An investigation can be conducted conscientiously and still be constrained by the organization that commissioned it. The investigator may be independent in assessing the evidence, but what has their independence actually allowed them to examine?
“The restaurant has had a few difficulties with Adrian,” the HR director told Nisha. “Make sure you get to the bottom of this one.”
Nisha had handled enough internal investigations to recognize the distinction between a complaint and a conclusion. She made a note of the comment, then turned to the referral pack.
The allegation was straightforward: Adrian, a shift supervisor, had used inappropriate language towards Mia, a junior service employee.
The investigation scope was equally clear. Nisha was to establish whether the alleged language had been used and whether it breached the hotel’s standards.
Under Background, the referral mentioned previous concerns about Adrian’s manner towards junior staff.
Nisha knew the hotel’s procedures. She was employed in central HR, reported through its management structure and regularly worked with managers who referred cases to her. Her role was to investigate disciplinary matters and provide findings for management to consider. She had no personal relationship with Adrian and no instruction to reach a particular outcome.
There was nothing unusual about an organization using its own HR staff to investigate complaints. Internal investigations can be appropriate, efficient and fair. But the absence of a personal conflict did not answer a different question: how much of the situation was Nisha actually independent to examine?
What the witnesses knew
Nisha began with Rachel, the restaurant manager who had received Mia’s complaint.
Rachel explained that Mia had approached her after an exchange with Adrian during service. Mia said Adrian had called her “useless” and asked whether she ever thought before doing something. The exchange had followed an order mistake.
“Did you establish exactly what Tessa heard?” Nisha asked.
“Not specifically.”
“Why not?”
“I didn’t think there was a reason to separate the two.”
Nisha looked down at her notes.
“So Tessa heard the argument, but Mia told her what Adrian had said?”
“Yes. Tessa was there. She heard them arguing and saw how upset Mia was afterwards. Mia told her what had been said.”
Rachel had recorded Mia’s complaint and spoken to Tessa the following morning. She had treated Tessa’s account as support for Mia’s description of the incident. She had not intended to misrepresent anything. To Rachel, the distinction had not seemed important at the time.
Nisha made sure it was recorded.
Tessa’s interview confirmed that she had been nearby. She heard raised voices and saw Mia upset afterwards. She had not heard Adrian use the word “useless”. Mia had told her what she said had happened.
Tessa believed Mia. She had no reason to think Mia was inventing the exchange. But she could not independently confirm the disputed words.
Ben’s account concerned a different occasion. He described Adrian speaking dismissively to a new employee some months earlier. Ben had found the incident uncomfortable and remembered the employee being upset. He had not witnessed the exchange between Adrian and Mia.
The employee involved in Ben’s account had not been interviewed.
Adrian accepted that he had corrected Mia over the order and that he had been impatient. He accepted raising his voice. He denied calling her useless.
“I should have handled it differently,” he said. “I was trying to get the order sorted before the whole service went sideways. That doesn’t mean I called her useless.”
Nisha recorded both the admission and the denial.
The interviews gave her a reasonably clear picture of what each person could establish. There had been an argument. Adrian had raised his voice. Mia had been upset. The disputed words remained disputed.
Tessa could confirm the argument and Mia’s distress, but not the words. Ben could describe an earlier incident, but not this one.
Nisha could distinguish those things. The question was whether the investigation’s wider framing would preserve the distinctions.
The request for context
After the interviews, Nisha spoke again with the HR director.
“Tessa supports Mia’s account, doesn’t she?” the director asked.
“She supports that there was an argument and that Mia was upset afterwards,” Nisha replied. “But she didn’t hear the words Mia says Adrian used. Mia told her what had been said.”
There was a pause.
“Right. But she supports the account of the incident.”
“She supports parts of it. Not the disputed words.”
The HR director did not argue with her.
“Make sure the report reflects the full context, including the previous concerns. Ben’s account is relevant to that.”
Nisha understood the point. The hotel was concerned not only with the exchange itself, but with what it might indicate about Adrian’s conduct towards junior employees.
That was not an unreasonable concern. Previous incidents can matter. A pattern of conduct may be relevant when an organization is assessing a complaint.
But there is a difference between recording previous concerns and treating them as established findings. There is also a difference between using earlier information as background and investigating whether that information is accurate, complete or representative.
Nisha’s terms of reference did not ask her to investigate the history of those concerns. She was to establish whether Adrian had used the alleged language towards Mia and whether that breached the hotel’s standards.
She could record Ben’s account. She could identify its limits. She could not turn the investigation into a review of every earlier concern, nor had she been asked to establish how those concerns had originally been raised, recorded or assessed.
The distinction mattered. The previous concerns were entering the investigation as context, but their status had not been established by the investigation itself.
What the report said
Nisha’s report included the accounts she had gathered. It also contained qualifications about what the witnesses had and had not heard.
But the summary brought the material together in a way that made the overall picture sound more conclusive:
The allegation was supported by the accounts of more than one member of staff.
Tessa confirmed that she had been present during the exchange, heard the raised voices and observed Mia’s distress afterwards. She subsequently supported Mia’s description of Adrian’s behaviour.
Ben’s account provided further context regarding Adrian’s manner towards junior employees. He described an earlier occasion on which Adrian had spoken to a new employee in a dismissive and intimidating manner.
