Why the total is not the argument
Citation counts appear in almost every NIW and EB-1A petition, and in a surprising number of requests for evidence. The reason is that a total—"the applicant's work has been cited 480 times"—does not by itself answer either legal question. For NIW, the second prong asks whether the applicant is well positioned to advance the endeavor, which turns on who has acted on the work. For EB-1A, the original-contributions criterion asks whether the work is of major significance in the field, and the final-merits review asks whether the applicant stands at the top of it. Officers have seen enough petitions to know that 480 citations means different things in oncology and in topology. The petition has to say what the citations show.
Independent citations and collaborator citations
The first distinction is between citations by the applicant's own group—self-citations and citations by co-authors—and citations by researchers with no connection to the applicant. Independent citations are the ones that demonstrate reliance by the field, and petitions that report them separately, with a clear method for how they were counted, are more credible than petitions that report only the total. A record in which most citations come from the applicant's former lab says something different from one in which most come from groups on three continents. Counting is tedious but mechanical: list the citing papers, mark the authors who have co-authored with the applicant, and report both figures with the date and the database used.
Field norms and percentiles
A citation count means little without a baseline. The useful baselines are field-specific and time-specific: how many citations a paper in this field, published in this year, typically receives, and what threshold places it in the top one or ten percent. Commercial databases publish such thresholds by discipline; a petition can use them if it identifies the field consistently and applies the right year. Two errors are common. One is comparing the applicant's total to an "average researcher" figure drawn from a different discipline or an all-time average. The other is claiming a percentile that the data do not support—"top 1% of scientists worldwide"—from a ranking that measures something else. An h-index can be presented with the same care: as a figure in context, with the field and career stage stated, not as a verdict.
Read the citing papers
The most persuasive evidence in a citation record is not a number but a sentence from a citing paper: a group that used the applicant's method, compared results against the applicant's benchmark, adopted the applicant's dataset, or built a system on the applicant's design. Reviews that mention the work in passing count as citations but prove little; papers that depend on the work prove reliance. For the applicant's main contributions, the petition should identify several citing papers of this kind, quote the relevant passage accurately, include the paper as an exhibit, and explain in plain language what the citing group did with the work. This is where we spend much of the time on a petition, and it is the part officers most often respond to.
Impact beyond journal articles
Citations in patents, clinical guidelines, standards documents, government reports and policy papers show that the work has moved from the literature into practice, and they are often stronger evidence of national importance than another journal citation. The same is true of software downloads and forks, dataset usage statistics, incorporation into commercial products, and adoption by named institutions. These are not always captured by citation databases, so they have to be found: by searching patent databases, by asking the applicant where the work has been used, and by asking letter writers to describe use they have seen. Each instance should be documented with a source the officer can check.
Presenting the record
Use one database consistently and state which one and the date; Google Scholar counts run higher than Web of Science or Scopus, and mixing them invites a credibility question. Present the data in tables an officer can read in a minute: citations by paper, independent versus collaborator citations, citing institutions by country, and the field threshold used for any percentile claim. Attach the underlying reports as exhibits. Make sure the field named in the citation analysis is the field named everywhere else in the petition. And describe what the numbers mean in one paragraph of plain language, because the officer will not do that work for you.
References
How Lumen Law & Co. can help
Citation analysis is strongest when it explains how other work used the applicant’s research—not merely how many citations appear in a database.
- Organize citation records and distinguish independent, collaborative, and duplicate references.
- Read relevant citing papers and identify actual use, validation, or development of the applicant’s work.
- Consider field- and publication-year context, along with patents, standards, and practical adoption.
- Integrate accurate citation examples into the petition’s broader evidence narrative.