Two different questions
EB-1A asks about you: have you risen to the top of your field, and can you show sustained national or international acclaim? The evidence is retrospective. It looks at what you have already done and how the field has responded.
EB-2 NIW asks about your work: is what you propose to do of substantial merit and national importance, are you well positioned to advance it, and would it benefit the United States to waive the job offer requirement? That framework comes from Matter of Dhanasar, 26 I&N Dec. 884 (AAO 2016), and it is forward-looking as much as backward-looking.
Both are self-petitioned, which is why they are so often compared. But "easier" is the wrong axis. A researcher with a modest citation record and a genuinely important research agenda may present far better under Dhanasar than under EB-1A, and the reverse is equally true.
EB-1A is about acclaim already earned. NIW is about the national importance of work still to come. Different questions, not different difficulty levels.
What each asks the publication record to do
Under EB-1A, publications are evidence of recognition. What matters is how the field responded — independent citations, adoption of your method, invitations to review or speak that followed from the work. The papers are the occasion; the response is the proof.
Under NIW, publications do two jobs. They establish that you are well positioned to advance the endeavour, and they help show the endeavour has substantial merit. A body of work in a field with clear national relevance can carry the first two Dhanasar prongs even where the citation counts would not, on their own, support an acclaim argument.
This is why the same set of papers gets framed differently in the two petitions. Nothing about the underlying research changes; the argument built on top of it does.
Where the research plan matters more than the count
For NIW especially, coherence beats volume. A record that traces a clear line — a problem identified, approaches developed, results others have taken up — supports the "well positioned to advance" prong far better than a scattered set of unrelated papers in the same period.
That has a practical implication for anyone still building. Choosing what to work on next with the eventual argument in mind costs nothing at the time and is very hard to retrofit later. It is the single thing we most often wish researchers had thought about eighteen months earlier.