Archived publications

You are currently in the publications archive, and may not be accessing the most up to date information. The more recent publications can be found in the publications A - Z.



Comments on the investigation and interviewing of 

children in the Ellis case

Graham M. Davies
Department of Psychology
Leicester University
United Kingdom

1. I, Graham Michael Davies, am a Professor of Psychology at Leicester University. I am a Fellow of the British Psychological Society and a Chartered Forensic Psychologist. My principal research and professional interests lie in the area of children's testimony. In this connection, I served on the original Pigot Working Party which developed the `Memorandum of Good Practice on Video Recorded Interviews with Child Witnesses' (Home Office, 1992) and lead the writing team which has drafted the new `Guidance' which will supersede the Memorandum next year. As part of the Inquiry into aspects of the Ellis case, I have been asked to consider whether `there are features of the investigation and/or interviews of the children which may have affected the reliability of the children's evidence, and if so, their likely impact' (letter of instruction from Sir Thomas Eichelbaum, 24.7.2000). In this connection, I have been supplied with (i) videotaped records of interviews with all children who were interviewed as part of the original investigation and who gave evidence at trial or depositions; and (ii) complete typed transcripts of these interviews (iii) transcripts of the examination and cross examination of the evidence given by interviewers, parents and children at depositions and trial. My report focuses upon the testimony provided by the 6 witnesses whose evidence led to the convictions against Mr Ellis.

2. My report is divided into three discrete sections as follows:

  • A summary of research findings on the reliability of evidence and vulnerability to suggestion as witnesses of children aged 5-6 years of age (sections 3-17)
  • Comments on the content and style of the interviews with each of the key witnesses, relative to current professional standards and research findings (sections 18-54)
  • More general comments on the investigation as a whole and in particular, issues relating to the reliability and suggestibility of the testimony of the witnesses (sections 55-60).

My report will focus on these procedural matters: it will not attempt to pass judgement on the guilt or otherwise of Mr Ellis nor to pronounce on the reliability of individual children's accounts. I perceive my role to be to provide independent advice and relevant information for others to draw their own conclusions, based on the wider evidence and circumstances of the case.

3. With the exception of one child, who was 7 years old, the remaining "conviction children" were all aged 5 - 6 years old at the time of their interviews. The research literature on the memory capacities of 5 and 6 year old children suggests the following generalities:

  • Children of this age are capable of observing events and recalling them accurately at a later time
  • Children's spontaneous statements will be limited and brief compared to those of older peers, but generally accurate
  • Statements concerning events which are repeated over time are likely to be more accurate than one-off events: the gist of such events is likely to be well recalled, including their temporal sequence
  • Children will show better recall for actions, than descriptions of persons; central features of events will be better recalled rather than peripheral features
  • Recall will be better for actions involving the child directly rather than as a bystander
  • Repeated events which take place in familiar contexts are likely to be better recalled than novel events in unfamiliar environments
  • Information about when events occurred (day, month, year) is likely to be very poorly retained unless linked to a milestone of significance to the child (e.g. a birthday or religious festival)
  • Likewise, the frequency with which events occur is likely to be poorly estimated: most children of this age operate on a `one; two; many' classification system
  • Elapsed time is generally poorly represented: children of this age have difficulty in estimating the duration of an event and how long ago it occurred.

All statements regarding accuracy assume that interviews with the child take place under conditions which minimise obvious sources of suggestion. In recent years, a great deal of psychological research has been devoted to identifying potential sources of suggestive responding in children in the age range 56 years and calibrating their effects.

4. It has long been demonstrated that, under experimental conditions, the reports of children aged 5-6 years may be disproportionately vulnerable to leading or suggestive questioning compared to the testimonies of older children and adults. A more recent finding, from the work of Ceci and his colleagues (e.g. Ceci & Bruck, 1996), is that the impact of such questioning may spill over into a child's subsequent spontaneous statements, such that the child may claim that a non-existent event occurred and even include spurious supportive detail to back up this claim. There is usually no question of the child's intending to mislead; the statements are made with evident sincerity, but appear to be based on information derived from sources other than direct experience. Some of these sources are listed below.

5. One source of suggestive responding is negative stereotyping: consistent denigration by an adult of another individual in the presence of a child may result in the child reinterpreting and reporting their interactions with that individual in a way which is consistent with the negative remarks.

Leichtman and Ceci (1995) arranged for a stranger, `Sam Stone', to meet briefly with children aged 3-4 and 5-6 years of age at a nursery school. Subsequently, when the children were interviewed concerning whether he had damaged a book or a toy, all age groups rejected this possibility with 93% accuracy. However, if the visit was preceded by stories about the stranger designed to foster negative stereotyping (` Sam is a clumsy person') 18�/a of the 5-6 year olds claimed he had damaged something and when pressed, 10% said they had seen him do it and one child stick to their story when gently challenged as to its accuracy.

6. Another source of suggestible responding in children of this age results from post-event misinformation: children recalling events from the past have problems distinguishing what they have seen and heard originally from what they have been told subsequently by others.

Poole and Lindsay (1995) staged science demonstrations for 3-5 year olds at nursery school and subsequently had their parents read them distorted accounts of what they had seen. Many of the children went on to incorporate the new and conflicting information into their answers when questioned about the original events.

