Research Methodology

This document codifies the rules the research-theory skill follows. The goal: produce reports that are epistemically honest, calibrated, and useful — not essays that either credulously repeat claims or reflexively dismiss them.

Source tiers

Stronger claims require higher-tier sources. A Tier 5 source can suggest a lead but cannot establish a fact.

Tier	Description	Examples
1	Primary documents	Declassified records, court filings, FOIA releases, contemporaneous government reports, original photos/audio/video with verifiable provenance
2	Peer-reviewed academic	Articles in reputable peer-reviewed journals; ideally not pay-to-play venues
3	Mainstream investigative journalism	NYT, WaPo, Reuters, AP, BBC, ProPublica, Guardian investigative pieces. Investigative > opinion > commentary.
4	Advocacy / partisan media	Outlets with a known editorial slant. Useful to understand what advocates argue; not sufficient to establish fact.
5	Social media / anonymous	Tweets, anonymous posts, unverified leaks. Treat as leads only. Almost never sufficient on their own.

When citing a source, the report should note its tier (e.g. [3] NYT, 2017 (Tier 3)).

Framework: Steelman → Hypotheses → Evidence → Verdict

1. Steelman the theory. Write the strongest possible version using advocates' actual framing — not a strawman. Build it from the strongest-sourced proponent material available (primary-document archives, named experts, specific documented sub-claims), and explicitly flag any element that rests only on Tier 4–5 circular sourcing. Capture the strongest documented sub-claims the evaluation will have to address head-on. This forces engagement with the theory's strongest form before any evaluation.

2. Decompose into atomic empirical claims, and enumerate the competing hypotheses. Distinguish factual claims ("X met Y on date Z") from interpretive ones ("This shows coordination"); each claim is researched separately. Separately, list the mutually-exclusive explanations of the event as a whole (the official/mainstream account, the theory, and any serious alternatives) — see "Competing hypotheses" below.

3. Evaluate each claim against the source tiers. Seek primary sources first; corroborate load-bearing claims against ≥2 independent sources. Note where evidence converges and where it conflicts. For each claim, separate two questions: (a) is the cited evidence real and accurately represented, and (b) does the inference from that evidence to the conspiracy conclusion actually hold?

4. Render a verdict per claim using calibrated vocabulary (below).

5. Render an overall verdict by weighing which hypothesis the full body of evidence best fits — not by confirming or denying the theory in isolation. The theory may be partially correct: some claims supported, others contradicted, others open. Concede documented kernels explicitly, and when the verdict is Contradicted, state what actually happened.

Competing hypotheses

A claim is never evaluated in a vacuum. Enumerate the mutually-exclusive explanations of the event — at minimum the official/mainstream account, the conspiracy theory, and any other serious alternative — and reach the verdict by asking which explanation the weight of evidence best fits.

• Weight diagnostic evidence most. Evidence that discriminates between hypotheses (consistent with one, inconsistent with the others) is worth far more than evidence merely consistent with the theory. A fact the theory "explains" but which the official account explains equally well is not support for the theory.
• State why rejected alternatives are rejected. Don't just name the winner; record what each losing hypothesis fails to account for.
• Use a compact matrix for complex cases (optional): key evidence × hypothesis, marking each consistent / neutral / inconsistent. Skip it when there are only two clean hypotheses.

Calibrated verdict vocabulary

Use these terms — and only these — for per-claim and overall verdicts:

• Supported — strong evidence from high-tier sources confirms the claim
• Partially true — kernel of truth, but the theory's framing overstates or distorts what's actually established
• Contradicted — strong evidence from high-tier sources contradicts the claim
• Unfalsifiable — claim is not testable as stated (e.g. "the conspirators covered their tracks so well no evidence exists")
• Insufficient evidence — genuinely open question; evidence too thin to call

Calibrated confidence levels

Use these terms — and only these — for confidence:

• High — multiple independent Tier 1–2 sources converge
• Moderate — convergence among Tier 2–3 sources, or single strong Tier 1 source
• Low — sources are mostly Tier 3–4, or there's significant disagreement among credible sources
• Insufficient — too little to assign confidence

Named failure modes

When a claim's support depends on one of these recurring techniques, name the technique explicitly in the report — naming it is part of the rebuttal:

• Anomaly-hunting — treating any unexplained loose end as proof of a plot. A complex real event always leaves loose ends; an anomaly is a question, not an answer.
• Cherry-picking / quote-mining — citing only confirming fragments, or quoting a source out of context to reverse its meaning. Always check the surrounding context of a quoted primary source.
• Fake or collapsed peer review — citing something with the form of peer review or officialdom without the substance. Verify the venue and the actual review process, not just that "a paper exists."
• Fake experts — credentials that are real but irrelevant to the specific technical question.
• Unfalsifiability / cascade logic — every disconfirmation re-absorbed as further proof ("they covered it up," "that witness was a plant"). If the claim is structured to be immune to disconfirmation, the verdict is Unfalsifiable.
• Moving the goalposts — a standard of proof is demanded, met, then replaced with a new one.
• Gish gallop — many weak claims presented so fast that not rebutting all of them reads as conceding. Identify the load-bearing claims, address those, and state explicitly that unaddressed volume is not conceded truth.
• Circular sourcing — secondary sources citing each other with no primary at the bottom. Trace at least one link toward a primary; flag claims that bottom out only in other proponent secondaries.
• Establishment-dismissal-as-confirmation — treating every official rebuttal as part of the cover-up, making mainstream evidence inadmissible by definition.

Hard rules

• No false certainty. Avoid "definitely", "obviously", "everyone knows", "debunked" without citing the debunking.
• No false balance. When one side has overwhelming evidence, say so. Don't manufacture symmetry where none exists.
• Distinguish "no evidence found" from "evidence of absence." These are different epistemic states.
• Steelman must precede evaluation. Always.
• "Where I could be wrong" section is required. Surface the strongest counterarguments to your own conclusion, sources you couldn't access, and assumptions you're relying on.
• Every secondary web source (Tier 3–5) needs a Wayback snapshot URL and an access date. Sources rot. Tier 1 primary sources hosted in permanent institutional archives (*.gov, court PACER, declassified-document repositories) may record Wayback: n/a if a snapshot isn't applicable or possible.
• Use the source tier annotation on every citation.
• Judge conspiracy theories particularly, not categorically. Never assign a negative verdict merely because a claim "is a conspiracy theory." Documented conspiracies exist (COINTELPRO, MKUltra, the Tuskegee study, the Gulf of Tonkin incident) — judge each atomic claim on its own evidence.
• Concede the kernel. When a claim contains a verifiable kernel of truth, grant it explicitly and route it to Partially true, separating the documented fact from the larger conclusion the theory draws from it.
• Separate evidence from inference. Evaluate both whether the cited evidence is real and whether the inference to the conspiracy conclusion holds. True evidence can carry an invalid inference.
• Corroborate load-bearing claims. Each load-bearing empirical claim needs ≥2 independent sources commensurate with its weight; flag any claim resting on a single source (a single decisive Tier 1 primary may suffice, but say so).
• A "Contradicted" verdict must supply the alternative. State what actually happened — refutation alone leaves the false account in place.
• Name the failure mode. When support rests on one of the techniques in "Named failure modes," name it in-text.
• Build the steelman from the best-sourced proponent material, and flag steelman elements that rest only on Tier 4–5 circular sourcing.
• Trace media evidence to its origin. For a photo, video, or document, identify the earliest/original source; don't treat a re-upload as if it were the original.