EB-1A RFE Rates in 2025: The Numbers USCIS Doesn't Publish

The Statistics Nobody Publishes

If you've been researching EB-1A and found that every article quotes the same few statistics ("EB-1A has a 40-60% approval rate," "30% of petitions receive an RFE"), you're not alone. USCIS publishes aggregate petition-level outcomes but not the granular data that would actually help you prepare: which criteria get challenged, at what rates, for which professions, from which countries, with what outcomes.

Lumova's dataset of 10,000+ AAO non-precedent decisions fills this gap. It's not a substitute for official USCIS data — it's drawn from appeals, which are themselves a filtered population — but it's the most comprehensive source of per-criterion, per-field RFE and denial patterns currently available outside USCIS itself. This article reports what the data actually shows. The numbers are rounded and anonymized, and they're current as of Q1 2025.

The Headline Number: RFE Rates Are Trending Up

Across the Lumova dataset, the estimated overall RFE rate for EB-1A petitions in 2024 was 31%. That's up from approximately 27% in 2022 and 24% in 2020.

This trend is consistent with what experienced practitioners report: USCIS has been applying the Kazarian Two-Step framework more rigorously since 2019-2020, and adjudicators are issuing RFEs earlier and on more criteria than they did a decade ago. If you're filing in 2025, you should plan for roughly a one-in-three chance of receiving an RFE — and you should not assume that a strong petition is automatic protection.

By service center: The Texas Service Center issues RFEs at approximately 34%, Nebraska at approximately 28%. This is consistent with what many attorneys have observed over the past decade: Texas tends to be more aggressive on C5 and C7 challenges specifically. You cannot choose your service center, but knowing which one your case is likely to be routed to can shape how much pre-emption work you invest in the brief.

Per-Criterion RFE Rates

Not all criteria are challenged equally. Here's what the Lumova dataset shows for the top-challenged criteria, measured as the percentage of RFEs that cite each criterion:

• Criterion 5 — Original Contributions of Major Significance: 62% of RFEs
• Criterion 4 — Judging the Work of Others: 38%
• Criterion 7 — Leading or Critical Role: 34%
• Criterion 6 — Scholarly Articles: 27%
• Criterion 1 — Nationally Recognized Awards: 22%
• Criterion 8 — High Remuneration: 19%
• Criterion 2 — Memberships: 14%
• Criterion 3 — Published Material About Petitioner: 11%
• Criterion 9 — Commercial Success in Arts: 7%
• Criterion 10 — Display at Exhibitions: 4%

(RFEs can challenge multiple criteria simultaneously, so these percentages sum to more than 100%.)

The takeaway: if you're claiming C5, you should expect it to be challenged. Not might — should. Build your C5 section with the assumption that the adjudicator will issue a template RFE against it, and pre-empt that template in the brief itself.

By Field: RFE Rates for Each Major Profession

The field you work in affects which criteria you're most likely to have challenged. Here are the approximate field-specific RFE rates across the dataset:

Software engineers and data scientists: Overall RFE rate ~34%. Most commonly challenged criteria: C5 (Original Contributions, ~68% of RFEs in this segment), C7 (Leading Role, ~46%), C8 (High Remuneration, ~19% — often because the petition didn't use BLS data properly).

Academic researchers (STEM): Overall RFE rate ~28%. Most commonly challenged: C5 (~59%), C6 (Scholarly Articles, ~34% — often because the petition didn't contextualize journal Impact Factors against field norms), C4 (Judging, ~31%).

Physicians and medical researchers: Overall RFE rate ~29%. Most commonly challenged: C5 (~56%), C8 (High Remuneration, ~28% — physicians often over-rely on salary alone without BLS comparison), C7 (~33%).

Startup founders: Overall RFE rate ~38% — the highest of any segment. Most commonly challenged: C5 (~70%), C7 (~51%), and unusually high rates of challenges to C3 (Published Material) because founders often submit press releases and startup blog coverage rather than genuine major media.

Artists and performers: Overall RFE rate ~26%. Most commonly challenged: C9 (Commercial Success, ~41%), C10 (Exhibitions, ~29%), C3 (~28%).

These numbers have a clear implication: your field shapes your audit priorities. A software engineer should spend disproportionate effort on C5 and C7 pre-emption. A physician should spend disproportionate effort on C5 and C8 (specifically using BLS data correctly). A founder should build C5 and C7 evidence with extra independence and rigor because this segment has the highest RFE rate in the dataset.

By Country of Birth: Patterns That Persist

The Lumova dataset also shows country-level patterns. These are not statements that USCIS adjudicators are biased — rather, they reflect systematic differences in the kinds of petitions submitted from different countries and the documentation challenges each country's applicants tend to face.

India: Overall RFE rate ~29%, approval rate after RFE response ~71%. Indian applicants are disproportionately represented in the technology segment, which itself has a higher RFE rate than average. The most common pattern: strong quantitative credentials (high citation counts, prestigious employers, high compensation) but weaker Step 2 final merits argumentation.

