Investigative Interviewing of Children and Vulnerable Witnesses

The intuition that children are unreliable witnesses is common and partially correct. Children, particularly younger children, are more susceptible to leading questions, more likely to acquiesce to authority figures, and more likely to incorporate post-event information into their accounts. But the research record is substantially more nuanced than a blanket unreliability claim.

Free recall accuracy. Across a substantial research literature, children's free recall of events they witnessed or experienced is generally accurate, though less complete than adults. In the absence of suggestive questioning, children as young as three to four years can provide accurate accounts of experienced events (Goodman and Reed, 1986; Baker-Ward, Gordon, Ornstein, Larus, and Clubb, 1993). The problem arises not in free recall but in response to leading and closed questions.

The suggestibility gradient. Younger children are more suggestible than older children, and older children are more suggestible than adults. The mechanism involves multiple factors: greater deference to authority figures; less understanding of the questioner's fallibility; a more limited event vocabulary requiring reliance on offered words; and less developed source-monitoring ability (the capacity to distinguish between what was experienced and what was suggested or imagined). The Bruck, Ceci, and Hembrooke (1998) landmark review of the McMartin and similar cases documented the systematic effect of repeated, suggestive questioning by adults committed to a particular outcome.

The developmental calibration imperative. Interview protocols designed for adults perform poorly with children. Adults who are trained for adult interrogation will ask abstract, multi-part, and leading questions that overwhelm the child's linguistic and cognitive processing. Structuring a child interview requires understanding what cognitive and linguistic competencies are present at each developmental stage, and using only the questioning forms that the child can handle accurately.

Trauma-informed considerations. Children disclosing abuse are not in a neutral cognitive state. They may be frightened of the consequences of disclosure, loyal to an abusive carer, ashamed, or dissociated. Trauma responses including avoidance, re-experiencing, and emotional dysregulation can all affect the child's account and the way the interview should be structured. A trauma-informed interviewer does not treat the child as a neutral information source to be questioned; they understand that the child's emotional state is relevant to how information will be retrieved and communicated. This does not mean believing the child's account uncritically; it means understanding that non-disclosure or partial disclosure is not equivalent to absence of the alleged event.

The NICHD Protocol was developed by Michael Lamb (originally at NICHD, later at Cambridge University), Kathleen Coulborn Faller, Yael Orbach, Irit Hershkowitz, and other collaborators through a series of studies in the 1990s and 2000s. The Protocol is structured around the principle that open-ended questions yield more accurate and more complete information from children than closed or leading questions, and that the ratio of open to closed questioning in a well-conducted interview should be the inverse of what untrained interviewers typically produce.

The pre-substantive phase: rapport building and narrative practice. The Protocol opens with an extended rapport-building phase that includes an exercise Lamb called "episodic memory practice" or "narrative practice": asking the child to describe a recent neutral event (a birthday, a school trip, dinner last night) using open-ended prompts and explicitly training the child that "I don't know" and "I don't understand" are acceptable responses. This phase serves three functions. First, it calibrates the interview to the child's developmental level. Second, it trains the child in the elaborative narrative mode the substantive phase will require, rather than the yes/no response mode that most child conversations with adults reinforce. Third, it establishes the interviewer's willingness to accept "I don't know" answers, which reduces acquiescence to misleading questions.

The substantive phase: questioning hierarchy. The Protocol specifies a strict hierarchy of question types, from most to least open. The sequence is:

Open invitations first: "Tell me everything that happened, from beginning to end, as much as you can remember." Extended narrative prompts: "And then what happened?" "Tell me more about that." Directive questions (closed but non-leading, referencing the child's own words): "You mentioned [child's own word]. Tell me more about that." Option-posing questions (yes/no or forced choice): used sparingly and only to clarify. Leading questions: to be avoided; used only where absolutely necessary to establish basic facts that have not emerged from open questioning.

