Where the two-step analysis comes from
In Kazarian v. USCIS, 596 F.3d 1115 (9th Cir. 2010), the court set out a two-part approach that USCIS adopted for adjudicating extraordinary ability petitions. Step one asks whether the petitioner has submitted evidence meeting at least three of the ten regulatory criteria. Step two is a final merits determination.
At step two the officer weighs the record as a whole, asking whether it demonstrates sustained national or international acclaim and that the petitioner is among the small percentage who have risen to the very top of the field.
The two steps ask genuinely different questions. The first is largely mechanical — does this evidence fit this criterion. The second is evaluative, and evidence that technically satisfies a criterion can carry very little weight there.
Three criteria met is a threshold, not an approval. The final merits determination is where the case is actually decided.
Why petitions fail at step two
The most common pattern is a record assembled to tick boxes. Each criterion is satisfied by the thinnest evidence that arguably qualifies — a couple of reviews, a publication list, a membership — and nothing in the record speaks to standing in the field. Every box is ticked and the overall picture is of a competent researcher rather than an exceptional one.
The second pattern is evidence with no independent component. Letters from supervisors and collaborators, citations from the immediate circle, work published but not taken up. Sustained acclaim is a claim about how the field responded, and a record with no independent voices in it cannot evidence that.
The third is absence of context. Numbers presented without field baselines, journals named without evidence of their standing, awards listed without explaining what they are or who else receives them. An officer who cannot situate the evidence cannot weigh it favourably.
What a record that survives step two looks like
It has depth on at least some criteria rather than the minimum on all of them. A sustained reviewing record across recognised venues, a body of work with genuine independent uptake, letters from people with no stake in the outcome who can point to specific effects of specific work.
It supplies context throughout, so that every number and venue can be located against what is normal in the field. And it tells a coherent story — a line of work, a problem the field cared about, evidence that others followed.
That kind of record is built, not assembled. Which is the honest summary of this entire cluster: almost everything that makes a publication record persuasive at step two is a consequence of decisions taken years earlier about what to work on and where to publish it.