China: Overall RFE rate ~33%, approval rate after RFE response ~68%. Chinese applicants face specific challenges around independence of expert letters (when recommenders are from the same home institution network) and citation context (self-citations inflating raw citation counts). The dataset also shows elevated scrutiny for petitions in sensitive research fields (quantum computing, advanced materials, AI safety).

Nigeria: Overall RFE rate ~27%, approval rate after RFE response ~73%. Nigerian applicants in the dataset are disproportionately physicians and academic researchers. The most common pattern: strong credentials and genuine achievement, but thin documentation of the prestige and recognition of West African professional organizations that could strengthen C2 claims.

Mexico: Overall RFE rate ~25% — the lowest of the top-four filing countries. Mexican applicants show a more diverse professional mix (engineering, arts, academia, medicine). The most common RFE pattern is unrelated to nationality — it's C5 specificity, the same as other populations.

RFE Outcomes: What Happens After You Respond?

Here's the other number that matters: once you receive an RFE, what happens next?

Across the dataset, RFE outcomes approximately distribute as:

• ~63% Approved after response: The response satisfied the adjudicator's concerns, and the petition was approved.
• ~22% Denied after response: The response was insufficient, and the petition was denied. Of this group, most denials cite either continued inadequacy of the challenged criterion or failure to make a Step 2 final merits argument.
• ~12% Second RFE or NOID: The adjudicator issued another request for evidence or escalated to a Notice of Intent to Deny. A second RFE is rare and typically means the response addressed the wrong issue.
• ~3% Withdrawn or abandoned: The petitioner gave up, ran out of time, or switched strategies.

The 63% post-RFE approval rate is actually encouraging — most responses do work, if they're well-structured. The 22% denial rate is concerning, and it's the primary reason pre-filing audits matter so much: the cost of getting a denial is massive (refile costs, time, visa status implications), and most of those denials would have been preventable with better pre-filing preparation.

The Pre-Emption Effect

Here's the statistic the rest of this article has been building toward. Across the dataset, petitions that included explicit pre-emption paragraphs for the most common RFE templates (as described in our per-criterion article) had an approximate RFE rate of 19% — compared to the dataset average of 31%.

A 12-percentage-point reduction in RFE rate from a single structural change to the brief. No new evidence required. Just the explicit acknowledgment that certain challenges are likely and the pre-written responses to them.

This is why the pre-filing audit framework in our audit playbook article specifically emphasizes pre-emption as an hour in its own right. The data shows it works.

What the Numbers Don't Show

A caveat, because every statistics article should include one: the Lumova dataset is drawn from AAO non-precedent decisions, which are themselves filtered (only appealed cases generate AAO decisions). This means the dataset overrepresents contested cases and underrepresents petitions that were approved on first adjudication without any dispute.

What this means practically:

• The 31% RFE rate is probably roughly accurate as an overall figure, because USCIS publishes aggregate RFE data that can be cross-checked.
• The per-criterion breakdowns are directionally accurate but should be interpreted as "if you receive an RFE, here are the probabilities it cites each criterion" rather than "every petitioner has a 62% probability of a C5 challenge."
• The post-RFE outcome percentages are most useful for understanding what happens after you receive an RFE, not as a prediction of your base-rate outcome if you never receive one.

Statistics are guides, not guarantees. Use these numbers to calibrate your preparation — not to make binary decisions about whether to file.

How Lumova Generates Your Personal Risk Profile

When you run the Lumova audit, the system doesn't just reference the dataset-wide averages above. It compares your specific profile — your field, country of birth, claimed criteria, evidence strength, and expert letter network — against the subset of the dataset most similar to your case. The output is a field-percentile number (e.g., "your profile is in the top 18% of software engineers in our dataset") and a personalized RFE likelihood range (e.g., "35-55%").

This is the kind of granular, comparative analysis you cannot get from aggregate statistics. It's also the kind of signal that separates petitioners who file confidently from petitioners who file hoping for the best.

FAQ

Q: How recent is the Lumova dataset?

A: Current as of Q1 2025, with rolling updates as new AAO decisions are published. Trends tend to shift slowly — yearly updates are generally sufficient for planning purposes.

Q: If C5 has a 62% RFE rate, should I just not claim it?

A: No. C5 is the most-claimed criterion in the dataset because it's the most broadly applicable. The high RFE rate reflects the difficulty of documenting "major significance," not that C5 is strategically weak. The right move is to build your C5 section with the expectation that it will be scrutinized and pre-empt the standard template challenges.

Q: Does the overall 31% RFE rate apply to attorneys as well as self-petitioners?

A: Partially. Attorney-represented petitions in the dataset have somewhat lower RFE rates than self-represented petitions (~28% vs ~36%), but the gap is smaller than many petitioners assume. The largest differences appear in pre-emption quality and Step 2 argumentation, not in evidence collection.

Q: Can I use these statistics in my own petition brief?

A: You can reference them as general context, but they're not authoritative legal citations. What you can do is use them to guide your own preparation — spend your pre-filing hours on the criteria most likely to be challenged in your segment.

Try Lumova free →