In most workplace investigations, the outcome turns on credibility. Two people describe the same event differently, no one else was in the room, and the investigator must determine whose account is more reliable. Last spring, Abhishek Ramaswami provided Checkoff readers with a solid overview of investigation best practices, covering investigator selection, interview methodology, documentation, and the investigation process from beginning to end. This article picks up where that discussion left off.
Credibility is an aspect of workplace investigations that is often challenged in subsequent litigation. The framework that follows is designed to provide Florida practitioners with a structured approach to making and documenting credibility determinations that withstand scrutiny. This approach is useful when advising employers on investigation design as well as when evaluating an investigation after the fact.
Why Credibility Determinations Matter
Under the framework established by the United States Supreme Court (SCOTUS) in Faragher v. City of Boca Raton and Burlington Industries, Inc. v. Ellerth (Title VII cases), an employer may avoid liability by demonstrating that it exercised reasonable care to prevent and correct harassing behavior and that the employee unreasonably failed to take advantage of preventive or corrective opportunities. A prompt and reasonable investigation is the central showing under the first prong of that defense. Both private and public employers in Florida operate under this framework because the Florida Civil Rights Act (FCRA) is interpreted in a manner consistent with Title VII. The “reasonable investigation” defense has limits that practitioners on both sides of the bar must understand. It is unavailable when the supervisor’s harassment culminates in a tangible employment action, and it is unavailable when a constructive discharge follows an official act of the employer. However, when the defense is available, the quality of the employer’s investigation is the central battleground.
The Eleventh Circuit’s decision in Baldwin v. Blue Cross/Blue Shield of Alabama is the controlling authority on what “reasonable” means in this context. The court recognized the inherent difficulty of the task, citing the Seventh Circuit’s observation that it is difficult for an employer to sort out charges and countercharges of harassment among employees with competing accounts. Importantly, the standard is reasonableness, not perfection.
In Meritor Savings Bank v. Vinson, SCOTUS noted that the question of whether alleged conduct was unwelcome “turns largely on credibility determinations.” When the investigation’s credibility assessment is challenged as unreasonable, the employer’s entire affirmative defense is at risk. It is important for management-side practitioners to understand how to structure investigations so that credibility findings are defensible. Plaintiff-side practitioners evaluating whether an investigation was reasonable or pretextual will look at how the investigator handled competing accounts. For both, the credibility determination is the point at which the investigation either withstands scrutiny or falls apart.
Framework for Evaluating Credibility
Credibility assessment is not a single judgment. It is an analytical process that requires the investigator to evaluate competing accounts against multiple established factors, determine how those factors interact in the specific case, and reach a conclusion that can be explained and defended. The distinction matters. A checklist of factors to consider is useful as a starting point, but it is not a framework. A framework tells the investigator how to move from raw information to a reasoned determination when accounts conflict.
The analytical sequence begins with corroboration, the single most important factor. Corroboration is external confirmation of a witness’s account: documents, electronic records, video, audio, testimony from someone who was present, or testimony from someone who received a contemporaneous report of the incident. 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 any other credibility factor, is to identify and exhaust all available sources of corroboration. Baldwin itself reinforces this point: the court held that an employer is not required to credit uncorroborated statements disputed by the alleged harasser.
When corroboration is absent or inconclusive, the remaining factors require more extensive analysis. In addition to corroboration, practitioners and courts have long applied a set of established evidentiary principles for evaluating witness credibility: inherent plausibility, demeanor, motive to falsify, and past record of similar conduct. These factors did not originate with any single agency or guidance document; they are analytical tools drawn from the laws of evidence, jury instructions, and decades of adjudicative practice. Their utility in workplace investigations persists regardless of shifts in agency policy.
In that regard, on January 22, 2026, the Equal Employment Opportunity Commission (EEOC) voted 2-1 to rescind its 2024 Enforcement Guidance on Harassment in the Workplace. That rescission, combined with an earlier nationwide injunction against the guidance’s gender-identity provisions, means no formal federal guidance on investigation methodology is currently in effect. However, as noted above, credibility factors derive from the laws of evidence and adjudicative practice rather than from any single agency’s guidance. Thus, they remain analytically sound and are widely applied by courts and trained investigators, irrespective of changes in EEOC enforcement policy.
Investigator neutrality is the precondition for reliable credibility analysis. An investigator who approaches the process with a predetermined conclusion or who treats the accused as presumptively guilty will produce findings that cannot withstand scrutiny. The EEOC’s Acting Chair recently reinforced this principle in the federal sector context, observing that a presumption of innocence for the accused furthers the truth-finding goal at the heart of the investigative process. That principle applies with equal force in private-sector investigations under Title VII and the FCRA. The credibility framework that follows assumes a neutral investigator applying established analytical principles to competing accounts. The factors interact. They do not operate in isolation, and a competent investigator does not simply check each box and tally a score.
