An HR investigator and a homicide detective ask similar questions for similar reasons: who, what, when, where, in what sequence, with what corroboration. The methodology that wins suppression hearings in criminal court is the same methodology that produces investigations defensible at EEOC, in arbitration, and at trial. Most workplace investigations underperform because they do not use it.
The Six Mistakes That Sink Workplace Investigations
1. Leading questions.
A leading question implants a fact and asks the witness to confirm it. "He raised his voice at you, didn't he?" is a leading question. So is "When he made the comment about your hair, were others present?" — because it presupposes that the comment was made.
Loftus's misinformation research (Loftus, 1975; Loftus & Palmer, 1974) demonstrated that even minor wording changes alter what witnesses later report remembering. In a workplace investigation, leading questions create an evidentiary trail that opposing counsel will use to argue the investigator authored the account. Open-ended invitations — "Tell me what you observed" — produce the witness's words, not the investigator's.
2. Witness contamination.
Witnesses talk. They text. They share notes about what they were asked. Once two witnesses have compared accounts, both statements are partially compromised. Investigators should: interview each witness separately and as soon as practicable; instruct each witness, on the record, not to discuss the substance of the interview with other employees during the pendency of the investigation; and conduct interviews in a sequence that minimizes shared exposure.
3. Confirmation bias.
Investigators commonly enter the interview with a working theory — usually drawn from the complaint itself. The risk is that the working theory drives question selection: questions that confirm the theory get asked, questions that might disconfirm it do not.
The corrective is structural. A written interview plan should include open-ended questions that, if answered honestly, would tend to support each plausible version of events. Investigators should also document the questions they did not ask and why — this is rarely done, but it is the documentation that survives bias allegations.
4. Confusing rapport with interrogation.
Rapport is not a softening tactic to manipulate someone into talking. It is the working condition under which accurate information is exchanged. The investigator who is warm but rigorous, who treats the interviewee as a person with information rather than a target with a story to break, gets a fuller account.
The flip side is also true. Investigators trained in confrontational interrogation models — accusation, theme development, false-evidence ploys — generate documentation that looks coercive on review. Workplace investigations are not criminal interrogations. Treating them as such produces unenforceable findings, retaliation claims, and constructive-discharge exposure.
5. Combining roles without disclosing them.
When the same HR business partner serves as the witness's prior coach, the investigator on the matter, and the recommender of discipline, the investigation cannot survive an impartiality challenge — even if the substance was fair. Investigators should disclose role boundaries at the beginning of each interview and avoid combining adjudicative and advisory functions on the same matter.
6. Documentation that paraphrases.
A summary memo that says "the witness confirmed the complainant's account of the meeting" is unusable in litigation. Defensible documentation captures: the question asked, in the words asked; the answer given, in the words given; specific dates, times, and locations referenced; the witness's stated basis for knowledge (observed, heard about from someone, inferred); and the investigator's contemporaneous notes, signed and dated.
The EEOC and Litigation Backdrop
The EEOC's enforcement guidance on harassment and the Faragher/Ellerth affirmative defense (Faragher v. City of Boca Raton, 524 U.S. 775 (1998); Burlington Industries, Inc. v. Ellerth, 524 U.S. 742 (1998)) make clear that an employer's investigation must be prompt, thorough, and impartial. "Thorough" and "impartial" are not abstract standards. They are operationalized in the interview record.
In Vance v. Ball State University, 570 U.S. 421 (2013), and the steady line of harassment cases since, the courts have repeatedly looked at the investigation file to determine whether the employer met its obligations. The interview transcripts and contemporaneous notes are usually the load-bearing evidence.
Where the investigation produces an adverse finding, the employee subject to discipline frequently challenges that finding through arbitration, internal appeal, or civil suit. The investigation file is produced. If the interviews show leading questions, missing follow-up, contamination, or paraphrased documentation, the underlying finding becomes very difficult to defend.
A Defensible Interview Structure
A workplace investigation interview that holds up under review tends to follow a recognizable shape:
- Opening administration. State the purpose, scope, and your role. Confirm the participant understands the process is impartial. Address confidentiality realistically — most jurisdictions limit blanket nondisclosure instructions, so the language should mirror current NLRB guidance.
- Rapport and orientation. Establish the interviewee's role, tenure, and reporting relationship. This is also where the witness's credibility-relevant baseline is established.
- Free narrative. An open invitation: "Tell me what you observed about [specific event or matter]." Allow uninterrupted account before targeted questioning.
- Targeted, open-ended follow-up. "What did you see next?" "Who else was present?" "How did you know that?"
- Clarification of inconsistencies. Specific, neutral, late in the interview. "Earlier you said X. A moment ago you said Y. Help me understand."
- Closing inventory. "Is there anything I have not asked about that I should know? Is there anyone else I should speak to? Is there a document I should see?"
- Documentation. Verbatim where possible, contemporaneous, signed and dated.
Investigators Need Different Training Than They Are Getting
Most HR investigator training programs allocate the bulk of their hours to law and policy — what counts as harassment, what the regulatory tests are, what the documentation templates look like. The interview itself is often a small module taught generically.
The result is investigators who can recite the law but cannot conduct an interview that tests their hypothesis without contaminating the witness. The methodology used by senior investigators in serious criminal matters — the Enhanced Cognitive Interview, structured non-confrontational questioning, deliberate sequencing, and rigorous documentation — translates almost cleanly into the workplace context. It is, in fact, what the leading workplace-investigation guidance documents describe in different language.
The Bottom Line
A workplace investigation does not need to look like a criminal investigation. It needs to be conducted with the same discipline. Leading questions, contaminated witnesses, biased question selection, and sloppy documentation are not policy questions. They are skill questions. Skills are trainable.
Key Sources
- Faragher v. City of Boca Raton, 524 U.S. 775 (1998).
- Burlington Industries, Inc. v. Ellerth, 524 U.S. 742 (1998).
- Vance v. Ball State University, 570 U.S. 421 (2013).
- U.S. Equal Employment Opportunity Commission, Enforcement Guidance on Harassment in the Workplace (April 2024).
- Loftus, E. F., & Palmer, J. C. (1974). Reconstruction of automobile destruction. Journal of Verbal Learning and Verbal Behavior, 13(5), 585–589.
- Fisher, R. P., & Geiselman, R. E. (1992). Memory-Enhancing Techniques for Investigative Interviewing. Charles C. Thomas.
- Society for Human Resource Management (SHRM), Conducting Workplace Investigations resource library.