Which category is denied most often?

Key answer

USCIS's own figures show an enormous spread, and it is widening. In the first three quarters of fiscal 2026, 59.3% of EB-1A decisions and 50.9% of EB-2 NIW decisions were denials, against 6.2% for EB-1B and 2.8% for EB-1C. A year earlier, in fiscal 2025, it was 33.1% and 44.8%; in fiscal 2024, 27.8% and 29.0%. This is the share of decisions issued that were denials, not the share of petitions filed.

Check your NIW profile against the Dhanasar test

Breakdown from the data

The answers below are translated from the original Russian - the language of the community chats.

EB-1A: 12,231 approved, 4,703 denied - 27.8% of decisions are denials.
EB-2 NIW: 27,526 approved, 11,256 denied - 29.0%.
EB-1B: 2.0%. EB-1C: 3.8%. More than a tenfold difference from EB-1A.

This breakdown is built on real USCIS notices and public sources, linked below. It is reference material, not legal advice: officers decide similar cases differently.

Why EB-1B and EB-1C are refused several times less often

The difference is not that these categories are 'easier'. It is who files them and how. EB-1B and EB-1C are filed by an employer rather than by the beneficiary, and almost always through a legal department or outside counsel who has done it many times. The packet is assembled from a proven template.

EB-1A and EB-2 NIW are self-petition categories. They contain many filings by people assembling documents for the first time, often without an attorney. On top of that, EB-1A has no objective threshold like 'so many years in a managerial position': the assessment is qualitative, and therefore contestable by nature.

Hence a practical conclusion: 27.8% is not your probability of refusal. It is an average across a very heterogeneous mass of applicants. Your risk depends on which half of that mass you resemble.

How to read these percentages correctly

What is counted is decisions issued during the year, not petitions filed during the year. Adjudication lags filing, sometimes by more than a year, so the denominator holds cases filed at different times. You cannot say from this data that 'such-and-such a percentage of petitions filed are refused'.

Second: a denial and an RFE are different things. A request for evidence is invisible in these statistics. A case that drew an RFE and was answered successfully lands in the approved column. So the share of people who met difficulty along the way is noticeably higher than the share of denials.

Third: this is data for applicants worldwide, with no breakdown by country or profession. For your situation it sets context, not a forecast.

What to do with this when choosing a category

If you are choosing between EB-1A and EB-2 NIW, the difference in denial share between them is small - something else matters far more: EB-1A needs three closed criteria out of ten plus sustained acclaim, while NIW is judged against the three prongs of the Dhanasar test and requires no criteria at all.

If you have an employer willing to file and you fit EB-1B or EB-1C, the statistics speak fairly loudly in their favour. But you cannot file EB-1B yourself: the employer is the petitioner, and that is a condition, not a formality.

And in general: the difference between categories is smaller than the difference between a well-assembled and a badly-assembled packet within one category.

Sources

USCIS: Form I-140 case-status data (quarterly)
The agency publishes per-quarter counts of received, approved, denied and pending petitions for each sub-category.
USCIS: the EB-1 category
Official description of EB-1A, EB-1B and EB-1C.
USCIS Policy Manual, Volume 6, Part F, Chapter 2
The EB-1A evaluation standard.

All questions on: EB-1A All questions on: EB-2 NIW All questions on: Choosing a category All questions on: Process and timelines