Why EB-1A Petitions Get Denied: What 1,165 USCIS Appeal Decisions Show

Most denied EB-1A petitions fail before the final review, and meeting three criteria rarely saves them. What 1,165 appeal decisions show about each of the ten criteria.

The TeamPublished 11 min read

We read every EB-1A decision USCIS’s Administrative Appeals Office (AAO) published from January 2023 to October 2026 — 1,318 in all — and coded why each petition failed. The main sample is the 1,165 decisions on appeal. These are appeal outcomes: they cover petitions USCIS had already denied, not USCIS’s overall EB-1A approval rate.

77%

of dismissed appeals failed before the final review

94%

of appeals that met three or more criteria were still dismissed

15%

of appeals lost criteria USCIS had already granted

What we found

  • Appeals rarely win. Of 1,165 appeals, 902 (77.4%) were dismissed, 250 (21.5%) were sent back to USCIS, and 13 (1.1%) were approved.
  • Most petitions never reach the final review. 77% of dismissed appeals failed at the first step: meeting at least three of the ten evidentiary criteria. Of those, 37% fell exactly one criterion short.
  • Meeting three criteria is not enough. Of 219 appeals with three or more criteria met, 206 (94%) still lost at the second step, the final merits determination.
  • Two criteria fail almost every time; two rarely do. Memberships failed in 93% of the decisions that ruled on them, and original contributions in 91%. Judging failed in 20%, and scholarly articles in 25%.
  • An appeal can take criteria away. In 172 appeals (15%) the AAO withdrew criteria USCIS had already found met — most often published material, high salary and leading role.
  • Since October 2024, more petitions lose on final merits. In decisions after USCIS’s October 2024 policy update, 27% of dismissals came at final merits, up from 6% before September 2023. The data shows the shift, not its cause.

How EB-1A is decided

EB-1A is the green card for people with “extraordinary ability” in the sciences, arts, education, business or athletics. Applicants can petition for themselves, without an employer or a job offer.

USCIS decides in two steps, a framework set out in Kazarian v. USCIS, 596 F.3d 1115 (9th Cir. 2010):

  1. The criteria step. The petitioner shows a major, internationally recognized award, or evidence meeting at least three of ten criteria in 8 C.F.R. § 204.5(h)(3): awards, memberships, published material about them, judging others’ work, original contributions of major significance, scholarly articles, exhibitions, a leading or critical role, high salary, and commercial success in the performing arts.
  2. The final merits determination. USCIS weighs all the evidence together and asks whether the person has sustained national or international acclaim and is among “that small percentage who have risen to the very top of the field of endeavor” (8 C.F.R. § 204.5(h)(2)).

A denied petitioner can appeal to the AAO, which reviews the whole record again from the start. The AAO publishes its decisions with personal details redacted.

Methodology

SourceEvery decision USCIS published under its AAO topic “I-140 – Immigrant Petition for Alien Worker (Extraordinary Ability)” with a decision date from January 2023 to early October 2026: 1,318 decisions. USCIS posts decisions with a lag of a few weeks, so October 2026 is incomplete.

SampleThree decisions filed under that topic concerned other visa categories and were excluded. Of the remaining 1,315 EB-1A decisions, 1,165 are appeals and 150 are motions to reopen or reconsider. Motions review the AAO’s own earlier decision rather than the petition, so the analysis uses the 1,165 appeals and reports motions separately.

Rules periodsUSCIS changed its EB-1 guidance twice in this window, on September 12, 2023 and October 2, 2024. We split decisions by date into three periods: before September 12, 2023 (121 appeals); September 12, 2023 to October 1, 2024 (270); and from October 2, 2024 (774).

CodingEach decision was coded against a fixed coding manual: the outcome; for each of the ten criteria, whether it was claimed, how USCIS and the AAO ruled, and the reasons for failure; whether the AAO reached final merits and why it denied there; and other grounds such as evidence dated after filing. Over 99% of coded criterion findings are tied to the passage in the decision that supports them, and a script confirmed that each passage appears word for word in the decision.

VerificationThe coding was done by AI, with three checks:

  1. Two independent coding passes over all 1,318 decisions. The passes agreed on 100% of outcomes, 99.7% of final-merits calls, and 94–98% of individual criterion findings, depending on the criterion.
  2. Every disagreement was resolved against the decision text, and each ruling is recorded in a published log.
  3. A third blind review of a random 50 appeals agreed with the final dataset on 100% of outcomes, final-merits calls and failed criteria, and on 272 of 275 individual criterion findings (98.9%). The one final-merits difference it found exposed a coding gap, which we corrected across the whole dataset before these figures were run.

