Approach the CFI as a decision exam: given a moment in an interview, choose the prompt, the phase, and the documentation standard that protect the account. The working chain is: funnel from open invitations down to closed questions; keep cognitive interview retrieval separate from credibility judgment; present evidence non-accusatorily; harden vulnerable-witness interviews with ground rules; record key exchanges verbatim with observations separated from inferences. You are ready to move from review to timed practice when you can fill the decision table blank, rewrite an accusatory exchange into a non-accusatory funnel, and mark every contaminating question in a vulnerable-witness script.
Choosing the Next Prompt: Funnel Order Over Question Lists
Plan every interview as a funnel: open invitations first, then specific cued questions, then closed questions last. Practice scenarios hinge on matching prompt type to the interview's current phase, not on reciting definitions.
An open invitation — tell me everything you saw at the loading dock — asks the subject to generate an account without supplied details. Free recall is the highest-quality material an interview produces: it is organized by the witness's own memory rather than your assumptions. Every specific detail you introduce before that recall is complete risks contamination, because the witness may fold your words into the account and later cannot say which details were originally theirs.
Specific cued questions come second, and they are built only from the witness's own words: you mentioned a red van — what did you notice about it? Closed questions come last and carry a single fact each: was the door locked? A leading question states or implies the answer — did he leave through the fire exit? — and is the prompt type to flag in any practice scenario, because it supplies content the witness may simply accept rather than recall.
Cognitive Interviewing and Credibility Assessment Do Different Jobs
The cognitive interview is an information-gathering method built on varied retrieval; credibility assessment is an evaluation of content against case facts. Treat them as separate phases so retrieval quality is never sacrificed to suspicion.
Core cognitive interview moves include context reinstatement (walking through the setting and your state at the time), varied retrieval (recalling events in a different order or from a different perspective), and focused active listening that follows up only on what the witness actually said. These techniques raise the amount and specificity of recall without introducing new facts, which is why they anchor the information-gathering phase.
Credibility assessment compares what the account says against what the case can verify: timelines, physical evidence, other statements. Behavioral observation — fidgeting, gaze aversion, hesitation — belongs in a separate, cautious category, because deception research has repeatedly found such cues weak and unreliable. Treat an observed behavior as a reason to gather more information, never as a conclusion, and let content checked against evidence carry the weight.
Scenario: A Fact-Finding Interview That Slides Into an Accusation
In a theft interview, jumping from open questions to a direct accusation pressures the subject, invites retraction, and weakens the record. Stay in information-gathering until the account itself, tested against evidence, justifies a shift.
Paper scenario: a subject has given an open account of their shift. The interviewer interrupts: let me stop you — we know the drawer was $340 short. Did you take it? That is a plausible instinct after a vague answer, but it stacks three problems at once: it is closed, it is leading, and it moves the interview from gathering to accusing in a single breath. Any admission that follows is now entangled with the accusation, and the record shows the pressure that produced it.
The better sequence keeps the non-accusatory posture: present the discrepancy as something to explain — the count shows $340 short on Tuesday, walk me through how the drawer was handled — then follow the answer with open prompts. Record the exchange verbatim, note the time, and avoid promises or minimization that could bear on voluntariness. The difference matters because a clean, voluntary account survives later scrutiny, while an admission extracted mid-accusation may not, and your documentation will show exactly how it was obtained.
Scenario: A Child-Witness Interview Contaminated by One Question
With a child or vulnerable adult, a leading question about the central allegation can shape the account permanently. Set ground rules first, ask one fact per question, and return to open invitations after any correction.
Paper scenario: a child mentions the garage and stops. The interviewer asks, did he touch you in the garage? That question introduces the act, the person, and the location in one turn. Vulnerable witnesses are especially susceptible to suggestion, and a yes can be assembled from the question rather than from memory — a defect that no later re-interview fully repairs, because the added details now sit inside the child's own account.
The stronger approach opens with rapport and ground rules: some questions I get wrong, so if I make a mistake, tell me, and if you don't know, say don't know. Then return to the child's own words: you said garage — tell me about the garage. Follow-ups carry one fact each and use only details the child has already offered, with every question and answer documented verbatim. The ground rules give the witness a script for correcting you, which is the practical safeguard against contamination.