Adrian accepted that he had corrected Mia during the exchange and that he had raised his voice, although he denied using the specific words alleged.
Taken together, the accounts indicate that the incident was consistent with concerns previously raised about Adrian’s conduct towards junior staff.
The summary did not say that Tessa had heard the disputed words. It did, however, say that she supported Mia’s description of Adrian’s behaviour.
It did not say that Ben had witnessed the incident. It described his account as further context.
Each statement had a basis in something Nisha had been told. The difficulty lay in what the statements collectively appeared to establish.
The referral had described previous concerns as background. In conversation, HR had called them relevant context. The report now said the incident was consistent with concerns previously raised. Its opening sentence said the allegation was supported by the accounts of more than one member of staff.
The language had moved.
Each formulation was doing different work. Background placed information alongside the complaint. Relevant context invited the investigator to take it into account. Consistent with concerns previously raised went further: it drew an inference connecting the present incident to an existing picture of Adrian.
No one had formally established that the earlier concerns were findings. Yet the summary now used them to make the current allegation appear part of a corroborating pattern. The language had moved from describing what the organization knew to suggesting what the evidence showed.
No one had needed to tell Nisha to misrepresent the evidence. She had been conscientious in gathering and recording it. She had identified the limitations of the witness accounts. Yet the summary brought those accounts together under a description that could leave a reader with a stronger impression of corroboration than the evidence warranted.
The problem was not simply that one sentence was imprecise. It was that the investigation had inherited a picture of Adrian in which the current complaint and previous concerns were already connected.
Nisha could examine the evidence supplied to her. She could not necessarily examine the decisions that determined what evidence would be supplied in the first place.
What Nisha could not examine
The employee involved in Ben’s account had not been interviewed.
Nisha could record that limitation. But she had not been asked to establish why the employee was absent, whether the original concern had been accurately recorded, whether other people had seen the incident, or whether there was information that might alter Ben’s account.
Nor had she been asked to examine how the earlier concerns had accumulated into the description of Adrian in the referral.
That history had entered the investigation as context. Its accuracy and completeness had not been established by the investigation itself.
This is where the distinction between an investigator’s independence and an investigation’s independence becomes important.
Nisha was free to assess the evidence she received. She could challenge a witness, distinguish direct knowledge from second-hand information and record where an account did not establish a disputed fact.
But she depended on the organization for the referral, the scope, the access to witnesses and the material made available to her. She worked within its reporting structure and knew that management would make the final decision.
None of that automatically made her biased. It did, however, shape the boundaries within which her independence operated.
An investigator can be independent in how they assess the evidence, yet dependent on others for what evidence they get to assess.
And an investigator can be independent of the person being investigated without being independent of the organization whose handling of the situation may itself need examining.
The organization may have a legitimate reason to limit an investigation. Not every complaint requires a full review of every previous disagreement or management decision. Scope has to be proportionate.
But if previous concerns are being used to strengthen the case against someone, their status and provenance matter. If the investigator is not authorized to examine them, that limitation should be visible in the report. Otherwise, material that has not been tested can acquire the authority of material that has.
The investigation may be careful within its boundaries while leaving the most consequential assumptions outside them.
Questions worth asking
When reviewing an internal investigation, it is not enough to ask whether the investigator was personally impartial or whether the interviews were conducted properly.
It is also worth asking:
- Who defined the scope? What was the investigator asked to establish, and what was excluded before the investigation began?
- What material did the investigator receive? Original complaints and witness accounts, or summaries prepared by people who had already interpreted them?
- Who selected the background information? Could the investigator check its accuracy, completeness and original context?
- What status was given to previous concerns? Were they allegations, informal reports, established findings, or simply matters that management had heard about?
- Could the investigator examine the organization’s own handling of the situation? Or was the organization’s account of that handling treated as part of the starting material?
These questions do not assume that an internal investigator is compromised. They help establish what kind of independence the investigation actually had.
A report should also state what the investigation was not asked to examine. That is not an admission of failure. It allows the reader to distinguish findings from assumptions, and tested evidence from background material whose accuracy or completeness remains unverified.
If you are the person being investigated, it is reasonable to ask what the investigator’s terms of reference cover, what previous concerns have been supplied, and whether those concerns are being treated as established facts or as untested background. If the organization is relying on broader concerns, ask whether you will be told what they are and given an opportunity to respond.
Where the consequences are serious, consider seeking advice from a union representative or an appropriately qualified employment lawyer. The relevant procedures and legal rights will depend on the circumstances and jurisdiction.
Nisha had identified the limits of the evidence. She had recorded them in the investigation. But the summary brought the material together as support for a wider picture of Adrian’s conduct.
The investigator had assessed the evidence she was given. The investigation had not examined how that picture had been assembled in the first place.
That is the difference between an investigator’s independence and an investigation’s independence.
Investigations – How Conclusions Are Formed
Part of a five-article series examining how investigations frame questions, assess evidence, form judgements and affect the people and organizations involved.
- When an Investigation Starts with Its Answer
- When the Evidence Starts Agreeing
- When the Person Becomes the Allegation
- When the Investigator Is Already Inside the Story– You are here.
- When the Investigation Becomes the Burden
Related reading: When You Wonder if You’re Becoming the Scapegoat — examining the distinction between legitimate accountability and becoming the explanation for a wider organizational failure.