Field studies by Hershkowitz, Horowitz, and Lamb (2007) examining Israeli police interviews conducted with and without the Protocol found that Protocol interviews produced substantially more information from free recall, longer and more detailed narratives, and fewer responses to leading questions, compared to interviews conducted without the Protocol.

Field validation across jurisdictions. The Protocol has been field-validated in Israel, Sweden, Norway, the United Kingdom, Canada, and South Korea, among other countries. A large comparative study by Lamb et al. (2007) across North American, European, and Israeli samples found consistent superiority of Protocol-structured interviews over unstructured interviews in measures of narrative length, proportion of free-recall information, and consistency of account across interview phases.

The Memorandum of Good Practice on Video Recorded Interviews with Child Witnesses for Criminal Proceedings (Home Office, 1992) was a landmark document in cross-jurisdictional terms. It established the principle that child abuse investigations should involve structured video-recorded interviews conducted by trained officers, and that the resulting recording could serve as the child's evidence-in-chief at trial, sparing the child from live examination in the courtroom phase.

The shift from Memorandum to ABE. The Memorandum was substantially revised in 2002 as Achieving Best Evidence in Criminal Proceedings: Guidance on Interviewing Victims and Witnesses and Guidance on Using Special Measures (Home Office and others, first edition 2002; revised 2011 and 2022). ABE expanded the framework from child witnesses to a broader population of "vulnerable or intimidated witnesses," incorporating provisions from the Youth Justice and Criminal Evidence Act 1999 (YJCEA 1999). Under the YJCEA 1999, child witnesses (under 18) in sexual or violent offence cases automatically qualify for special measures, including the admission of pre-recorded video evidence as evidence-in-chief (YJCEA 1999 s 27) and live link questioning from outside the courtroom (s 24).

ABE interview structure. ABE interviews follow a four-phase structure: rapport, free narrative, questioning, and closure. The questioning phase requires adherence to a questioning hierarchy substantially similar to the NICHD Protocol, with open-ended prompts prioritised and leading questions explicitly restricted. ABE interviews are conducted by trained Achieving Best Evidence interviewers, typically police officers who have completed the National Policing Improvement Agency (now College of Policing) ABE interview training programme.

The Special Measures regime. Under YJCEA 1999, eligible witnesses (children, adults with mental or physical disabilities, and witnesses subject to fear or distress) may apply for special measures including: live link (testifying via video link from outside the courtroom); screens (to shield the witness from seeing the defendant); pre-recorded video evidence-in-chief; pre-recorded cross-examination; examination through an intermediary; and aids to communication. The pre-recorded cross-examination provision (YJCEA 1999 s 28), piloted extensively before national rollout, allows child witnesses to give their cross-examination evidence at a pre-trial hearing, removing the need for the child to appear at trial at all.

The intermediary role. One of ABE's most significant contributions is the formalised intermediary role, available under YJCEA 1999 s 29. An intermediary is a communication specialist, trained in child development, speech and language therapy, or related disciplines, who attends the interview and trial to ensure that questions are put to the child in a developmentally appropriate form and that the child's answers are accurately interpreted. The intermediary's presence converts the abstract legal question from "what did the witness say?" to "was the question the witness was asked the same question the questioner intended to ask?" Research by Plotnikoff and Woolfson (UK 2015) found that intermediary involvement substantially improved the quality and completeness of evidence from vulnerable witnesses.

The Protection of Children from Sexual Offences Act 2012 (POCSO) is the central Indian statute governing sexual offences against children under 18. Its procedural provisions for child-witness examination are among the most protective in the world at the statutory level, though their implementation varies substantially across states.

POCSO § 35: child witness examination procedure. Section 35 requires that the evidence of a child must be recorded within a period of thirty days of the Special Court taking cognizance, and that the Special Court must endeavour to complete the trial within one year. While § 35 does not specify the interview protocol in detail, the Statement of Objects and Reasons and subsequent Ministry guidelines emphasise the need for trained and child-friendly examination. The Rules under POCSO require that the examination of a child be conducted in the presence of a translator, interpreter, or special educator where necessary.

