How Facticity Thinks About Truth: Using h-index to Measure Evidence

When most people hear the word “truth,” they think of something binary, true, or false. But in the real world, especially online, truth is often about evidence. How many credible sources support a claim? How well-established are those sources? And is the claim being repeated because it is well-supported, or just because it is loud?

That is where the h-index becomes useful.

The h-index, also called the Hirsch index, was originally created to measure the impact of researchers. In academia, a scholar has an h-index of h if they have published h papers that have each been cited at least h times. It is a simple idea, but a powerful one, because it captures both volume and influence at the same time. One highly cited paper is not enough. Neither is a long list of papers that nobody references. The score only rises when there is both breadth and consistent recognition.

At Facticity, we think this same logic can help evaluate claims.

Instead of asking whether an author has influential papers, we ask whether a statement has strong supporting evidence. A claim can be said to have an h-index of h if h independent sources support it, and each of those sources has itself been cited, referenced, or relied on at least h times. In plain English, that means the statement is not just supported by a few random articles. It is supported by multiple sources that themselves matter.

That makes the concept surprisingly relatable. Imagine hearing a claim online and trying to decide whether to trust it. If only two obscure websites mention it, that is weak support. If twenty respected sources support it, and those sources are themselves widely referenced, that is a much stronger signal. The h-index gives us a way to express that difference with a number.

This matters because the internet is full of claims that can look convincing on the surface. A post can go viral, get repeated, and still rest on very little real evidence. Facticity’s goal is to make that visible. Rather than asking people to trust a vague verdict, the system can show how much support exists behind a statement and how strong that support really is.

There is also an interesting parallel here with PageRank, the system that helped make Google so effective in its early years. Before Google, search engines were often bad at ranking results because they relied heavily on simple keyword matching. PageRank changed that by looking at links between pages. A page became more important if other important pages are linked to it. In other words, authority flowed through the network.

The h-index is different, but complementary. PageRank helps identify which sources are authoritative by looking at how influence passes between them. The h-index looks at whether a claim is supported by enough independent sources, each with enough influence of its own. One is about the quality of the nodes in the network. The other is about the depth and consistency of support behind a specific statement. First, collect the supporting sources for a statement. Then look at how often each of those sources is cited or referenced. Sort those counts from highest to lowest. The h-index is the largest number where at least that many sources each have at least that many references.

Take a simple example. Suppose a statement is supported by sources with reference counts of 50, 30, 22, 10, 8, 5, and 3. Once those numbers are sorted, you check how far the pattern continues. There is at least 1 source with 1 or more references, at least 2 sources with 2 or more, at least 3 with 3 or more, at least 4 with 4 or more, and at least 5 with 5 or more. But there are not 6 sources with 6 or more. So, the h-index is 5.

What that means, in practice, is that the statement has five independent supporting sources, and each of those sources has enough recognition to count as meaningfully established.

That is a much more useful signal than simply saying “five sources mentioned this.” It tells us something about the quality and consistency of the support, not just the quantity.

Like any metric, the h-index is not perfect. It does not replace judgment, context, or source-level analysis. Some fields cite more heavily than others. Some true claims may be new and not yet widely referenced. Some false claims may briefly appear well-supported if poor-quality sources echo one another. But as part of a broader system, it offers something valuable: a transparent, data-based way to estimate how well a claim stands up.

That is really the bigger idea behind Facticity. In a world overflowing with information, people need more than labels. They need to understand why a claim should be trusted, how strong the evidence is, and where that confidence comes from. Metrics like the h-index help move fact-checking in that direction. Not as a replacement for reasoning, but as a clearer way to show the structure of evidence behind what we read.