Application of the Framework
Consider a common scenario: a complainant reports that a supervisor made sexually explicit comments during a one-on-one meeting. The respondent denies it. No one else was present. The investigator’s analysis should proceed through the factors in a structured sequence.
First, the investigator should look for corroboration. Did the complainant tell anyone about the incident at or near the time it occurred? Contemporaneous corroboration (a text message to a friend, a complaint to a coworker the same day, a journal entry) is not as strong as direct witness corroboration, but it is meaningful because it may occur before a motive to fabricate might develop. Did the respondent make similar comments to other employees? If so, pattern evidence corroborates the complainant’s account even though no one else witnessed the specific incident.
Second, the investigator should evaluate inherent plausibility. Is it reasonable, given the workplace context, organizational structure, and what is known about the parties, that the alleged conduct could have occurred? Plausibility is not proof. It is a threshold assessment of whether the account makes sense on its face.
Third, the investigator should consider the motive to falsify. Both parties may have motives that cut against their credibility. The complainant may have a pending performance issue, and the respondent has an inherent motive to deny. Having a motive does not mean the person lied, but it is a factor the investigator must assess alongside the other evidence. Closely related is witness bias; that is, whether a witness has a personal, professional, or financial relationship with either party that could influence what the witness reports. Motive and bias are distinct concepts, but both bear on whether a witness’s account can be taken at face value.
Fourth, the investigator should examine internal consistency. Is each witness’s account consistent across multiple statements and within any single account? Effective questioning technique drives this analysis. By revisiting key topics at different points during the interview and framing questions from different angles, the investigator develops a sufficient record to assess whether the witness’s account holds together on the points that matter. An investigator should look for inconsistencies on significant points, not minor peripheral details, and should give the witness an opportunity to explain any discrepancy before drawing a negative inference. Memory is imperfect, and people do not tell stories identically every time. Inconsistency on a core element of the account is analytically different from inconsistency on a background detail.
Fifth, the investigator should consider the specificity and detail of the accounts. Specificity and detail are often relevant to evaluating a witness’s account, though they are not dispositive. An account that includes specific contextual details (what the person was wearing, what else was happening in the room, what was said immediately before and after) may carry more weight than a vague or conclusory account, but the investigator must be cautious about treating detail as a proxy for truthfulness.
Sixth, the investigator should evaluate any relevant past record. Has the respondent been the subject of prior complaints or substantiated findings involving similar conduct? Has the complainant filed prior complaints that were found to be without merit? Past history may be relevant, but it is not determinative. A respondent with a prior substantiated complaint is not automatically guilty of the current allegation; a complainant who filed a prior unfounded complaint is not automatically dishonest now. The investigator must evaluate past history as one factor among several, not as a shortcut to a conclusion.
Seventh, the investigator may consider demeanor. How did the witness present during the interview, and how did the witness respond to questions? A witness who is evasive, who repeatedly fails to answer direct questions, or whose manner of response changes markedly when the topic shifts to the contested events may be less credible than one who responds directly and consistently. That said, demeanor is the weakest credibility indicator and the most subjective. Research consistently demonstrates that many people cannot accurately detect deception simply by observing behavior. A deceptive witness may appear calm and credible; an honest person may be anxious simply because the investigation process is stressful. Cultural and individual differences in communication style further complicate demeanor assessment. Investigators who rely heavily on demeanor expose their findings to challenge on exactly this basis. Demeanor may be treated as supplemental, not foundational, and should never be the sole basis for a credibility determination. For these reasons, some investigators choose not to consider demeanor at all.
The credibility framework’s strength lies in the interaction of these factors. Consider an investigator facing a one-on-one dispute with no corroboration. The complainant’s account is plausible, detailed, and internally consistent. The complainant has no apparent motive to fabricate. The respondent’s denial is vague and inconsistent with other known facts. Under these circumstances, the investigator is in a strong analytical position, not because any single factor is dispositive, but because the cumulative weight of the analysis supports a credibility finding. The determination rests on that cumulative weight.
Documenting Credibility Determinations
The investigation report will be closely scrutinized in litigation. How the investigator documents credibility findings determines whether the employer can defend the investigation’s outcome.
The most common documentation failure is conclusory language.“I believed the complainant” is not a defensible credibility finding. It tells the reader nothing about the basis for the belief. A defensible finding explains the analytical basis; e.g., the complainant’s account was more plausible because it was corroborated by a contemporaneous text message to a coworker, was internally consistent across two separate interviews, and was consistent with a pattern of similar complaints about the respondent, while, on the other hand, the respondent’s denial was vague, lacked specific detail, and was inconsistent with the respondent’s own prior written statement. That level of specificity connects the credibility finding to the analytical framework and gives the reader (and, if necessary, a court) a basis for evaluating whether the determination was reasonable.