POCSO § 24: survivor-friendly examination. Section 24 restricts how a child witness may be examined. The provision prohibits the child from being called repeatedly to testify unless absolutely necessary. It requires that the examination take place in a child-friendly atmosphere, in the presence of the child's parents, guardians, or any other person the child trusts, and that the magistrate or judge satisfy themselves that the child understands the questions being put. Section 24(4) expressly prohibits the court from asking questions about the child's character or prior sexual experience in sexual-offence cases.

The gap between statute and implementation. The gap between POCSO's protective text and operational practice is well-documented by Indian child-rights organisations. A 2018 report by the HAQ: Centre for Child Rights found that dedicated child-friendly rooms for POCSO examination were absent in many Special Courts; that trained child-friendly police officers were not routinely assigned to POCSO cases; that delays far exceeding the thirty-day statutory target were common; and that the examination of children in open court rather than through intermediary or live-link arrangements was standard practice, exposing children to the confrontational adversarial environment that POCSO's provisions were designed to mitigate.

BNSS 2023 and child witness provisions. BNSS 2023 § 392 (successor to CrPC § 327) requires that inquiries and trials in cases involving rape and sexual offences not be open to the public, and that examination of witnesses in such cases must be conducted with appropriate safeguards. The intersection between BNSS and POCSO creates overlapping procedural frameworks; in practice, POCSO's specialised provisions govern in child sexual offence cases.

Comparison with ABE. The most significant practical difference between POCSO implementation and the ABE framework in England and Wales is the absence of a national trained-interviewer programme and the absence of a technology infrastructure for pre-recorded video evidence. ABE interviews are conducted by officers specifically trained for that role, using standardised training materials, with the video recording serving as the evidence-in-chief at trial. POCSO does not have an equivalent statutory and training infrastructure, and the in-court examination of child witnesses remains the norm.

The Cognitive Interview (CI) was developed by Ronald Fisher (Florida International University) and Edward Geiselman (UCLA) in 1984 as a structured interview technique for adult witnesses, drawing on encoding specificity principles from memory research. The original CI comprised four components: mental context reinstatement (mentally revisiting the environmental context of the witnessed event); reporting everything (reporting all details, even seemingly trivial ones); recalling events in multiple orders (including reverse chronological order); and changing perspectives (imagining the event from different viewpoints).

Efficacy in adult witnesses. Meta-analyses consistently show the CI produces 25-45 percent more correct information from adult witnesses than standard police questioning, with minimal increase in incorrect information (Kohnken, Milne, Memon, and Bull, 1999 meta-analysis of 42 studies). The CI has become standard practice in UK police interviewing of adult witnesses, embedded within the ABE framework, and is recommended in the NIJ guidelines for adult witness interviews.

The enhanced CI. Fisher and Geiselman subsequently developed the Enhanced Cognitive Interview (ECI), adding focused concentration instructions, active witness participation encouragement, and rapport-building components. The ECI is the version in current use in most English and Welsh police forces.

Adaptations for children. The original CI performs inconsistently with children under approximately ten years, primarily because mental context reinstatement and perspective change are cognitively demanding tasks that younger children cannot execute effectively. Adapted versions developed by Memon, Cronin, Eaves, and Bull (1996) and by Brown and Pipe (2003) simplify the CI for use with child witnesses by using visual context reinstatement (presenting photographs of the environment), removing the perspective-change instruction, and building in more extensive rapport and ground-rules phases. These adaptations improve CI performance with children compared to standard interviews, though the effect sizes are smaller than in adult samples.

