An investigation needs a working theory. It also needs a deliberate way to discover that the theory is wrong.
Confirmation bias becomes a concern when an investigator gives greater attention or weight to information supporting an existing belief while overlooking information that challenges it. During an interview, that can influence which questions are asked, which answers receive follow-up, and how uncertainty is described afterward.
The practical response is to make alternative explanations part of the work. The College of Policing identifies “consider the opposite” as one possible strategy for reducing confirmation bias during investigative hypothesis development. College of Policing: Making Good Decisions
Separate the event from the explanation
Consider a hypothetical investigation into an unauthorized entry.
An access record shows that a credential assigned to an employee was used after normal operating hours. A supervisor believes the employee returned to remove confidential material.
Those are separate propositions:
- A credential was recorded as being used.
- The assigned employee personally used it.
- The entry was unauthorized.
- Confidential material was removed.
- The person intended to act improperly.
The access event may justify further investigation. It does not establish the remaining propositions by itself.
Before the interview, identify which propositions are supported, which are disputed, and which have not yet been tested. That separation helps prevent an allegation from becoming an assumed fact inside the question plan.
Write down what would change the working theory
A useful preparation exercise is to complete this sentence:
“If our current explanation is wrong, we would expect to find…”
Possible answers in the hypothetical case might include an approved maintenance task, evidence that another person used the credential, an incorrect system clock, or records showing that no material left the premises.
These are possibilities to test, not excuses to accept automatically. Their value is that they give the investigation a route away from its first explanation.
For each plausible explanation, identify:
- What information would support it.
- What information would contradict it.
- Which source could distinguish between competing accounts.
- What remains unknowable from the available records.
An alternative that cannot explain the known facts may be rejected. The important point is to assess it rather than leave it unexamined.
Ask for an account before supplying the conclusion
Compare these questions:
“Why did you return to take the documents?”
“Please describe your movements after your scheduled shift ended.”
The first question assumes both a return and a purpose. The second allows the person to provide an account that can then be clarified and checked.
Open questions do not eliminate the need for precise follow-up. They create space to learn what the person actually says before narrowing the discussion. Guidance from the College of Policing also warns that misleading question wording can introduce information into a witness’s account. College of Policing: Non-Leading Questioning
A disciplined sequence might be:
- Invite an account of the relevant period.
- Clarify locations, times, people, and actions.
- Identify the person’s source of knowledge.
- Explore discrepancies against specific records.
- Record unresolved differences without forcing agreement.
The objective is an account that can be tested.
Keep observations separate from findings
Interview notes can quietly turn an interpretation into a fact.
“The employee became evasive when asked about the records room” is an interpretation. It does not explain what happened or allow another reviewer to assess it.
A more useful record describes the exchange: the employee asked which date was under discussion, paused, and said they could not recall entering the room. The investigator can then identify what requires clarification and which records may assist.
The same discipline applies to voice-based review. A potential stress-related indicator does not establish its cause. Treating it as confirmation of an existing suspicion would add an unsupported inference to the investigation.
Use a second reviewer to test the reasoning
A second review is most useful when it has a defined task.
Rather than asking whether a colleague agrees with the conclusion, ask them to identify:
- Assumptions presented as established facts.
- Questions that contain an allegation.
- Explanations that were raised but not checked.
- Evidence inconsistent with the preferred theory.
- Conclusions that exceed the underlying record.
Where practical and consistent with agency procedures, the reviewer can assess the factual sequence before reading the first examiner’s interpretation.
Agreement between colleagues is not independent proof. The purpose is to expose gaps in reasoning and identify further checks.
Where AISIS and VASIS fit
AISIS and VASIS are intended to support trained, examiner-led investigative work. Their use does not remove the need to test alternative explanations or distinguish an observation from a verified fact.
An examiner should be able to explain the next investigative step without claiming that a system independently resolved the case. That step might be a clearer question, a document comparison, another witness interview, or a decision that the available material does not support further interpretation.
The standard is whether the investigation can show how it tested the explanation, including information that challenged it.
The hypothetical example and suggested review questions illustrate an investigative approach. They are not a validated VIPRE examination protocol.
