Credibility Is Not Character: A Better Way to Write Credibility Findings

“I just didn’t find her credible.”

I have heard some version of this sentiment more times than I can count, and it is almost always offered as a conclusion based on a gut feeling. Whenever I hear it, I always ask the same, simple question: Why?

That question is what this article is about. “Credibility” is a critical factor for investigators to consider in their analysis, but a credibility determination must be grounded in a thoughtful analysis of objective information. Too often, credibility findings in investigations get used to mean trustworthy, likeable, consistent, reliable, not lying, and sometimes just the witness I believed. When investigators conflate those things, the report stops being an analysis, becomes a character reference, and loses its defensibility.

The real question for an investigator is whether a particular statement, from a particular witness, on a particular point, is reliable enough to support a finding. This is a much narrower question than, “is this person credible?”

Credibility is about statements, not souls

The most common error I see in investigation reports—including ones I have been asked to pressure-test for clients before a hearing or litigation—is the global credibility finding. The report says something like:

“The Investigator finds the Complainant to be credible. The Investigator finds the Respondent to be not credible.”

That sentence, standing alone, is almost never defensible. It treats credibility as a status the witness carries into the room, rather than something the evidence reveals statement by statement. It also invites the reader to disagree on the same global terms: I found him credible. I found her credible. The investigator has just turned a fact-finding process into a popularity contest when they could have shown the reader the basis for their conclusion.
A more disciplined finding sounds like this:

“The Investigator finds the Complainant’s account of the April 20, 2025, incident reliable on the point of non-consent based on the contemporaneous text to her best friend, the corroborating apology message from Respondent, and the internal consistency of her account across two interviews. The Investigator finds Complainant’s account less reliable on the timing of the conversation with her parents because she gave inconsistent dates and the timeline is contradicted by the dated photographs in Exhibit 14.”

A credibility analysis may reveal that the same witness is reliable on one point and unreliable on another. The next step is to weigh the reliability of each piece of information to determine the facts by the applicable standard of proof.

Inconsistency is evidence, not a conclusion

The second discipline problem in credibility writing is treating inconsistency as a finding rather than as a data point that still has to be weighed.

I have read reports where the investigator or fact finder wrote, effectively: the witness gave a different date in the second interview, so the witness is not credible. This reflex can be tempting for practitioners, especially where it confirms someone’s gut instinct (bias) about what they think happened and who is “telling the truth.”

Inconsistency is a piece of evidence that needs to be examined alongside everything else, including:

• Whether the inconsistency is on a material point or a peripheral one
• Whether the witness had a plausible reason for the inconsistency (memory, trauma response, time passage, leading question in the first interview)
• Whether the witness self-corrected, or only changed the story when confronted
• Whether the inconsistency runs in one direction (helpful to the witness’s interest) or in directions that cut against the witness
• Whether the rest of the account, including the parts the witness had no reason to invent, is corroborated

A witness can be inconsistent on the date and entirely reliable on what happened. A witness can be perfectly consistent and still wrong, because they rehearsed a flawed account. Treating “consistency” as a proxy for “truth”—or “inconsistency” as a proxy for “lie”—is one of the fastest ways to produce a finding the parties will live with but the record will not support.

The standard credibility factors the EEOC, California CRD, and the standard California jury instructions have historically identified—plausibility, motivation and bias, demeanor (with caution), consistency, corroboration, and the witness’s opportunity to perceive the event—are a good inventory to work from. An investigator must exercise discipline applying them to the specific statement at issue, not to the witness as a person.

A witness can be imperfect and still be reliable

This is the principle I would want every decision-maker who reads my reports to walk away with. It is also the part that is hardest to write well, because it requires the investigator to acknowledge weakness in a witness without abandoning the witness’s account.

In real cases, witnesses are almost never clean. The complainant was drinking. The respondent has a documented temper. The corroborating friend dislikes the respondent’s family. The supervisor missed a meeting last year and is defensive about it. None of that disqualifies their accounts. All of it should be acknowledged, weighed, and explained.

A report that pretends a favored witness has no weaknesses is less credible. A report that disqualifies a witness because of an unrelated flaw is less defensible. The version that is both honest and defensible is the one that says: here is what cuts against this witness, here is why I do not find that dispositive on this point, and here is the evidence that ultimately moved me.

That is showing your work—the same discipline that makes the rest of the investigation defensible. Credibility is no different. It just feels harder, because the investigator or fact finder is being asked to make a judgment call in writing, in the middle of a contested matter, knowing that the disappointed party will read it carefully looking for a reason to reject it.

What this means for decision-makers and counsel

If you are a general counsel, Title IX coordinator, HR executive, or board member reading a report before you have to act on it, here are three questions worth asking the investigator before you adopt the findings:

• Are credibility findings made at the statement level, or only at the witness level? Global findings are a red flag, not a feature.
• For each disputed material fact, can you point me to the specific evidence—corroboration, inconsistency, motive, opportunity to perceive—that the investigator weighed? The report should let you answer that without a follow-up call.
• Where the investigator credited one account over another, did the report acknowledge the weaknesses in the credited account and explain why they were not dispositive? If not, the finding will not hold up under serious review.

These are the same questions a hearing officer, reviewing court, OCR investigator, or opposing counsel will ask. Better to ask them yourself, while the report can still be tightened.

---

A credibility finding fully exposes an investigator’s judgment, laid out on the page for any reader, including a hostile one, to test. And that is uncomfortable, but it is also a necessary part of the work. A report that shirks the analysis and hides behind, “I found the witness credible,” is asking the reader to take one on faith. A report that names the statement, identifies the relevant evidence, weighs it openly, and explains what moved the needle is doing a proper job.

Credibility is not character. It is reasoning, made visible, one disputed fact at a time. Build the report that way, and the finding will hold.

If your organization is facing a sensitive matter and you want an investigator whose credibility analysis will hold up under scrutiny—or a thinking partner to pressure-test one that has already been written—reach out to Grey Zone.

---

About the Author

---