What evidence actually counts

The instinct is to submit everything. Resist it. Officers have limited time per petition, and hundreds of pages of loosely connected material makes your strongest evidence harder to find, not easier.

Every exhibit should serve an identified proposition. If you cannot say in one line what a document proves, cut it.

Match the evidence to the endeavor

This is the decision that matters most, and getting it wrong is a documented failure mode.

Your endeavor is…	Officers weigh	De-emphasise
Academic research	Independent citations, peer-reviewed publications, journal review invitations, competitive grants, evidence others build on your work	Commercial metrics
Applied / industry	Adoption of your tools or methods, documented outcomes, patents, commercialisation, recognised-outlet media	Publication counts
Entrepreneurial	Traction, revenue, users, funding, letters of interest, business plan, evidence the venture operates	Academic markers
Teaching / practice	Licensure, documented shortage in the specialty, institutional demand, outcomes	Citations
Arts / culture	Exhibitions, performances, commissions, critical reception, national reach	Publications

denied.md is explicit on the mismatch: if the endeavor involves entrepreneurship, industry leadership, or applied professional services, publications and citations are less important, and emphasising achievements unconnected to the proposed work dilutes the narrative.

Citations: useful, not required

For research endeavors, independent citations are the clearest objective marker that your work influenced the field beyond your own group. Self-citations are discounted.

But low counts are not disqualifying, and this is worth internalising because it causes needless despair. In our case database:

• A radiation oncology assistant professor was approved in two months with an explicitly reported zero citations and no RFE.
• A materials science professor was approved with 190 publications and 30 citations, a ratio well under one.
• A doctoral researcher was approved with one undergraduate publication and two citations, having deliberately never argued citations at all, treating it as the weakest part of the profile.
• Meanwhile a petitioner with 250+ citations and NIH-funded research was denied, on prong 1.

Citations are one signal among several. Deciding what not to argue is part of strategy.

Substitutes for citations

Where the citation record is thin, independent recognition can come from:

• Peer review or editorial service for journals or conferences
• Competitive grants awarded by peer review
• Senior membership in a professional body where selection is merit-based (IEEE Senior Member, for example, appears in an approved profile in our corpus that explicitly reported few citations and no prizes)
• Adoption of your tools, methods, or standards by others
• Media coverage in recognised industry outlets
• Speaking invitations at established venues
• Awards where the selection basis is documented
• Third-party public records naming you

That last category is underused and disproportionately valuable, because it is the one thing you cannot generate yourself.

Letters of interest

For a new venture with limited traction, these are what reduce an officer's uncertainty. They are not required in every case, a long documented track record can substitute, but the closer your endeavor is to "ambitious and new," the more they matter.

The clearest example in our corpus: a software engineer's petition drew an RFE on national importance, and was approved after the response added a state Chamber of Commerce letter showing local demand, an independent economist's analysis, and expert opinion letters confirming feasibility. Not more credentials. Third-party validation of demand.

Organise for the officer, not for completeness

• Number exhibits systematically and add a short explanation of what each one proves. case.md recommends exactly this, and the reason is that officers have limited time.
• Provide an exhibit index that maps each item to the element it supports.
• Reference exhibits inline in the petition letter, so the officer never has to search.
• Paginate the assembled packet continuously.

Build a master timeline first

Before assembling anything, build one timeline of your education and employment: dates, titles, duties, locations. Then reconcile every document to it, résumé, letters, forms, and prior filings including the DS-160.

Inconsistencies across documents are a leading RFE cause. Officers read the documents together, and a four-month unexplained overlap between two jobs reads as an error that puts the whole timeline under scrutiny. Where a real discrepancy exists, explain it in the petition rather than leaving it to be found.

Quality over volume, concretely

denied.md puts submitting too much evidence first in its list of denial causes. The reasoning is not that officers are lazy, it is that excessive material obscures the strongest arguments, and an argument an officer does not find is an argument you did not make.

A concise, well-curated packet in which every document serves a purpose is more persuasive than a disorganised mountain of paperwork.

Next: recommendation letters that work · browse the case database