Evidence Building
O-1A Evidence Strategy for Researchers With Preprint-Heavy Records
In machine learning, economics, and computational biology, the most significant work often lives in preprints before peer review completes. This article explains how to frame arXiv and bioRxiv research for O-1A criteria, and when to lead with original contributions instead of the scholarly articles criterion.
Why preprint records complicate O-1A petitions
In fields such as machine learning, computational biology, economics, and physics, the preprint has become the primary vehicle for scientific communication. Papers posted to arXiv, bioRxiv, medRxiv, or SSRN circulate months or years before formal peer review completes—sometimes accumulating hundreds of citations before journal publication. For O-1A petitioners whose most significant work exists only in preprint form, this creates a genuine presentation challenge. USCIS adjudicators evaluating the scholarly articles criterion are trained to look for peer-reviewed journals as the benchmark of scholarly credibility. A petition that leads with preprints without adequate contextual explanation will often draw a Request for Evidence asking for proof of journal acceptance or peer review.
The challenge is not insurmountable, but it requires a more deliberate framing strategy than a traditional publication-based record. The central argument the petition must make is that in the specific field at issue, preprints function as the operative scholarly publication—not a preliminary draft awaiting legitimacy, but the document that the field's researchers actually cite, debate, and build upon. This argument is strongest in sub-fields of machine learning and deep learning, where accepted papers at venues such as NeurIPS, ICML, ICLR, or ACL are simultaneously posted to arXiv and treated as definitive publications. It is weaker in fields where peer review remains the exclusive credentialing mechanism, such as clinical medicine or traditional humanities disciplines.
Two decisions shape the entire evidence strategy. First, identify whether the petitioner's preprints have accumulated citations—particularly from published journal articles and peer-reviewed conference proceedings—that demonstrate the field treats the work as authoritative. Citations in published work are far more persuasive than preprint-to-preprint citations. Second, assess whether any of the petitioner's preprints have since been accepted for publication or presented at a highly selective conference. Even one or two peer-reviewed publications in a selective venue provides the anchor that allows the rest of the preprint record to be framed as consistent with how this particular field communicates. Without that anchor, the petition faces a harder burden.
Scholarly articles and the preprint question
The scholarly articles criterion under 8 C.F.R. § 214.2(o)(3)(iii)(B)(5) covers scholarly articles in the field in professional journals or other major media. USCIS policy guidance has extended this language to cover publications in conference proceedings when those conferences are the primary peer-reviewed publication mechanism in the field. At NeurIPS, ICML, ICLR, and similar top-tier machine learning conferences, acceptance rates below fifteen percent and rigorous double-blind review processes make acceptance functionally equivalent to journal publication. A petition that includes these conference papers alongside an expert letter explaining the field's publication norms has a strong basis for satisfying the criterion, even if the petitioner's record otherwise skews toward arXiv preprints.
For fields where conference proceedings do not carry this weight—computational biology, economics, certain areas of chemistry—the petition strategy for the scholarly articles criterion should identify any journal publications, even limited in number, and frame those as the publication anchors. A single first-authored article in a leading journal is often stronger evidence than ten highly cited preprints, because it gives the adjudicator an unambiguous data point of peer-reviewed acceptance. The preprint record then becomes supplementary evidence of productivity and influence rather than the primary criterion evidence. Citation exhibits attached to each publication—showing how the work has been cited in subsequent peer-reviewed literature—should be standard in any scholarly articles exhibit.
When the petitioner's record is entirely in preprints with no peer-reviewed publications, the scholarly articles criterion becomes the petition's most vulnerable point. In that scenario, the petition should consider leading with original contributions—where the citation record for preprints is more directly probative—and treating the scholarly articles criterion as a secondary argument. Expert letters from established researchers in the field, explaining why the specific preprints at issue are regarded as scholarly publications in this sub-field, carry significant evidentiary weight and should be tailored to the publication norm question, not reserved for a generic statement of the petitioner's overall importance.
