Research: Deaths and Disappearances of U.S. Defense and Advanced-Research Scientists
An evidence-based investigation into a cluster of deaths and disappearances of U.S. scientists and researchers tied to defense, aerospace, nuclear, and advanced-research programs.
Methodology
Source tiers
Every source cited in this repository is categorized by tier. Tiers are not a fixed list of three — finer distinctions are used where the evidence warrants.
Tier 1 — Primary sources. Law enforcement press releases and case bulletins (LAPD, BCSO, NMSP, LA County Sheriff, Taos County); court filings and charging documents; White House press-briefing transcripts on .gov or equivalent archival sources; official statements from DOE/NNSA/FBI/Air Force on .gov domains; Congressional records; institutional statements and obituaries published directly by the employer (Caltech, MIT PSFC, NASA JPL, LANL, KCNSC, AFRL); direct statements from family members (Facebook posts from verified family accounts, family-run search pages, on-camera press conferences) — primary for the family's own statements, not for facts about the investigation itself.
Tier 2 — Named expert commentary. On-the-record statements from identifiable subject-matter experts (e.g., former FBI officials, CSIS / NTI / think tank analysts, UAP-disclosure advocates with documented professional backgrounds). An expert speaking on the record stakes their professional reputation on a specific claim — this carries more weight than filtered reporting. Evaluated individually on the expert's credentials, independence, and the relevance of their claim.
Tier 3 — Local and beat reporting. Local news outlets with geographic or institutional proximity to the cases (Albuquerque Journal, Taos News, Los Alamos Reporter, KRQE, KOB, CBS LA, Boston Globe); specialty trade press; reporters with a documented beat covering the relevant institutions or subject matter. These reporters typically have direct sources, community knowledge, and sustained coverage that national outlets lack.
Tier 4 — National mainstream reporting. National news outlets (CNN, CBS, ABC, NBC, Reuters, AP, Washington Post, NYT) covering cases outside their regular beat. Network "Dateline" and "what we know" roundups fall here, not at the primary level, even when they quote primary sources. National outlets have editorial infrastructure but often lack domain expertise and local context; parachute coverage is particularly prone to paraphrasing errors and decontextualized quotes.
Tier 5 — Tertiary / aggregator. Wikipedia; news roundups citing other news outlets without original reporting; "here's what we know so far" compilations.
Tier 6 — Reporting relying on anonymous sources. Claims that appear in Tier 3 or Tier 4 outlets but trace to a single anonymous source and are then repeated across other outlets. The Daily Mail's KCNSC employment claim for Steven Garcia is the canonical example.
Tier 7 — Independent commentary, Substack, YouTube, TikTok, podcasts, social media. Useful for finding leads and seeing what claims are in circulation. Not used as evidence for factual claims.
Tier 8 — Foreign state-affiliated press. Included when geopolitically relevant. Country of origin and known editorial orientation noted. Not dismissed for being foreign; not elevated for being foreign.
Quote provenance rule
When a factual claim or direct quote is found only in national mainstream reporting (Tier 4) or lower, researchers should attempt to trace it to a higher-tier origin before relying on it. Paraphrasing in national coverage is common and sometimes inaccurate — a quote attributed to a source in a national article may not reflect what the source actually said. If a quote cannot be traced to a primary source (Tier 1), an on-the-record expert statement (Tier 2), or a local reporter's direct interview (Tier 3), that limitation must be noted alongside the claim.
Confidence ratings
Separate from source tier, every factual claim in this repository carries one of the following labels:
- Confirmed — Multiple independent primary sources, or a single primary source making a self-evidencing statement (e.g., a sheriff's press release stating the sheriff issued an alert).
- Reported — Appears in credible secondary reporting, often with named sources, but not directly confirmable via primary documentation.
- Alleged — Claim made by a specific identifiable source but not independently confirmed.
- Speculated — Inference, theory, or pattern-matching not directly asserted by any source.
The tier tells you provenance. The rating tells you weight. These are independent. A Tier 1 source can make a "may have been a danger to himself" statement that rates only "Reported." A Tier 3 local news article can cite a rock-solid court filing that rates "Confirmed."
Inclusion criteria for cases
Cases are included based on this repository's own criteria, not based on any external list (including the White House review list):
- Employment in or documented association with U.S. defense, aerospace, nuclear, or advanced-research programs
- Death or disappearance between 2022 and the present
- At least one of: unexplained circumstances; unresolved investigation; primary-source law-enforcement or institutional involvement
Cases that fit the pattern are included regardless of whether they appear on any public list. Cases on public lists that do not fit the pattern are still included, with a clear "inclusion rationale" explaining the weaker fit (e.g., Amy Eskridge, Jason Thomas).
Neutrality and bias handling
- No country or political entity is pre-excluded or pre-implicated. Hypotheses involving any state actor (including allies such as Israel and the United Kingdom, adversaries such as Russia, China, and Iran, and the United States itself) are evaluated on evidence alone.
- Sensational framing ("assassinations," "silenced," "targeted") is avoided in this repository's own prose. Such framing from sources is quoted and attributed, not adopted.
- White House, FBI, and other U.S. government statements are Tier 1 as documents (they are primary evidence of what was officially said) but the factual claims inside them are evaluated on the same standard as any other source.
- Exotic hypotheses (foreign intelligence targeting, UAP/UFO-related theories, other speculative frameworks) are evaluated as hypotheses: what evidence would support them, what evidence contradicts them, and what is the base rate. Not dismissed a priori; not endorsed without evidence.
No contact policy
No attempts are made to contact any individual, family member, agency, employer, or journalist in connection with this research. All sourcing uses already-public material only.
How to read this repository
- the Dossier — The top-level synthesized document covering all cases and analysis.
cases/— One markdown file per case with structured facts, sources, and confidence ratings.appendices/— Supporting material: primary-source excerpts, named expert commentary, and foreign coverage with provenance.analysis/— Cross-case connection analysis, pre-registered hypotheses, and dedicated analysis layers.data/— Structured JSON specs for diagrams and timelines, with schemas.logs/— Research log, contradiction tracker, and known-unknowns register.
PDFs are generated separately (prompt 002) into pdf-output/ and are gitignored. The website rendering is handled separately (prompt 003).