Using Regulations, USCIS Policy, and AAO Decisions in Extraordinary Ability Research

Separate authority from evidence#

Extraordinary ability research uses two broad source families. Legal and administrative sources explain what a classification requires and how an agency or court interprets the governing text. Factual sources establish what happened in the beneficiary's career: award rules, judging records, publications, contribution data, contracts, compensation records, organizational evidence, and independent context. A regulation cannot prove that a specific award was selective, while an award certificate cannot establish the regulatory standard. A useful research record labels the function of each source and avoids asking one document to do both jobs.

This distinction improves drafting and review. The EB-1A evidence-planning guide shows how to map primary and contextual proof to claims. The O-1A assembly guide adds the petitioner and proposed-work record. This guide addresses the research layer: finding, classifying, preserving, and citing the rules and decisions used to evaluate those facts.

Build a deliberate source hierarchy#

Begin with the complete governing text#

Read the relevant statutory and regulatory provisions before searching outcomes. Capture the classification definition, evidence threshold, every element of the criteria under consideration, comparable-evidence language, petitioner requirements, and proposed-work requirements. Follow cross-references and definitions. Note whether a phrase belongs to EB-1A, O-1A, or both. Small differences—such as leading or critical versus critical or essential, the number of listed criteria, or the work-documentation rules—should not be smoothed over in a combined summary.

Record the source URL, jurisdiction, version or effective date if available, access date, relevant subsection, exact proposition, and surrounding qualification. A quotation without its condition or exception can invert the rule. Avoid using a search-result excerpt as the source. Open the official or reliable publication, confirm the text, and save enough context for another reviewer to reproduce the research. When an unofficial codification is used for navigation, compare important language with the official source when practical.

Use the USCIS Policy Manual with version awareness#

The USCIS Policy Manual provides centralized agency guidance on EB-1 extraordinary ability and O-1 extraordinary ability. Use the chapter navigation, footnotes, appendices, alerts, and update history—not only a sentence copied into a memorandum. Determine which classification, criterion, and adjudicative step the passage addresses. For EB-1A, distinguish initial evidentiary findings from final merits. For O-1A, distinguish criterion evaluation, totality discussion, and the separate documentation for the petitioning and work arrangement.

Policy can be revised. A research log should include the page title, volume, part, chapter, access date, and any visible revision information. Before relying on an older saved version, check the current site and identify material changes. If a policy passage cites a regulation or decision, review that underlying source instead of relying exclusively on the paraphrase. If a source and policy appear inconsistent, do not resolve the conflict by silently selecting the preferred text; flag it for careful legal analysis.

Treat forms and instructions as time-sensitive sources#

Form I-140 and Form I-129 materials answer operational questions that a doctrinal research memo may miss: current editions, signatures, supplements, classifications, filing locations, fee information, required initial evidence, and presentation details. The O-1 instructions also interact with contracts, itineraries, consultations, and petitioner arrangements. Check the official form landing page as well as the PDF instructions because alerts or edition information may be posted separately. Never rely on a downloaded form from an unofficial archive when preparing a current filing.

Identify the status of every AAO decision#

The Administrative Appeals Office publishes precedent, adopted, and non-precedent decisions in different contexts. A precedent decision designated through the applicable process has an authority status that a routine non-precedent decision does not. An adopted decision supplies policy guidance within its stated scope. Non-precedent decisions resolve individual cases and can illuminate recurring evidentiary issues, but USCIS explains that they do not announce new constructions of law or establish agency policy binding in other proceedings.

Before citing a decision, record its identifier, date, classification, decision type, disposition, procedural posture, issue, facts material to the reasoning, exact holding or analysis relied upon, and whether later authority changed the landscape. Read the full decision. A favorable sentence may describe the petitioner's argument rather than the agency's conclusion; a remand may not be a final approval; and a decision may turn on a missing document unique to that record. Do not infer a general approval formula from a small number of selected outcomes.

