Paranormal link analysis: finding patterns without overclaiming

Define a question that the records can answer

Link analysis starts with a question, not an impressive network diagram. Instead of asking whether all unusual experiences are connected, ask which accessible accounts in a particular region describe a particular combination of details. That gives you something you can inspect, revise and explain.

Keep the reported experience separate from its category. “Humming sound” is a descriptive phrase; “UFO encounter” is a classification. Searching both may be useful, but forcing every account into a category before reading it can hide connections across different kinds of reports.

A hypothetical investigation: missing time, humming and Utah

This is a made-up search scenario, not a claim about a real Utah case. Imagine that a researcher wants to compare accounts mentioning missing time and a humming sound in Utah. Start with each detail separately, then combine them. Inspect the strongest overlaps and keep track of which detail each passage actually supports.

If one record contains all three terms, check whether the place refers to the event rather than the author’s address, and whether the sound and time gap belong to the same episode. If another record contains only two details, it may still be useful—but the missing third detail must not be supplied by inference.

Where a location is approximate, do not turn it into an exact map pin or a claim that two events occurred five miles apart. A map is only as precise as its underlying location evidence. An unfamiliar detail may be a promising lead without being a geographic cluster.

Distinguish relationships from resemblance

Write down what kind of connection you mean. “Both reports mention a hum,” “both summaries quote the same newspaper” and “both accounts name the same place” describe different relationships. Only the first is a shared reported detail; the second may explain why the wording is similar.

A useful comparison ledger has one row per claimed connection and a passage reference on each side. Add a note about source dependence and uncertainty. This makes it possible for someone else to disagree with the link without having to reject your entire investigation.

• Which detail is shared?
• Which passage supports it on each side?
• Does the passage affirm, question or deny the detail?
• Are the sources independent or copied?
• What alternative could explain the resemblance?

Use AI-sourced dossiers as labelled research leads

An AI-sourced dossier can suggest a case name, phrase or reference worth retrieving. Its prose is not an original witness statement or documentary evidence. Follow the references and identify which have actually been retrieved before relying on their contents. A plausible-looking reference is still a lead until checked.

For a cross-domain starting point, the Fouke Monster and Boggy Creek tracks dossier is clearly labelled as sourced using AI. Use it to frame questions and seek the underlying material, not as proof that an account or physical trace has been independently verified. The label describes provenance; it should remain visible when a record is shared.

Ask what is missing from your pattern

A collection of similar accounts does not tell you how common the detail is in all accounts. Search results reflect the records available, their wording and your query. They are not a representative survey of every experience in a region or period.

Look for counterexamples and near misses: accounts in the same place without the sound, accounts with the sound but a different setting, or a passage explaining how a widely repeated phrase entered later retellings. Those comparisons can help distinguish a genuinely useful lead from a pattern your query was designed to produce.

Produce a research brief, not a verdict

Keep a shortlist with the strongest passage-backed overlaps, known source dependencies and explicit gaps. State what would change your assessment. A next step might be retrieving a newspaper page, clarifying a date or checking whether two records are retellings of one event.

The goal is to make investigation easier: fewer unexplained leaps, clearer routes to sources and better questions for any responsible follow-up. A connection is valuable when it helps decide what to examine next, even if the eventual explanation is ordinary or the available evidence remains inconclusive.