How to Assess Credibility in Workplace Investigations

When workplace misconduct allegations involve conflicting accounts and no direct witnesses, employers can face significant legal and operational risk if investigators rely on instinct, demeanor, or assumptions to decide whom to believe, explained Shana Bridgeman, a partner at Weiss Serota Helfman Cole + Bierman in Boca Raton, Fla.

Bridgeman conducts workplace investigations for private and public employers throughout Florida and nationally. She discussed credibility determinations in workplace investigations with SHRM.

Specialty Credential: Workplace Investigations

SHRM: How can investigators use corroboration, plausibility, consistency, and motive to evaluate conflicting accounts?

Bridgeman: Start with corroboration in every case. Corroboration is external confirmation of a witness’s account: a text message sent that afternoon, a coworker’s testimony from the day of the incident, a security camera recording, an email thread, a journal entry. When corroboration exists, it carries more weight than any other factor because it moves the analysis beyond one person’s word against another’s. The investigator’s first task, before weighing anything else, is to identify and exhaust every available source. Even in a one-on-one dispute, corroboration is often available. Did the complainant tell anyone at or near the time? Has the respondent made similar comments to other employees?

When there is no corroboration, the other factors carry the analysis in combination. Plausibility is a threshold assessment, not proof: whether the alleged conduct makes sense given what is known about the workplace and the parties. Internal consistency is whether the account holds together across multiple interviews and within a single interview, which is why effective questioning technique, revisiting key topics, and framing questions from different angles, is important.

The investigator should focus on inconsistencies on core points rather than peripheral details and should give the witness an opportunity to explain discrepancies before drawing a negative inference. Motive to falsify is whether either party has a reason to misrepresent what occurred. However, having a motive does not automatically mean the person lied.

The factors interact. They are not a checklist. The question I hear most often from clients is what to do when there are no witnesses and no documents, and the two accounts simply conflict. Many employers treat that as the end of the analysis and find no violation because the complainant has nothing to “back it up.” That is a mistake. Uncorroborated does not mean unproven. In a one-on-one dispute with no corroboration, a plausible, detailed, internally consistent account from a complainant with no apparent motive to fabricate, combined with a vague respondent denial that is inconsistent with other known facts, can support a credibility finding — not because any single factor is dispositive, but because the determination rests on the cumulative weight of the analysis, and the totality of the evidence. The investigator should be able to explain how each factor contributed.

SHRM: Why is demeanor one of the least reliable indicators of credibility?

Bridgeman: Demeanor measures how a person appears, not what happened. Research consistently demonstrates that most people, including experienced professionals, cannot reliably detect deception by observing behavior.

A deceptive witness may present as calm and confident. An honest witness may seem anxious, evasive, or rehearsed simply because being interviewed is stressful, because the witness fears retaliation, or because the witness has never been questioned in a formal setting. Cultural background, personality, mental health, and individual coping style all shape how a person presents, and none of those characteristics is evidence of truthfulness or deception.

Because demeanor is unreliable as evidence, a finding that rests on demeanor is difficult to defend. Plaintiff’s counsel will argue that the investigator substituted instinct for evidence, and that argument is often successful because nothing in the record connects directly to the finding. Demeanor observations may be included in an investigation report, but they must be tied to a specific analytical significance in the specific case, not offered as a general impression, and they must never be the sole basis for a credibility determination. In the investigations I conduct, demeanor is supplemental at most, and some experienced investigators do not consider it at all.

Seminar: U.S. Employment Law & Compliance

SHRM: What mistakes can make an investigation appear biased or predetermined?

Bridgeman: Confirmation bias is the most common error and the most difficult to detect in one’s own work. This can occur when an investigator forms a preliminary theory and then reads ambiguous evidence as confirming it. The antidote is structural. Good intentions are not enough. Look for evidence that contradicts the working theory. Interview every relevant witness before drawing any conclusion. At each stage, ask what evidence would change your mind, and then find out whether that evidence exists.

Asymmetry is the second common error. Every witness must receive the same scrutiny. If the investigator tests the respondent’s account through three follow-up interviews but accepts the complainant’s initial statement at face value, the process is uneven and the finding is open to challenge. The same is true in reverse.

A third error is drafting the report as an argument. This occurs when the investigator, having reached a conclusion, writes the report to support that conclusion. Favorable evidence is emphasized. Contrary evidence is minimized or omitted. The credibility analysis reads as though the outcome were inevitable. The final report is not a brief. It must read as a balanced analysis of what the investigator actually found, including the evidence that points to a different conclusion.

SHRM: How should employers document credibility determinations in anticipation of litigation?

Bridgeman: The report should explain the analytical basis for every credibility finding. Under the Faragher-Ellerth defense [in which a company can block a workplace harassment lawsuit if it proves it tried to stop the harassing behavior, and the worker did not report it], a prompt and reasonable investigation is the employer’s central showing, and courts evaluate the overall reasonableness of that investigation, not its perfection. The written report is the employer’s best evidence that the standard was met. The most common documentation failure is conclusory language. “I believed the complainant” is not a credibility finding; it tells the reader nothing about why. A defensible finding explains the analytical basis: the complainant’s account was corroborated by a contemporaneous text message to a coworker, was consistent across two separate interviews, and matched a pattern of similar complaints, while the respondent’s denial was vague, lacked detail, and contradicted the respondent’s own prior written statement. That level of specificity connects the finding to the underlying evidence and, if the matter goes to court, it provides a basis for a judge or jury to evaluate whether the determination was reasonable.

Equally important, the report must address the evidence that weighs against the finding. An investigator who reaches a determination and omits the contrary evidence produces a document that reads like advocacy. A report that acknowledges the contrary evidence and explains why it was not enough to change the outcome demonstrates the balanced analysis that courts expect.

When credibility cannot be resolved, say so. Not every investigation produces a clear answer. When the accounts are equally plausible and no corroboration favors either one, document that a determination could not be made on the available evidence. That is not a failure of the investigation. It is an honest conclusion. The employer can still take appropriate preventive measures, such as additional training, monitoring, or schedule adjustments, without a substantiated finding. When I advise employers on report structure, I remind them that the reader will be a court, an agency investigator, or an arbitrator who has never met the witnesses, and every finding should be traceable to something in the record.