The field has settled into two camps, and both are wrong.
One treats trauma-informed practice as a reason to soften scrutiny. Inconsistencies are not probed. Difficult questions are not asked. The account is accepted largely as given, because testing it is understood to be harmful.
The other treats the whole idea as advocacy wearing the costume of method. Interviews are conducted as though the objective were to catch someone out, on the theory that rigour requires coldness.
Both produce bad investigations, and both misunderstand what the practice is for. Trauma-informed interviewing is not a concession to the complainant. It is how you get accurate information from a person whose recall is affected by stress.
Care in the room and rigour on the page are not competing obligations.
What the practice actually claims
Stated narrowly, and this is the only version worth defending: people under acute stress encode and retrieve memory differently from people at rest. Accounts given under those conditions are frequently fragmented, non-linear, strong on sensory and peripheral detail, and weaker on sequence and duration. An interview conducted without accounting for that will produce a less accurate account, not a more honest one. The Department of Justice puts the purpose the same way: a trauma-informed interview is designed to give the person the best opportunity to recall accurately, and it improves the evaluation of the evidence rather than softening it.DOJ, Framework for Prosecutors, 2024
That is a claim about interview technique. It is not a claim about who is telling the truth. The same federal guidance frames the objective as building provable cases in a trauma-informed manner while ensuring due process, which is precisely the pairing the two camps treat as impossible.DOJ, Framework for Prosecutors, 2024
What it does not claim
It does not claim that inconsistency is evidence of trauma. It does not claim that a fragmented account is therefore truthful. It does not claim that the investigator should believe the complainant. And it does not license the abandonment of testing, corroboration, or credibility analysis.
Practitioners who overstate the science do real damage here, because the overstatement is the part that gets attacked. An investigator who writes that inconsistencies are consistent with trauma, and stops there, has offered a conclusion that a competent cross-examination will take apart, and has weakened a finding that may well have been correct.
The defensible position is narrower and stronger: an inconsistent account is not, by itself, a dishonest one, and the investigator has to establish which it is rather than assuming either.
An inconsistent account is not a dishonest one, and refusing to test it serves nobody.
Interviewing that is both humane and testing
The two goals are met by structure, not by temperament. What makes an interview humane is largely the same thing that makes it productive.
Open the account before you shape it
Begin with an uninterrupted free narrative. Let the person describe what happened in their own order, without questions, without correction, and without being redirected to chronology. Interrupting to establish sequence early is the single most common way investigators lose detail they will never recover.
Ask open questions before closed ones
Move from open to specific, not the reverse. Closed and leading questions contaminate recall, and once contaminated the account cannot be uncontaminated. The order of your questions is part of the evidentiary quality of the answer.
Return to inconsistencies rather than avoiding them
Every material inconsistency must be put to the person, clearly, with the opportunity to address it. Doing this respectfully is a matter of framing rather than avoidance. You are asking because your report has to explain it, and saying so plainly is usually enough.
Give control back where it costs nothing
Breaks, pacing, the order of topics, who else is in the room, how long the session runs. None of these affect the accuracy of the account and all of them affect whether the person can give it.
The respondent interview is where trauma-informed practice is most often abandoned, and the abandonment is usually unexamined. A respondent may also be under acute stress, may also give a fragmented account, and is owed the same interview quality. An institution that uses one method for complainants and another for respondents has built a process that will not survive scrutiny on exactly that point.
Credibility, and what you may rely on
This is where investigations are won and lost, and where the two failed camps do the most damage.
What an investigator may rely on: internal consistency on material points, consistency with contemporaneous accounts, corroboration by other evidence, plausibility against known facts, and the presence or absence of a motive to fabricate where the record actually supports assessing it.
What an investigator may not rely on: demeanour as a proxy for truthfulness, the absence of emotion, the presence of emotion, delay in reporting treated as inherently probative, or an assumption in either direction about how a person who experienced this would behave.
Demeanour deserves particular attention because it is the most seductive. Investigators believe they can tell. Federal guidance on investigative bias is direct about the risk: assumptions about how a person who experienced this would behave shape credibility judgements and how a report is handled, and policy, training, and supervision have to be built so that they do not.DOJ, Identifying and Preventing Gender Bias, 2022 A determination that rests visibly on how someone seemed is a determination with an obvious line of attack.
The delay question
Delayed reporting is where the two errors meet.
The unrigorous version treats delay as expected and therefore unremarkable, and does not examine it. The hostile version treats delay as evidence that nothing happened.
Neither is analysis. Delay is a fact that requires explanation like any other. Ask why. Record the answer. Assess it against the rest of the record. A reason that is coherent and consistent with the surrounding evidence strengthens the account; one that shifts under examination is relevant, and should be treated as relevant, without being treated as dispositive.
Writing it so both are visible
A report is read by someone who was in neither room. The care that went into an interview is invisible to them unless the report shows it, and so is the rigour.
- Record the method, briefly: how the interview was structured, what was asked openly, and when specific questions were introduced.
- State what was put to the person, particularly the difficult parts. A report that never shows an inconsistency being tested reads as a report that never tested one.
- Explain inconsistencies rather than absorbing them. Say what the inconsistency was, what the person said about it, and what weight it carries.
- Separate what the person said from what the investigator concluded, visibly, so the reader can disagree with the second without doubting the first.
- Say what you did not rely on. A line stating that demeanour was not treated as probative closes a line of attack before it opens.
What an appeal looks for
Appeals in this area rarely turn on whether the investigator was kind. They turn on four things.
- Whether material inconsistencies were identified and addressed, or passed over.
- Whether the credibility determination rested on permissible grounds, stated on the page.
- Whether both parties received the same interview quality and the same opportunity to respond.
- Whether the reasoning can be followed from evidence to conclusion without the investigator present to explain it.
A report that was gentle and unrigorous fails on the first two. A report that was rigorous and brutal often fails on the third, because the respondent or the complainant did not receive a fair opportunity in the room. The method that survives is the one that is careful about both.
Why this is not a compromise
The framing of a trade-off is the error underneath both camps. It assumes that care and rigour draw from the same pool, so that more of one means less of the other.
They do not. A person who is not overwhelmed gives more detail. More detail gives more to corroborate. More corroboration produces a determination that rests on evidence rather than on impression. The humane interview is the one that produces the more testable account.
Conversely, an investigator who avoids difficult questions in the name of care has not protected the complainant. They have produced a finding that will be overturned, in a process the complainant will have to go through again.
Avoiding the hard question does not protect anyone. It produces a finding that will not hold.
Trauma-informed practice, done properly, is not a softer standard. It is a better method for getting at what happened, applied to everybody in the process, and written up so that a reader who was not there can see both the care and the testing.
— Albert Roberson, Fractional Coordinator