Documentation Standards Your Testimony Can Actually Support
Documentation should let you testify to what was asked, what was answered, and in what order. Verbatim question-and-answer records beat polished summaries, because summaries erase the exact wording testimony turns on.
Write contemporaneous notes and quote key exchanges word for word, especially invitations, corrections, and any admission. Separate observation from inference in the record: the subject paused before answering and looked at the floor is an observation; the subject was being deceptive is an inference, and a report that dresses inferences up as facts will read badly under cross-examination. Timestamped sequencing also protects you when testimony asks what came first.
Testimony then follows from the record: you can refresh your memory from your notes, quote the questions you asked, and acknowledge the limits of what you observed. If your credibility assessment relied on content rather than behavior, say so plainly; if you noted an impression, present it as an impression. Ethical duties here — candor about the record, fair handling of recordings, disclosure obligations — vary by role and jurisdiction, so anchor your standard to the rules that govern your own work.
A Decision Table for Prompt Choice and Phase Shifts
Drill this pairing until it is automatic: each row is a moment in an interview, the default move, and the error that moment tempts. Defaults are conditional — the subject, the evidence, and your legal context can change the right choice.
Cover the middle columns and quiz yourself: given the moment, what is the default move and why? Then reverse the drill — take a flawed exchange from any practice script and identify which row it violates. The table's value is as a decision drill rather than a memorization list, because the reasoning being trained is matching a messy scenario to the right row.
Note the conditionality built into every row. A closed question is the wrong default early with a cooperative witness but the right tool when you need a precise amount or time. Presenting evidence non-accusatorily assumes you have evidence worth presenting and a subject able to respond to it. Where your jurisdiction's rules restrict questioning practices or require particular procedures with minors, those rules override any default here — which is why the legal and ethical foundations topic deserves its own study pass.
| Interview moment | Default move | Why it works | Error to avoid |
|---|---|---|---|
| Cooperative witness, interview start | Open invitation: tell me everything about... | Maximizes uncontaminated free recall | Opening with a specific allegation |
| Witness gives a partial account | Cued prompt built from their own words | Extends recall without adding facts | Suggesting details they never mentioned |
| Precise fact needed (time, amount) | Closed question, one fact only | Fast and unambiguous | Bundling two questions together |
| Account conflicts with case evidence | Present the discrepancy, ask for an open explanation | Tests the story without accusation | Direct accusation framed as a yes/no |
| Child or vulnerable adult | Ground rules first, then open invitations | Builds correction norms, lowers suggestibility | Skipping rapport to reach the allegation |
| Admission or correction occurs | Record verbatim, note the time, no minimization | Preserves voluntariness and wording | Promising leniency or softening the record |
| Preparing the report | Verbatim Q&A for key exchanges | Preserves what testimony depends on | Converting everything to a polished summary |
A Four-Week Prep Sequence With a Self-Check Rubric
Rotate through the six topic areas while drilling scenario decisions weekly. Score yourself on decisions made, not terms recalled, and treat any rubric score as a learning milestone, not a pass prediction.
A four-week sequence that adapts to your timeline: in the first week, map the six topic areas and write your own one-line definition plus one example for each named technique — open invitation, cued prompt, leading question, context reinstatement, varied retrieval. In the second week, drill the decision table until you can fill it from a blank grid. In the third week, rewrite flawed dialogues: script a five-line accusatory exchange, then convert it into the non-accusatory funnel. In the fourth week, practice documentation and run a mixed review against your practice-question results.
The core exercise: take any practice question you answered by elimination, reconstruct the underlying scenario, and write the interviewer's next three prompts yourself. Expected observations in a strong attempt: prompts move from open to specific, no prompt introduces a fact the subject has not mentioned, a correction from the witness triggers a return to open recall, and any evidence is presented without an accusatory label. Miss one of those and you have located a specific technique to redrill, not a vague weakness.
- You can define open invitation, cued prompt, closed, and leading questions, each with a written example, without notes.
- You can rewrite a five-line accusatory exchange into a non-accusatory funnel and justify each change in one sentence.
- In a child-witness script, you can mark every question that introduces a fact the witness never mentioned.
- Your practice report quotes key exchanges verbatim and separates observations from inferences.
- You can state in one sentence why retrieval techniques and credibility judgments occupy separate phases.
References and further reading
Use these references to explore the concepts and check the latest information from the relevant organizations.