• Goal: preserve evidence quality through records and corroboration while reducing story drift.
• Primary threats: memory, suggestion, rumor, motivated interpretation, and missing baseline rates.
• Best outputs: well-documented timelines, source separation, and clearly bounded claims.
• Bottom line: field methods can strengthen or weaken a case, but they do not substitute for controlled experimentation.

• Evidence preservation: captures what was said, when, and by whom, before later reinterpretation changes the story.
• Corroboration mapping: separates witnesses who heard the story independently from those who heard it from the same source.
• Context reconstruction: identifies stressors, expectations, cultural cues, and environmental factors that may explain the report.
• Hypothesis generation: points to variables worth testing in the lab, such as mental states, triggers, or social dynamics.

• Guarantee causal inference: uncontrolled settings leave many ordinary explanations open. You cannot rule them all out.
• Estimate true frequencies: reporting is selective. We rarely know how many similar events happened and were never reported.
• Eliminate hindsight: even honest witnesses may unconsciously reshape their memory after learning the outcome.

• Write a one-sentence claim that can be checked. Specify what happened, to whom, when, and what would count as verified.
• Separate the raw experience from the interpretation placed on it afterward.
• Identify what the witness knew at the time versus what they learned later.

• Create a chronological timeline with timestamps as precise as possible.
• Log who knew what when. Include witnesses, family members, professionals, and investigators.
• Record first disclosures: when the witness first told someone, and what exactly was said at that moment.

A field investigation gets stronger when it can rely on records that were created before the confirming event, or independently of the witness’s later story. Useful examples include dated notes, letters, time-stamped messages, call logs, diaries, medical records, travel receipts, and third-party documents.^[2] A record made before the outcome was known is much harder to dismiss than a memory recalled afterward.

• Interview witnesses separately when possible. Shared interviews let one person’s account shape another’s.
• Ask for verbatim recall before asking for interpretation. Avoid leading questions.
• Document uncertainty explicitly. A witness saying “I think” is different from “I know.”
• Track contamination pathways: who spoke to whom after the incident, and in what order.

Before concluding anything unusual happened, a good investigator works through four categories of ordinary explanation:

• Perceptual: misidentification, sleep transitions, hypnagogia (the drowsy state between waking and sleep), or seeing patterns in random noise.
• Informational: rumors, subtle cues, prior knowledge, or unconscious inference from available facts.
• Statistical: coincidence, base-rate neglect (forgetting how often similar things happen), or selective memory.
• Social: suggestion, conformity, or story escalation through repeated retelling.

• Early documentation: a record that predates the confirming event. This is the single most important quality marker.
• Independent corroboration: witnesses or documents that do not share the same story pathway.
• Specificity: clear, checkable details rather than broad statements that could fit many outcomes.
• Bounded opportunity: limited realistic routes for ordinary information transfer.
• Transparent gaps: uncertainties are recorded rather than glossed over.

• The first account appears only after the outcome is widely known.
• All witnesses learned the details from the same storyteller.
• Key specifics emerge gradually through retelling, getting sharper each time.
• Records were reconstructed from memory after the fact rather than written at the time.

Many classic field reports involve crisis apparitions, strong impressions, or meaningful coincidences reported around a death, injury, or other high-stakes event. These cases are evaluated mainly on timing and record quality. The key question is: what was documented before the witness learned the outcome?^[1]

Read: Spontaneous Cases Collection and Crisis Coincidence Reports

These investigations focus on environmental checks, deception controls, careful event logging, and witness separation. Some cases allow controlled observation windows, but the setting is hard to standardize. Ordinary explanations such as building noise, misperception, and social contagion must be ruled out first.

Field mediumship investigations focus on information control. The key questions are: what could the medium have learned through normal means, what did the sitter disclose, and how are hits counted against misses?^[2] Stronger approaches borrow ideas from lab research, such as blinding (keeping the medium unaware of who the sitter is), using proxy sitters, and controlling feedback. Real-world sittings, though, often include many opportunities for ordinary information to leak through.

• Critique 1: Hindsight reconstruction

  Skeptical critique based on memory research.

  Details in a story often sharpen after verification. Witnesses unconsciously fill in gaps to match the known outcome. This is not dishonesty. It is how human memory works.

  This is a well-established finding in cognitive psychology and applies to all eyewitness testimony, not just psi reports.

• Critique 2: Selection bias

  Skeptical critique based on reporting patterns.

  Striking cases get shared. Mundane misses are forgotten. This means the cases that reach investigators are not a random sample of all experiences.

  Without knowing the denominator (how many similar experiences produced no match), it is impossible to judge whether the hits are above chance.

• Critique 3: Social contagion

  Skeptical critique based on social psychology.

  Shared storytelling creates agreement among witnesses that looks like independent corroboration but is not. Once people discuss an event, their accounts converge.

  Investigators counter this by interviewing witnesses separately and tracking who spoke to whom before the interview.

• Critique 4: Base-rate neglect

  Skeptical critique based on probability.

  Coincidences are expected when many people have many experiences every day. A striking match between a dream and a real event may simply reflect the large number of dreams people have that are never checked against reality.

  Without a baseline rate for how often similar experiences occur without a match, a single striking case cannot be evaluated statistically.

Five questions cut through most field reports quickly:

• Timeline integrity: does the documentation clearly show the order of impression, disclosure, and verification?
• Record quality: are there records made at the time, or only retrospective testimony?
• Independence: are corroborators truly independent, or part of the same story network?
• Information pathways: what ordinary routes could have delivered the key detail to the witness?
• Claim boundedness: is the claim specific enough that it could clearly have been wrong?

If a case’s key specifics cannot be shown to exist before the verifying event, the case is weak. If witnesses are not independent, the case is weak. If plausible ordinary information routes remain open, the case is weak.

Cases that preserve records made before the fact and show genuine witness independence become more informative. They narrow the space of ordinary explanations. That is true whether or not the case ultimately supports a psi interpretation.