Search decisions by issue, not just outcome#

Use neutral issue terms: award selectivity, membership admission standards, published material focus, completed judging, originality, major significance, distinguished reputation, critical role, salary comparison, comparable evidence, sustained acclaim, or proposed work. Review both favorable and unfavorable reasoning where available. Build an issue table that separates the rule stated, evidence described, analysis, outcome, and limits on comparison. This reduces confirmation bias and makes it easier to see which missing facts—not just which conclusion—affected the decision.

Research court decisions with jurisdiction and posture in view#

Court opinions may address standards of review, agency reasoning, procedural questions, or interpretations relevant to extraordinary ability. Their usefulness depends on the court, jurisdiction, publication status, issue presented, and later history. A district court opinion, appellate opinion, unpublished disposition, and settlement do not carry identical implications. Confirm whether the quoted passage is a holding, background, party argument, or dicta. Also identify what record was before the court; judicial review may be limited to an administrative record developed earlier.

Apply source discipline to factual evidence#

Case-specific web evidence deserves the same care as legal research. For an award, preserve official eligibility and judging rules, categories, number of recipients, date, organizer, and results. For media, preserve the complete article, title, author, publication date, outlet information, and evidence that the material is about the beneficiary and work. For salary or metrics, preserve the dataset, methodology, geography, occupation, experience level, unit, period, and limitations. For organizational reputation, prefer independent and specific records over self-promotional slogans.

A source's independence is a factual question. Record ownership, sponsorship, employment, client, collaborator, and commercial relationships when known. Do not omit disclaimers, methodology, or date ranges that qualify the claim. Dynamic pages should be saved in a form that preserves the visible content and retrieval details when permitted. Screenshots can help preserve presentation but may be insufficient without URL, title, publisher, date, and surrounding context. Foreign-language records require a complete, accurate translation consistent with applicable filing requirements.

Maintain a reproducible source log#

• ○Give each source a stable identifier and record title, publisher or issuing body, URL, publication or effective date, access date, and stored-copy location.
• ○Classify authority level: governing text, policy, form instruction, precedent, adopted decision, non-precedent decision, court opinion, government data, or commentary.
• ○State the precise proposition supported and preserve surrounding language that limits or defines that proposition.
• ○Record jurisdiction, decision status, later history, dataset methodology, affiliation, and other limitations relevant to weight.
• ○Schedule a currency check before filing and update links, form editions, policy text, dates, and time-sensitive factual records.
• ○Keep legal authority citations separate from exhibit citations so each claim can be traced to both the rule and the proof.

Write citations that show their work#

A citation should make the source and proposition verifiable. Introduce the authority accurately, quote sparingly, paraphrase without expanding the meaning, and give a pinpoint location when practical. Explain how the authority relates to the particular claim. Do not string-cite decisions whose different facts are never discussed. Do not state that USCIS always approves or denies a fact pattern based on isolated decisions. A transparent citation may acknowledge that a source is non-precedent or that a dataset covers only one segment of the field.

Resolve apparent conflicts carefully#

When sources appear to conflict, first determine whether they address the same classification, issue, time, jurisdiction, and procedural posture. Check whether a regulation changed, policy was updated, a decision was superseded, or a form instruction addresses a narrower operational question. Rank sources by authority, but do not ignore a lower-level source that exposes a practical issue. Document the conflict and the reason for the chosen interpretation. Complex conflicts are a signal to obtain qualified legal analysis, not an invitation to select the most favorable quotation.

Audit the research before publication or filing#

Run a final citation audit against the current source, not only the saved draft. Confirm that every URL opens, every quotation is accurate, every decision status is labeled, every form edition remains current, and every factual exhibit supports the claim made. Search for absolute words such as always, automatically, guarantees, must approve, or proves and test whether the source truly supports them. Remove unsupported outcome predictions. Verify that internal Library links point to the intended guide, including the EB-1A and O-1A comparison and the RFE issue-mapping guide.