concern. The claim that they displayed extraordinary behavior is an interpretation. It depends on range, geometry, platform motion, sensor characteristics, and the relationship between each light and known traffic. Write these layers separately before judging either one.
3. Check provenance and continuity
The official case page establishes an identified government publication path for the displayed imagery and report. That is stronger provenance than an unattributed upload, but it does not give a public reader unlimited access to the original sensor stream, every processing step, or all supporting data. Mark that limitation. Do not fill a chain-of-custody gap with confidence merely because a file is dramatic or official.
4. Test independence
Five visible lights are not automatically five independent sources. If they share one video, sensor, platform, clock, and viewing geometry, they are multiple targets within one observation system. The reported alignment with separate commercial flight and radar information is more useful because it adds a different dataset. Independence still requires asking whether the datasets were genuinely separate and correctly synchronized.
5. Establish geometry and baseline behavior
AARO reported that the objects were at great distance and aligned with established air corridors. Distance matters: faraway aircraft can appear as small dots with little visible structure, and platform motion can change apparent motion. The baseline comparison is therefore not an imagined exotic vehicle but ordinary aircraft at the reported ranges and headings. For a deeper treatment of sensors and geometry, continue to the Evidence and Technology hub.
6. Read the strongest counterargument
The strongest ordinary explanation is the official aircraft assessment, not a generic dismissal. Test whether it accounts for the number and spacing of lights, their pace, the flight corridors, and the radar correlation. Also ask what the public report omits. A counterargument is strongest when it predicts the observed features and names its evidence; it is weak when it merely says that something looks familiar.
7. Calibrate confidence
Assessment: The claim that the public imagery requires an anomalous explanation is contradicted by the currently available official analysis. Confidence in the narrower aircraft explanation is high but not absolute for an outside reader, because AARO uses the phrase almost certainly and the public does not independently hold every supporting datum. This wording reports the institution’s confidence without turning it into personal certainty.
8. Preserve falsifiability
The assessment should change if authenticated, time-aligned data showed that the lights could not correspond to the cited aircraft; if range or motion data demonstrated performance incompatible with those tracks; or if AARO withdrew or materially revised its resolution. A claim protected from every possible contrary result is not being tested.
Confidence rubric
- Documented: the source establishes that an event, record, or statement exists, without necessarily proving the proposed explanation.
- Well supported: strong provenance, independent corroboration, and tested alternatives converge on a conclusion.
- Plausible: consistent with the record, but important tests or data remain incomplete.
- Unresolved: available evidence does not discriminate among live alternatives.
- Weakly supported: the claim relies on major assumptions, poor provenance, or dependent retellings.
- Contradicted: reliable evidence conflicts with the claim’s necessary prediction.
These labels describe the current evidence, not a person’s honesty or intelligence. The Sources and Methodology page explains how the site separates an account from its ultimate explanation, while the Critical Analysis hub collects resolved, disputed, and skeptical material.
Stop conditions
Pause rather than force a conclusion when the original source cannot be located, essential metadata are missing, range or timing cannot be established, supposedly independent sources trace back to one account, or the claim changes each time contrary evidence appears. Stop immediately if further collection would require trespass, unsafe conduct, harassment, unlawful access, or disclosure of private or classified material. The correct interim result may be insufficient public evidence.
Reusable claim worksheet
- Exact claim: Write one testable sentence.
- Claim type: Observation, identity, performance, program, document, or physical material?
- Earliest source: Creator, date, full title, identifier, and stable URL.
- Direct observations: What does the source record without interpretation?
- Inferences and assumptions: What must be true for the claim to follow?
- Provenance: Where did each item come from, and what transformations occurred?
- Independence: Which witnesses, sensors, or documents are genuinely separate?
- Geometry and baseline: What range, timing, field of view, platform motion, and ordinary comparisons are available?
- Best counterargument: What alternative explains the most evidence with the fewest unsupported assumptions?
- Unknowns: Which missing facts prevent a stronger conclusion?
- Confidence: Choose a rubric label and explain why.
- Change condition: Name the evidence that would raise, lower, or reverse the assessment.
Keep a copy of the source and your completed worksheet. If a Countdown to Disclosure page omits decisive evidence or uses a stronger conclusion than its sources support, send the page URL, passage, and record through the corrections channel. Terms that need clarification are collected in the Glossary.