He Said, She Said: Making and Defending Credibility Determinations

The standard: more likely than not, not beyond reasonable doubt

Ensure you apply the correct standard for review. An investigator who imports criminal-law instincts paralyzes themself. A workplace investigation asks whether it is more likely than not that the conduct occurred. This is a preponderance of the evidence standard, not proof beyond a reasonable doubt. “Inconclusive” has become the refuge of investigators applying the wrong standard. This happens when an investigator is unable to be certain and declines to decide, and the organization treats the non-findings as exonerations. Sometimes an inconclusive finding is honest. But it must be a conclusion reached, not a decision avoided. When reaching an inconclusive finding, the file should show the same rigor behind “we cannot determine” as behind any other finding. And, an inconclusive finding may mean remedial measures should still be imposed like office-wide or department-wide training, clarified expectations, and heightened attention moving forward.

The factors, in the order you should trust them

 Review recognized credibility factors.

1. Corroboration. Review not just the eyewitnesses, but also the quieter forms of evidence like the contemporaneous outcry to a friend, the text sent that night, the calendar entry, and/or the pattern of avoiding a particular shift. Contemporaneous evidence is powerful precisely because it often predates any motive to litigate or complain.
2. Inherent plausibility. Ask yourself, which account better fits the undisputed facts, the documents, the timeline?
3. Consistency. Has each party’s account held its shape across tellings?
4. Motive to fabricate. Examine this in both directions, honestly. The complainant facing a performance review and the accused facing termination both have stakes.
5. Pattern. Prior complaints about the same actor, properly considered, are legitimate evidence.
6. Demeanor. Confident liars and nervous truth-tellers are both abundant. Cross-cultural demeanor readings are notoriously unreliable. If demeanor appears in your analysis at all, describe behavior, not character.

These factors align with how the Equal Employment Opportunities Commission (EEOC) directs its investigators to assess credibility.

Writing the determination

The credibility section of your report should read like a path, not a verdict (i.e., here are the accounts, here is where they conflict, here is the evidence bearing on each conflict, here is how the factors cut, and here is the finding). Name the evidence that cuts against your conclusion and say why it did not carry the day. A file that acknowledges its hard points is a file that survives hostile reading, while the one that avoids them is the one that looks rigged. Never resolve credibility by rank (i.e., a director would not lie), by likability, or by consequence (i.e., a finding would end his career). Each of these reasoning patterns have sunk real employers in real cases.

What can employers do now to enhance their credibility assessments?

1. Put the standard for evaluation in writing in your investigation protocol and train investigators out of the certainty instinct.
2. Build a credibility-factors analysis into your report template. Include things like plausibility, corroboration, consistency, motive, and pattern so the analysis is structured, uniform, and visible.
3. Require a history check in every interpersonal-misconduct investigation. Prior complaints involving either party should be reviewed and considered, with the consideration documented.
4. Treat every inconclusive finding as requiring a documented response plan so a “couldn’t determine” outcome never reads as “didn’t care” or a “less than thorough evaluation.”

Next installment: Writing the report. What to write, what to leave out, who reads it, and how a document meant to end the matter becomes the backbone of the defense.