Original contributions from preprint evidence
The original contributions criterion under 8 C.F.R. § 214.2(o)(3)(iii)(B)(6) requires evidence of original scientific, scholarly, or business-related contributions of major significance in the field. This is often the strongest criterion for researchers whose work lives primarily in preprints, because the evidence of significance—citations by peer researchers, adoption in downstream work, incorporation into widely used software frameworks or datasets—does not depend on publication format. A preprint that has been cited extensively in published journal articles and conference proceedings demonstrates major significance in the same way a journal article does; USCIS adjudicators looking at citation exhibits do not discount them based solely on the source document's peer review status.
The citation exhibit for original contributions should be constructed carefully. Google Scholar citation counts are the starting point but not the end. The petition should identify the most significant citing works—papers in top journals or selective conferences that explicitly credit the petitioner's preprint for a technique, dataset, finding, or framework they adopted—and summarize each citation's context. A hundred citations in obscure forums are less persuasive than twenty citations in high-impact venues that explain how the petitioner's work advanced the state of the field. If a widely used open-source library, benchmark dataset, or standard methodology traces directly to the petitioner's work, that adoption is primary evidence of major significance.
Expert letters for the original contributions criterion should come from independent researchers who have cited the petitioner's work or who work in adjacent areas of the field. An expert who can attest that they have used the petitioner's method in their own published work, and that it changed how their group approaches a specific problem, is more persuasive than an expert providing a general testimonial about the petitioner's abilities. Drafting these letters is therefore a collaboration: provide each expert with the petitioner's specific preprints, their citation data, and the names of downstream works that built on them. Letters drafted without this grounding tend toward boilerplate and are easily discounted by adjudicators trained to identify generic praise.
Judging and peer review in preprint-era research
Researchers who work primarily with preprints have often also participated in the peer review ecosystem—as reviewers for journals, as program committee members for conferences, or as discussants at workshops—because preprint norms do not displace peer review in the same field; they run alongside it. For the judging criterion under 8 C.F.R. § 214.2(o)(3)(iii)(B)(4), the relevant question is whether the petitioner has been recognized by journal editors or conference organizers as someone whose judgment is valued. Invitation to review for a selective conference such as NeurIPS, ICLR, or ICML—or for a top journal in the field—satisfies the criterion directly.
Quantifying peer review service strengthens the judging exhibit. A letter from the program chair of a major conference confirming the number of papers the petitioner reviewed in a given year, or an OpenReview profile showing extensive reviewing activity, supplements the documentation. For conference program committees, an invitation to serve as an Area Chair or Senior Program Committee member carries greater weight than general reviewer service, because these roles involve evaluating and synthesizing multiple reviews rather than simply providing one. Area Chair appointments at top venues are competitive and require recognition by the program chairs of the petitioner's expertise. These appointments should be documented with the original invitation correspondence and any acknowledgment in the proceedings.
An often-overlooked form of judging in computational research is reviewing grants for funding agencies in whose portfolio the research falls. NSF review panels, NIH study sections, and DOE program reviews invite researchers who have demonstrated expertise in emerging computational or biological fields. For a petitioner whose work is primarily in machine learning or computational science, an invitation to a review panel at NSF or NIH represents exactly the kind of expert recognition the judging criterion contemplates—independent acknowledgment from a credentialing institution that this researcher's judgment is valued. Invitation letters and program documentation should be organized into a separate exhibit.
Salary, memberships, and field recognition
For researchers at universities, national laboratories, or technology companies, the high salary criterion under 8 C.F.R. § 214.2(o)(3)(iii)(B)(8) is often achievable when compensation can be compared to Bureau of Labor Statistics benchmarks for the relevant occupation code. For a machine learning researcher at a well-funded technology company, total annual compensation—including base salary, equity, and bonuses—frequently exceeds the ninetieth percentile of BLS SOC 15-1221 earnings in the metropolitan area. Offer letters, W-2s, most recent pay stubs, and a comparison exhibit against BLS data for the appropriate occupation code and geography constitute the standard evidence package. This criterion does not require national recognition; it requires numerical proof of relative compensation standing.