Equally important, the report must address evidence that cuts against the finding. An investigator who reaches a credibility determination but omits discussion of contrary evidence produces a document that reads like advocacy, not analysis. The obligation to address contrary evidence is what distinguishes a reasoned conclusion from a predetermined one. As previously noted, Baldwin instructs courts to evaluate the overall reasonableness of the investigation. A report that acknowledges the contrary evidence, explains why it was insufficient to change the determination, and walks through the reasoning demonstrates the kind of balanced analysis that courts are seeking.
When credibility cannot be resolved, the report should say so. Not every investigation produces a clear winner. When the evidence is evenly balanced, the accounts are equally plausible, and no corroboration tips the scale, the investigator should document that a determination could not be made on the available evidence. This is not a failure of the investigation; it is an honest conclusion that reflects the limits of the evidence. The employer can still take appropriate preventive measures (additional training, monitoring, schedule adjustments) even without a substantiated finding.
Demeanor observations, when included, require context.“The complainant appeared nervous” means nothing without an explanation of why that observation is analytically significant in the specific case. Many witnesses are nervous during an investigation interview regardless of whether they are telling the truth. If demeanor is documented at all, it should be noted alongside the more reliable factors and should never carry disproportionate weight in the written analysis.
Common Pitfalls
Several recurring errors undermine credibility determinations even in otherwise competent investigations.
Confirmation bias is the most pervasive error. It arises when the investigator forms a preliminary conclusion early in the process and then interprets ambiguous evidence to confirm it. Confirmation bias is unintentional, which makes it difficult to detect in one’s own work. It affects both what the investigator looks for and how the investigator interprets what is found. The antidote is structural: the investigator should consciously seek evidence that contradicts the emerging theory and should interview all relevant witnesses before reaching any conclusion.
Equating a witness’s demonstration of emotion with credibility is a fundamental analytical error. A complainant who is visibly distressed is not necessarily credible; a respondent who is calm and composed is not necessarily truthful. Emotional presentation is not evidence. It can be a reaction to the investigative process and can be influenced by personality, culture, and individual coping mechanisms. The investigator who equates distress with truthfulness has substituted assumption for analysis.
Overeliance on demeanor is a related but distinct problem. As discussed above, demeanor is the weakest credibility indicator. An investigator who documents demeanor prominently and other factors superficially invites the argument that the determination was based on gut instinct rather than evidence.
Failure to investigate inconsistencies occurs when the investigator identifies a discrepancy in a witness’s account but does not follow up. Noting an inconsistency without giving the witness an opportunity to explain it, and without pursuing additional corroboration, leaves the credibility determination incomplete. The inconsistency may have a reasonable explanation, and the investigator will not know unless the question is asked.
Applying different standards to the complainant and the respondent is a fairness error that can undermine the entire investigation. For example, if the investigator conducts three detailed follow-up interviews with the respondent to test the respondent’s account but accepts the complainant’s initial statement without similar scrutiny (or the reverse), the process is asymmetrical, and the finding is vulnerable to challenge.
Drafting the final report as an argument is a common pitfall that occurs when the investigator, having reached a conclusion, writes the report in support of that conclusion rather than as a neutral analysis. The report reads like a brief: favorable evidence is emphasized, contrary evidence is minimized or omitted, and the credibility analysis is presented as though the outcome was obvious. This is distinct from confirmation bias, which affects the investigation itself; the problem here is the tone of the report. The final investigation report is not a brief. It must be read as a balanced, reasoned analysis, not as an advocate’s closing argument.
Conclusion
The legal standard for the adequacy of an investigation is reasonableness, not perfection. Credibility assessment is at the core of determining reasonableness. Investigators who use a structured analytical framework and document their reasoning transparently will produce investigations that withstand scrutiny. Those who rely on gut instinct and conclusory findings produce investigations that do not.
The framework is not complicated. It requires methodical analysis, not expertise in psychology. Start with corroboration. Work through the established analytical factors. Explain how the factors interact in the specific case. Document both the evidence that supports the finding and the evidence that cuts against it. When credibility cannot be resolved, say so.
Florida practitioners advising clients on designing investigations, or on evaluating investigations after the fact, should treat credibility methodology as the central question, not an afterthought. The credibility determination is where the investigation succeeds or fails, and it is the determination that courts will examine most closely when the investigation is challenged.
This article was originally published by The Florida Bar Labor and Employment Law Section in the June 2026 issue of The Checkoff.