Adaptations for adults with intellectual disabilities. A substantial minority of sexual and violent-crime complainants have intellectual disabilities, and this population presents specific challenges for investigative interviewing. A person with an intellectual disability may have difficulty with the open-ended narratives that NICHD and CI protocols favour, may show elevated acquiescence (the tendency to agree with questions regardless of content), and may have impaired source monitoring. Adaptations recommended by Milne and Bull (1999) and by Home Office guidance include using shorter, simpler sentences; breaking questions into single components; explicitly training the interviewee to say "I don't understand" and "I don't know"; and avoiding the reverse-order recall instruction.

ABE 2022 includes specific guidance for interviewing adults with communication difficulties, intellectual disabilities, and mental health conditions, incorporating the intermediary provision (YJCEA 1999 s 29) and specific questioning adaptations. In the US, there is no equivalent national framework for intellectually disabled adult witnesses, though the ADA accommodations and state-level protocols in several jurisdictions address some aspects.

The recovered-memory controversy of the 1980s and 1990s intersects directly with investigative interviewing because the protocols it generated, both those that produced false allegations and those later developed in response, shaped the entire field.

The McMartin preschool case. The McMartin Preschool case (Manhattan Beach, California, 1984-1990) began when a parent alleged that her son had been sexually abused by a teacher. The investigation expanded to include multiple teachers and hundreds of alleged child victims. The interviews, conducted primarily by the Children's Institute International, included repeated sessions with children, use of anatomically detailed dolls, focused questioning on specific accusations, and what critics characterised as a systematic disregard for children's denials and expressions of uncertainty. The resulting accounts included allegations of satanic ritual abuse, animals killed, and tunnel networks beneath the school, none of which were corroborated by physical evidence. After seven years of legal proceedings, all charges were eventually dropped or resulted in acquittals.

The Country Walk case. The Country Walk case (Miami, Florida, 1984) involved allegations against Frank and Ileana Fuster at a day-care centre. The interviews of child witnesses were conducted by a clinician who used intensive, leading questioning techniques. Ileana Fuster's testimony, obtained after prolonged imprisonment and intense pressure, has been the subject of sustained controversy about its reliability. The case has been analysed by researchers including Debbie Nathan and Michael Snedeker in Satan's Silence (1995) as an example of investigator-generated false allegations in child abuse cases.

Elizabeth Loftus and the suggestibility research response. The McMartin and similar cases directly stimulated the research programme described by Stephen Ceci and Maggie Bruck in their 1993 article "Suggestibility of the Child Witness" in Psychological Bulletin and the subsequent book Jeopardy in the Courtroom (1995). Ceci and Bruck's experimental programme demonstrated that children could be induced to report experiencing events that had not occurred, when those events were suggested repeatedly by authority figures over multiple sessions. The Loftus lost-in-the-mall paradigm extended this finding to older children and adults.

The recovered memory literature and its empirical status. The theoretical mechanism claimed by recovered-memory therapists, that childhood trauma is repressed and can be recovered through therapeutic techniques including hypnosis, guided imagery, and dream analysis, has not been supported by prospective research. Studies following known documented abuse survivors prospectively (Williams, 1994; Widom and Morris, 1997) found that some survivors did not recall the abuse at a later interview, but this finding is better explained by normal forgetting, motivated non-disclosure, and ordinary memory failure than by a specific repression mechanism. The British Psychological Society Division of Clinical Psychology (1995) and the APA Working Group on Investigation of Memories of Childhood Abuse (1998) both issued statements acknowledging that genuine trauma can impair memory recall while declining to endorse the repression mechanism as scientifically validated. Both statements also acknowledged the risk of false memories generated by suggestive therapeutic techniques.

Current practice implications. The legacy of the recovered-memory debate for contemporary investigative interviewing is primarily methodological. The NICHD Protocol, ABE, and equivalent frameworks are explicitly designed to minimise the use of leading and suggestive questions, to avoid multiple-session questioning about the same events, and to not use techniques (anatomical dolls in leading contexts, for example) that have been empirically associated with false-memory production. Investigative interviewers are now trained to document their questioning technique and to provide that documentation to the court so that the trier of fact can evaluate the conditions under which the account was obtained.