Memberships in selective professional associations remain available as a criterion anchor even when the petitioner's publication record skews toward preprints. In machine learning and related fields, selective memberships include IEEE Fellow status, ACM Fellow status, or invitation to an invitation-only program such as the Simons Institute for the Theory of Computing semester program, where selection is based on expert nomination. These memberships reflect recognized standing that does not depend on the format in which the petitioner publishes. The evidentiary presentation should include the selection criteria for each membership, distinguishing between associations that admit anyone and those that require competitive or expert review.
Press coverage for preprint researchers tends to cluster around research breakthroughs and results that attract mainstream or specialized science journalism before formal publication. Coverage in Nature News, MIT Technology Review, Wired, Science, or specialized outlets such as dedicated machine learning newsletters—when it is specifically about the petitioner's research and identifies the petitioner by name and role—satisfies the press criterion under 8 C.F.R. § 214.2(o)(3)(iii)(B)(3). The exhibit should document each piece with the publication's audience, reach, and editorial standing. Coverage that mentions the petitioner's institution rather than the petitioner personally is less persuasive and should be framed carefully.
Building the complete preprint-era evidence file
The practical sequencing of a preprint-heavy O-1A petition typically follows a four-criterion structure built around original contributions as the anchor, supplemented by scholarly articles including conference proceedings, judging, and high salary or press coverage as secondary criteria. The petition brief should contain a dedicated section explaining the preprint ecosystem to a USCIS adjudicator who may not be familiar with arXiv or bioRxiv. This section should be short, precise, and supported by an expert letter from a senior researcher in the field. The goal is not to educate the adjudicator at length, but to establish a factual premise that the rest of the petition builds upon.
Timing considerations are acute for preprint researchers. If any of the petitioner's most-cited preprints are currently under review at a journal, the filing timeline should be aligned with an anticipated acceptance—where possible—so the petition can include at least one publication acceptance as a bridge between the preprint record and the peer-reviewed world. If a paper has been accepted but not yet formally published, the acceptance letter from the journal editorial office constitutes proof of peer-reviewed publication for O-1A purposes. Filing four to six weeks after a journal acceptance allows the petition to anchor on that accepted paper while the preprint citation record remains current and comprehensive.
The most common Request for Evidence for preprint-heavy petitions challenges either the scholarly articles criterion or the original contributions criterion on the ground that citation counts without peer validation are insufficient. Having ready-made response materials prepared before filing—including additional expert letters, field-norm statements from journal editors, and expanded citation analysis—reduces the delay from RFE receipt to submission. An attorney experienced in O-1A petitions for researchers in computational and quantitative fields will have seen these RFE patterns and can structure the initial petition to preempt the most common objections. Proactive framing is always more efficient than reactive remediation.
What we typically gather for this kind of case
| Document | Where to source | Why it matters |
|---|---|---|
| Peer-reviewed publications | Web of Science / Scopus exports | Anchors original-contributions and authorship criteria |
| Citation analysis | Google Scholar profile + ESI top-1% data | Quantifies major significance in the field |
| Salary benchmark | BLS OEWS for SOC code + locality | Documents high-salary criterion at 90th-percentile or above |
| Critical-role letters | Direct supervisor + program director | Establishes role's importance, not just title |
What we see go wrong, again and again
- 01Treating extraordinary ability as a credentials checklist rather than a story of field-wide impact.
- 02Submitting bibliometric data (h-index, citation counts) without explaining what makes those numbers high relative to peers in the same sub-field.
- 03Relying on letters from collaborators or co-authors rather than independent experts who can speak to influence.