Coding rulesWhere USCIS found a criterion met and the AAO did not dispute it, we count the criterion as met. Where the AAO found fewer than three criteria met but added that the petition would fail final merits anyway, we count it as a criteria-step failure with a backup final-merits ruling (8 dismissals).

StatisticsDifferences between rules periods were tested with chi-square tests, and Fisher’s exact test where counts were small.

DataThe coded dataset, the coding manual, the analysis script and the adjudication log are published with this note: download all files (ZIP, 1.7 MB) or the dataset alone (CSV, 1.1 MB). Free to reuse under CC BY 4.0: credit Cooperative and link to this note.

What the evidence supports

Appeals rarely succeed

Of 1,165 appeals, 902 were dismissed (77.4%), 250 were sent back to USCIS for a new decision (21.5%), and 13 were approved (1.1%). A remand is not a win: USCIS decides the petition again.

Remands most often followed a gap in USCIS’s own decision. In 72 of 250 remands, USCIS had not explained its denial; in 51 it had overlooked evidence; in 35 it had applied the wrong standard; and in 76 the AAO raised an issue USCIS had not yet decided.

Most petitions fail at the first step

Of the 902 dismissals, 696 (77%) failed at the criteria step, and 206 (23%) failed at final merits. Many criteria-step failures were close: 260 (37%) had two criteria met and fell one short, 201 (29%) had one, and 235 (34%) had none. The median petitioner in this group claimed five criteria.

Three in four dismissed appeals failed before the final review

EB-1A appeals decided by the AAO, January 2023 – October 2026

All 1,165 appeals

  • Dismissed 902 77%
  • Sent back to USCIS 250 21%
  • Approved 13 1%

902 dismissed

  • Failed at the criteria step 696 77%
  • Failed at final merits 206 23%

696 failed at the criteria step

  • No criteria met 235 34%
  • One met 201 29%
  • Two met, one short 260 37%

Percentages are of the row above. These are appeal outcomes: petitions USCIS had already denied, not its overall approval rate. Criteria-step failures include 8 with a backup final-merits ruling.

Some criteria are far harder than others

The chart shows how often each criterion failed when the AAO ruled on it, in decided appeals (dismissed or approved); remands are excluded.

Membership and original contributions fail about nine times in ten

Share of AAO rulings on each criterion that found it not met

  1. Commercial success (performing arts)93% of 46 rulings
  2. Membership93% of 368 rulings
  3. Original contributions91% of 392 rulings
  4. Awards80% of 386 rulings
  5. High salary75% of 296 rulings
  6. Published material66% of 456 rulings
  7. Leading or critical role58% of 479 rulings
  8. Display at exhibitions43% of 167 rulings
  9. Scholarly articles25% of 392 rulings
  10. Judging others’ work20% of 522 rulings

Decided EB-1A appeals (dismissed or approved), January 2023 – October 2026. A ruling is an AAO finding of met or not met; remands are excluded.

Why each criterion fails

The reasons repeat across hundreds of decisions. Shares below are of the decisions where that criterion was not met; a decision can give more than one reason.

  • Membership (342 failures). The association did not require outstanding achievements of its members (90%), or experts did not judge applicants (59%). Associations open to anyone who registers, pays or holds a degree or licence do not qualify.
  • Original contributions (355). The work was not shown to be of major significance to the field (92%), the support letters were general or conclusory (75%), or use by others was not shown (47%). In 31%, the impact stayed within the petitioner’s employer or clients.
  • Awards (307). The award was not shown to be nationally or internationally recognized (85%) or given for excellence in the field (31%). In 22%, it went to a team, production or company rather than the person.
  • High salary (223). The pay was compared with the wrong group (77%), was not proven (44%), or was compared only with averages (33%).
  • Published material (299). The outlet was not shown to be major media (80%), the article was not about the person (57%), or it lacked an author, date, title or translation (52%).
  • Leading or critical role (279). The organization was not shown to have a distinguished reputation (76%), or the role was not shown to be leading or critical (65%).

Meeting three criteria rarely wins

Of 219 appeals where three or more criteria were met, 206 were dismissed (94%) and 13 approved. Where the AAO ruled on final merits itself, the reasons it gave were that the person was not shown to be at the very top of the field (95% of those dismissals), that the acclaim was not sustained (85%), and that the support letters were weak (78%). In 34% the record showed an early or recent career rather than a sustained one.

An appeal is a second review, not a second chance

The AAO reviews the whole record again. In 172 appeals (14.8%) it withdrew criteria USCIS had already found met: published material (55 times), high salary (47), leading role (45), judging (33), awards (28), membership (20), scholarly articles (16), original contributions (15) and exhibitions (9).

