Qualitative publication. The working set contains seven publication-development records. No population-level consistency percentage is reported because the validated source architecture does not support one.
What a longitudinal source comparison can—and cannot—tell us
Stories change. Memory changes. Questions change. Interviewers emphasize different details. A witness may remember something decades later that was never recorded in the earliest account.
That makes a seemingly simple question—does a witness keep telling the same story?—surprisingly difficult to answer.
The Sasquatch Research Observatory examined a small set of witness records for which the same reported encounter can be followed through more than one source layer. The purpose was not to decide whether the underlying encounters were genuine. It was narrower: when recoverable accounts of the same event are separated by time, what remains stable, what appears later, and what actually conflicts?
The strongest result is methodological before it is statistical:
A later detail is not automatically a changed story.
Across the strongest longitudinal records examined, the basic structure of an encounter can remain recognizable while later accounts add physical description, context or interpretation. That does not establish that the encounters occurred as reported. It does show why story stability cannot responsibly be measured by treating every later-added detail as a contradiction.
How the comparison was built
The analysis began with 21 records that automated screening identified as possible longitudinal witnesses. That number did not survive validation as a usable sample.
A witness was usable only when the Observatory could distinguish the same event across temporally separated source layers and compare actual claims. Duplicate publication, syndicated material, different encounters by the same witness and secondary retellings were not treated as independent longitudinal accounts.
The publication-development set contains seven records: Todd Neiss, Brian Gosselin, Doris Harrison, Paul Freeman, Bob Gimlin, Fred Beck and Sara Brown. Six have mature or partial claim-level analysis. Brown remains useful but requires further event separation because the Brown property record contains multiple distinct incidents.
These records are case studies, not a representative sample of Sasquatch witnesses. No population-level consistency percentage is calculated.
What counts as a change?
The Observatory separates several situations that are often collapsed together.
Stable means the same substantive claim can be recovered in separated source layers.
Added later means a later source contains information that the earlier recoverable source does not contain. Unless the earlier source explicitly says the opposite, this is not a contradiction.
Changed or contradicted requires materially incompatible claims about the same event and feature.
Unresolved is used when source coverage is insufficient.
Interpretation is also kept separate from observation. A witness's later theory about what an experience meant does not automatically alter what the witness said was observed.
This distinction is essential. If an early account says that a figure was dark and upright, while a later interview adds an estimate of eight feet, the height estimate is a later-documented detail. It cannot honestly be rewritten into the early source—and its absence from the early source does not mean the witness previously described a different height.
Todd Neiss: a recognizable physical-description core
Neiss reported seeing three figures during military training in Oregon's Coast Range in 1993. A detailed first-person account published in 2007 describes three dark upright figures, with the central figure larger than the two flanking figures, along with distinctive proportions.
Within the recoverable source chronology, several elements of that physical-description package recur in later material. Other context appears later.
The limitation is important: the earliest detailed source currently recovered by the Observatory dates to 2007, roughly fourteen years after the reported event. The case therefore helps examine later stability, but it does not give us a near-event baseline.
Bob Gimlin: unusually close to the event
The Patterson-Gimlin encounter provides a different source architecture. Roger Patterson and Bob Gimlin were interviewed by Jack Webster on October 26, 1967—six days after the reported Bluff Creek encounter.
That creates an unusually valuable early anchor. Decades-later Gimlin interviews can be compared against a source recorded almost immediately after the event rather than against another late retrospective account.
At the present stage, the basic event structure remains recognizable across that span while later interviews contain considerably more context and detail. The Observatory therefore treats the early source as the baseline and dates later elaboration to the source in which it is actually recoverable.
The point is not that Gimlin's story is therefore proven true. Narrative stability and event authenticity are separate questions.
Paul Freeman: added precision is visible in the record
A June 1982 UPI report says Forest Service employee Paul Freeman reported a reddish-brown, ape-like figure that walked upright, was hairy except for the face and was associated with a strong odor.
An October 1982 UPI follow-up supplies more detailed morphology, including thinner hair on the face and chest and visible skin. Later reporting adds still more precision.
This is a useful example of why chronological coding matters. The later details should not be backdated into the June report. But neither should their absence from June automatically be counted as inconsistency.
