How to Write Investigation Findings That Hold Up
The language investigators use in their conclusions — and how they connect findings to evidence — determines whether a report survives legal scrutiny.
An investigation report can document every interview, collect every relevant document, and reach the right conclusion — and still fail under adversarial scrutiny if the findings section is written poorly. Courts and agencies evaluating challenged investigations don't just ask whether the employer got the answer right. They ask whether the reasoning is traceable, the evidentiary basis is explicit, and the language used to express conclusions is calibrated to what the evidence actually supports. The mechanics of how a report is written matter as much as what it says.
The Core Problem: Findings That Outrun the Evidence
The most common vulnerability in investigation reports is a mismatch between what the evidence shows and the strength of the conclusion drawn from it. A finding that states "Respondent sexually harassed Complainant" when the evidence supports a narrower conclusion — say, that Respondent made two comments the Complainant found offensive — overstates the record. In litigation, that gap invites attack. Defense counsel can demonstrate that the stated finding isn't supported by the underlying facts, which undermines the entire report's credibility, including conclusions that are well-supported.
Calibrated findings language maps conclusions to evidence precisely. Where evidence is strong and corroborated, findings can be stated with corresponding confidence: "The weight of the evidence supports a finding that..." Where the record is mixed or credibility-dependent, the language should reflect that: "Based on the available evidence, the investigator finds it more likely than not that..." The preponderance-of-the-evidence standard that governs most employer investigations is a more-likely-than-not threshold — findings language should reflect that standard, not overstate it.
Connecting Findings to Evidence: The Citation Function
Every material finding should be traceable to a specific evidentiary basis. In practice, this means citing the interview, document, or exhibit that supports the conclusion — not in footnote-heavy academic style, but in direct prose: "Complainant's account is corroborated by the text messages exchanged on March 14 (Exhibit 3) and by Witness A's independent description of the same incident."
This practice serves two functions. First, it allows a reviewing court or agency to evaluate whether the conclusion is actually grounded in the record — courts have consistently distinguished between investigations that articulate a reasoned basis for their conclusions and those that assert conclusions without explanation. In Baldwin v. Blue Cross/Blue Shield of Alabama, 538 F.3d 1329 (11th Cir. 2008), the court noted the importance of employers documenting the basis for personnel decisions in discrimination cases. Second, explicit citation prevents a common post-hoc problem: when a case is litigated years later, the investigator may not remember why they believed one witness over another. A report that shows its work makes reconstruction possible.
The same principle applies to negative findings — conclusions that a violation did not occur. "The investigator finds insufficient evidence to substantiate the complaint" needs to explain why: what evidence was considered, what it showed, why it was not sufficient, and whether the complainant's account was found not credible, not corroborated, or simply inconclusive. Bare conclusions of non-substantiation are a recurring weakness that agencies flag during audits.
Timeline Documentation Within the Report
Investigation reports often include a narrative of events — what happened, in what order. Timeline accuracy matters beyond factual accuracy: courts look at whether the employer understood the sequence of events when it acted. A report that reconstructs a timeline incorrectly, or that omits key dates, can undermine a legitimate business justification by making it appear that the employer acted without full knowledge of the facts.
Timelines should be sourced. Rather than "the incident occurred in early March," a defensible report states "according to Complainant's account (Interview, March 22) and the badge-access log for March 7 (Exhibit 1), the incident occurred on the evening of March 7." The EEOC's enforcement guidance on harassment emphasizes that employers must have a system for investigating complaints that allows them to determine what occurred — timeline documentation is part of showing that the system worked.
Policy Application and Legal Standard Language
Findings sections typically include a determination of whether conduct violated company policy, applicable law, or both. These are separate questions, and conflating them creates risk. A report that concludes "Respondent violated the company's anti-harassment policy" is making a factual and policy determination the employer is positioned to make. A report that concludes "Respondent engaged in unlawful sexual harassment under Title VII" is making a legal conclusion — one that may be contested, that depends on legal standards outside the investigator's expertise, and that can create awkward inconsistencies if litigation reaches a different conclusion.
The SHRM guidance on investigation reports advises investigators to make findings relative to policy rather than law, reserving legal conclusions for counsel. Findings that track specific policy language — quoting the relevant policy provision and explaining how the conduct maps to it — are both more defensible and more practically actionable: they give the employer a basis for remedial action regardless of how a court would ultimately characterize the conduct.
What Good Practice Looks Like
A defensible findings section states what the investigator concluded, explains the evidentiary basis for that conclusion, acknowledges contrary evidence and explains how it was weighed, applies the relevant standard explicitly, and stops short of legal conclusions. It reads as a reasoned analytical document, not a narrative summary or an advocacy brief. Investigators who build these habits into their drafting — treating each finding as a mini-argument that must be supported, not asserted — produce reports that hold up not because they reached the right answer, but because the reasoning is visible and can be evaluated.
Marshal's structured findings builder prompts investigators to link each conclusion to a specific evidence reference before the report can be finalized — building the citation discipline that courts look for directly into the drafting workflow.
[ Field Notes ]
Once a week — investigation technique, employment law, and AI in HR practice. No fluff.