Since October 2024, more petitions reach final merits — and lose there

The share of dismissals decided at final merits rose from 6.0% before September 2023 to 16.6% between the two updates and 26.8% after October 2024 (p < 0.001). Remands fell from 32.2% to 16.9% between the second and third periods (p < 0.001). Approvals fell from 4.1% to 0.8% (Fisher’s exact p = 0.010), though the counts are small: 5, 2 and 6 approvals. These are changes over time, not proof of cause: the decisions cannot show whether the policy updates drove them, and which cases get appealed may also have changed.

Criterion failure rates changed little across periods. The clearest exception is exhibitions, which failed in 25% of rulings between the updates and 51% after October 2024. That is when USCIS clarified that non-artistic exhibitions count only as comparable evidence; the timing matches, but the data cannot show cause.

Avoidable mistakes appear again and again

Shares are of all 1,165 appeals:

  • Evidence dated after the petition was filed, which cannot count: 121 appeals (10.4%).
  • Inconsistent, unverified or questionable documents: 114 (9.8%).
  • Lawyers’ assertions in the brief, which are not evidence: 111 (9.5%).
  • Missing or defective translations and documents: 109 (9.4%).
  • An earlier O-1 or other approval argued as proof, which the AAO does not accept: 106 (9.1%).
  • New claims or evidence first raised on appeal, which the AAO will not consider: 101 (8.7%).
  • An appeal that did not identify any specific error: 78 (6.7%).

By field

Approval rates were low in every field: sciences 1.8% (392 appeals), business 1.3% (303), athletics 0.7% (144), arts 0.4% (260) and education 0% (34). Science petitioners who lost were the most likely to lose at final merits (28% of their dismissals, against 13–22% in other fields).

What to do now

This section is our interpretation of the findings, not a finding itself, and it is not legal advice.

  • Count only criteria you can win. Judging and scholarly articles succeed most often. Membership and original contributions fail at least nine times in ten, so treat them as extras, not as one of your three.
  • Build for the final review from the start. Three criteria only open the door. The decisions turn on whether the whole record shows sustained acclaim at the very top of the field, compared with the top of that field, not with ordinary peers.
  • Prove the measure, not just the fact. A membership needs its admission rules, a salary needs a comparison with top earners in the same role and place, and press needs circulation data and an article about you.
  • Make support letters specific. General praise is the most common reason contributions and final-merits claims fail. Letters need concrete results, uses by others, and why it matters to the field.
  • File only when the record is complete. Evidence dated after filing does not count, and new claims on appeal are not considered. An earlier O-1 approval does not help.
  • Treat an appeal as a full re-review. The AAO can take away criteria USCIS granted. Strengthen the whole record, not just the points USCIS rejected.

Limitations

  • Appeals are a filtered pool. The decisions cover petitions that USCIS denied and the petitioner then appealed. They show why denied petitions fail; they do not give USCIS’s overall approval rate, and approved petitions are not published.
  • Differences over time are trends, not causes. Which cases get appealed, and how USCIS decides them first, can change for reasons other than policy.
  • We see only what the decision says. Redacted decisions do not show the full evidence filed, and the AAO does not always rule on every criterion.
  • AI coding. The coding was done by AI and verified by two independent passes, a review of every disagreement, a word-for-word quote check and a third blind review of 50 appeals.
  • Incomplete final month. USCIS publishes with a lag, so October 2026 is incomplete.

Figures are from AAO decisions published through early October 2026. USCIS adds new decisions every month, and every case turns on its own record. Use this as a map of where petitions fail, not as advice on your own case.

Sources

  1. USCIS — AAO Non-Precedent Decisions, topic “I-140 – Immigrant Petition for Alien Worker (Extraordinary Ability)”, decisions dated January 2023 – October 2026.
  2. 8 C.F.R. § 204.5(h) — extraordinary ability: definition, the ten criteria and comparable evidence.
  3. 8 C.F.R. § 103.2(b)(1) — eligibility must exist at the time of filing.
  4. Kazarian v. USCIS, 596 F.3d 1115 (9th Cir. 2010) — the two-step review.
  5. Matter of Ho, 19 I&N Dec. 582 (BIA 1988) — inconsistencies must be resolved with independent, objective evidence.
  6. Matter of Furtado, 28 I&N Dec. 794 (BIA 2024) — new claims raised for the first time on appeal.
  7. USCIS policy update on EB-1 evidence, September 12, 2023 (summary).
  8. USCIS policy update on EB-1A evidence, October 2, 2024 (summary).
  9. Hun Lee and Stephen Yale-Loehr, A Review of Recent AAO EB-1A Extraordinary Ability Decisions (2020) — an earlier review of 39 decisions.

Related notes