Five files can still be one point of view

My EB-1A petition included five supporting letters and seventeen media and online mentions, including two awards.

Those counts helped describe the record.

They did not tell me how many independent perspectives the record contained.

Five files can repeat one announcement. Three pages can trace back to the same organization. Two letters can rely on the same secondhand account. A folder can grow while the underlying support stays almost unchanged.

That is why document count and corroboration are different questions.

The first asks how many items you collected.

The second asks who created each item, what that source was positioned to know, and whether it adds anything the other sources do not.

I needed both views of the evidence.

A source chain can look like a source set

Imagine an organization publishes an announcement.

An event page repeats it. A partner copies part of it into a newsletter. A profile later summarizes the same announcement.

You now have four files.

But the central fact may still come from one original statement.

That does not make the later pages useless. They may help establish dates, public visibility, or context. They may preserve details that disappeared from the first page.

They should not automatically be treated as four independent observations of the same result.

The useful question is:

**Where did this source get the fact?**

If the answer keeps leading back to the same announcement, the files belong to one source chain.

Counting the chain once does not mean discarding every page. It means being honest about the structure of the support.

Different sources were useful because they saw different things

The evidence in my petition did not all need to repeat the same sentence.

A media article could show how a piece of work was described publicly at the time.

An event or program record could help establish that something happened, when it happened, and how it was presented.

A supporting letter could explain responsibility from a writer's direct line of sight.

An award page could identify the recognition and the body that gave it.

Those sources become more useful when their jobs are kept separate.

The media article may not know the full internal division of responsibility.

The colleague may know the work closely but should not be treated as an independent authority on every claim about the organization.

The program page may document an event without explaining what changed because of it.

The award page may confirm a name on a list without proving the significance of the underlying work.

Corroboration does not require every source to say the same thing.

It requires each source to contribute within the limits of what it could know.

Repetition can hide behind different formats

Source dependence is not always obvious from the file type.

A PDF, a webpage, and a screenshot may feel like three kinds of evidence. If all three capture the same organization-written biography, they are three formats of one account.

The reverse can also be true.

Two documents that look similar may contain genuinely different perspectives. A contemporary article may describe an initiative from outside the organization. A letter written years later may explain the writer's direct knowledge of my role. Both discuss the same work, but their vantage points and evidentiary jobs are different.

This is why I would not classify independence by appearance.

I would trace authorship, origin, and line of sight.

A narrow source can add something important

It is tempting to prefer the source that makes the biggest statement.

The more useful source may be the one that adds one fact no other document can supply.

A dated page might add the period.

A participant record might add direct confirmation that an event happened.

A letter might add the responsibility the public page could not see.

A media article might add outside context that an internal record could not provide.

None of those sources has to prove the whole claim.

The set becomes stronger when the pieces have distinct, bounded jobs.

That approach also makes weak support easier to spot. If four files all prove the date but none shows my responsibility, adding a fifth date source does not close the gap.

The missing perspective matters more than the document count.

I would separate origin from contribution

For each source, I would record two different things.

First: where did it come from?

Second: what does it uniquely add?

Those questions prevent two opposite mistakes.

The first mistake is overcounting. Several files inherit the same fact from one source, but the petition presents them as separate confirmation.

The second mistake is underusing. A source gets dismissed because it does not prove the main conclusion, even though it provides a date, role, relationship, or outside description that the rest of the record needs.

A source can be dependent for one fact and still add an independent fact of its own.

For example, a partner page may repeat an organization's program description but independently show that the partner promoted the event on a particular date. The repeated description and the dated public record are not the same contribution.

The distinction belongs at the claim level, not only at the file level.

Build a source-dependency map

Choose one material claim from your petition.

Create one row for every source you plan to use and add these columns:

1. **Source creator:** Who wrote, issued, or maintained it?
2. **Fact used:** What exact fact are you taking from it?
3. **Basis of knowledge:** Did the creator observe the fact, record it as part of the work, or repeat it from somewhere else?
4. **Upstream source:** Does this item point back to another source in the set?
5. **Unique contribution:** What does this item add that would disappear if you removed it?
6. **Boundary:** What does the source not establish?

Then draw an arrow whenever one item depends on another for the fact you are using.

Look for two problems.

The first is a cluster of files with arrows leading to one origin and no other perspective.

The second is a material part of the claim with no source positioned to know it directly.

Do not solve the first problem by counting the cluster more loudly.

Do not solve the second by asking an existing source to say more than it can know.

Narrow the claim, find a different perspective, or state clearly what the current record supports.

A petition can contain many documents.

What matters during this audit is whether the documents give the reader more than one copy of the same point of view.