The record also contains a historical date discrepancy in later reporting. Rather than silently harmonize it, the Observatory retains it as a provenance issue.
Fred Beck and Ape Canyon: when interpretation expands
Fred Beck provides a particularly useful counterweight.
Contemporary 1924 reporting about the Ape Canyon miners preserves a recognizable event core: miners near Mount St. Helens, reports of large hairy figures, shooting incidents, rocks striking the cabin and the group's departure.
Beck's 1967 account retells the episode more than four decades later and contains a much more developed spiritual or paranormal interpretation.
That later interpretive framework matters to the history of the story, but it should not simply be merged with the physical observations attributed in 1924.
The case also demonstrates an attribution problem. Contemporary reports concern several miners. A statement appearing in a 1924 article cannot automatically be assigned to Beck unless the source identifies him as its speaker.
For that reason, Ape Canyon is valuable precisely because it resists a simplistic label of either “unchanged” or “changed.” Some event structure persists; later interpretation expands; and some claim-level comparisons remain limited by attribution.
Brian Gosselin and Doris Harrison: useful but asymmetric records
Brian Gosselin's 2018 account of the 1976 Abair Road encounter provides a detailed firsthand description, including a close spotlighted encounter. But a strong, claim-comparable 1976 source has not yet been recovered for the current analysis. That prevents the Observatory from pretending that every detail documented in 2018 was documented in 1976.
Doris Harrison presents almost the reverse problem. Contemporary 1972 reporting preserved by the Missouri Department of Conservation provides a valuable early description of a tall, black, shaggy figure roughly fifty feet from the house. Later retrospective material indicates that Harrison continued to stand by the core account, but the Observatory has not recovered an equally strong later firsthand transcript for detailed claim-by-claim comparison.
Both records remain informative. Neither is allowed to become stronger than its source architecture permits.
Sara Brown: why encounters must be separated
The Brown property record illustrates another problem: one witness can accumulate multiple experiences.
Brown has discussed years of reported activity, recordings, thermal material, tracks and multiple sightings. Recent interviews also contain newer incidents.
Those interviews cannot be compared wholesale and called a stable or changing “story.” Each event must first receive its own identity. Only then can the same event be traced through later retellings.
This is why the current Brown record remains partially validated rather than being forced into the denominator.
What the Observatory found
The present comparison supports a limited qualitative conclusion:
In the strongest longitudinal records examined, a recognizable core event narrative can persist while later accounts add physical detail, contextual information or interpretation.
It also supports a methodological conclusion with broader importance:
Later elaboration and contradiction are not the same thing.
This does not mean later additions are necessarily accurate memories. Human memory can change, and the present research does not establish why a later detail appeared. It means only that the evidence must be coded according to what the sources actually establish.
If an earlier source is silent about a feature, the responsible classification is not “contradiction.”
What this research does not establish
This analysis does not establish that Sasquatch exists, that any particular witness account is accurate, or that stable testimony proves an event occurred.
It also does not tell us what percentage of Sasquatch witnesses remain consistent over time.
The original automated screening pool contained 21 possible records. After event matching, source-family checks and claim-level validation, the usable publication-development set became much smaller and remained uneven in source quality. Turning that into a percentage would create an appearance of precision that the evidence does not support.
The absence of a percentage is therefore a result of the research process, not a missing result.
Why this matters
Eyewitness testimony is often discussed in extremes. A later detail may be treated as proof that a witness invented the story, while long-term repetition may be treated as proof that the event happened.
Neither conclusion follows automatically.
A better question is:
What was documented, when was it documented, and how does that specific claim compare with the same witness's other recoverable accounts of the same event?
That is the standard the Observatory will continue to use.
As additional early interviews, newspaper accounts, recordings and transcripts are recovered, RQ-001 can be revised. The analysis is designed to change when the source record changes.
The purpose is not to protect a story.
It is to preserve the difference between what the witness said, what the source actually documents, and what the evidence allows us to conclude.
Inspect the underlying record
This publication is derived from source-linked witness and event records. Exact source locations remain visible in the Observatory rather than being replaced by the article narrative.
Exact source locators · Analytical findings · Search the archive