7. These effects are more likely when the person relating the misinformation is a reliable adult rather than another child, suggesting that the credibility of the source may contribute to the effect.

Lampinen and Smith (1995) read a story to 3-5 year-olds and then showed them a video of a person talking about the story in a misleading way. The credibility of the person in the video was varied: in one condition the person was an adult, in the second a child, and in the final condition the person was an adult who was introduced as being 'silly'. Participants only incorporated the misleading information into their reports when the incorrect information was presented by the 'credible' adult; participants were not misled by the 'silly' adult or the child.

8. Effects of misleading information do not need to be derived from hearing detailed accounts in the interim: distortions can result from the persistent use of leading questions by the interviewer which result in the misinformation being subsequently incorporated into the child's own version of events.

In the Leichtman and Ceci (1995) study referred to above, some of the children aged 3-4. and 5-6 years who had experienced negative stereotyping went on to be interviewed four times by interviewers who asked leading questions implying that `Sam Stone' had damaged items at the nursery. Subsequently, the proportion of 5-6 year olds who claimed he had damaged nursery equipment was 38%, decreased to 15% when pressed, and 5% persisted their story when challenged.

9. It is not necessary to repeatedly use leading questions to create memory distortion. In some circumstances, simply the repetition of the same specific question across a number of interviews can cause a child to create a false memory which they may hold with apparent confidence.

Ceci, Huffman, Smith and Loftus (1994) questioned children aged 3-6 years of age on 7 separate occasions about a mix of real events from their lives and false events that they established the child had never experienced ( e.g. getting their finger caught in a mousetrap). Over the course of questioning, 32% of the older children claimed to have experienced one or more of the fictitious incidents, sometimes corroborating these admissions with spurious detail. Such effects were increased if the child was encouraged to visualise the experiences as they tried to remember them.

Questioning needs to be direct and specific to obtain the effect: asking the same question but in an open format (e.g. wh- questions: those that begin `Why, where, when' etc. ) which require the child to construct a response rather than answering `yes' or `no' do not produce the same distorting effects (Waterman, 2000).

10. Like all of the factors described above, such effects are not unique to children; similar, but smaller effects have been observed in studies with adult participants. However, the younger the child, the more vulnerable they are to such contamination. Moreover, there are additional modulating factors which can increase or decrease the power of the effect. For instance, the effects of misinformation are increased if along delay separates original observation and subsequent misleading information or if such misinformation concerns peripheral detail rather than central events. Conversely, it is more difficult to mislead a child witness about events they have themselves experienced or have undergone in a similar form on a number of occasions. Preschool children who are more socially assertive and more cognitively advanced are less susceptible to misinformation effects than other children of comparable age (Geddie, Fradin & Beer, 2000).

11. It is important to stress that factors influencing the reliability of the statements of very young children are not based purely on the results of laboratory experiments. In recent years there have been a number of instances where allegations of sexual and/or physical abuse have emerged from interviews with preschool children in day care which have subsequently proven to be unreliable.

In the Kelly Michaels case (State vs Michaels, 642, A 2d 1372,1385-91; New Jersey, USA, 1994) an investigation was sparked into possible abuse in a day-care centre for 3-6 year olds by a suggestive remark made by one child who was having his temperature taken rectally. This led to a school meeting and individual parents questioning their children which was followed by repeated formal interviews with police child protection officials. A total of 19 children gave evidence at trial against Michaels, alleging a range of bizarre sexual practices which had allegedly occurred during school hours. Despite the absence of any supporting medical evidence, Michaels was found guilty. However, the verdict was quashed on appeal, largely on the basis of the highly suggestive interviewing procedures employed and the lack of any corroborative evidence for the various allegations made.

12. Similar apparent miscarriages of justice have occurred elsewhere in other cases of alleged sexual abuse by adult carers of young children, notably the McMartin case in California (see Davies, 1991) and the Orkney Inquiry (Clyde, 1992) Bruck and Ceci (1995) list a number of features which emerged from their review of State of New Jersey vs Michaels which they feel may contribute to possible miscarriages of justice in such cases. In addition to those mentioned above, they instance:

  • Interviewer bias: interviewers having a pre-set agenda which they pursue single-mindedly, irrespective of the initial responses of the child. Eventually, some children comply with the interviewer's agenda
  • Differential reinforcement: any statement the child makes which alludes to an abuse agenda is reinforced by the interviewer giving verbal and non- verbal encouragement to the child.
  • Peer pressure: the interviewer tells children that other children have already provided statements mentioning this child's involvement in various acts: the interviewer looks to the child to confirm their role
  • Inappropriate use of anatomical dolls: the ability to see the doll as a model of self develops much later than play with dolls, typically around 4-S years of age. Demonstrations of actions on and with dolls by developmentally younger children may reflect suggestion rather than accurately mimic real events

13. However, for every case of interviewers eliciting erroneous information from young children, there will be instances where the sometimes bizarre allegations by children turn out to have a basis in solid fact. In the United Kingdom, Fred and Rosemary West committed a series of sexually motivated murders of men, women and children in a house in Gloucester over a number of years. The offences only came to light when a police officer took seriously the statements of a child familiar with the house. In Belgium in 1996-7, Marc Detroux murdered a number of children, including some who were starved to death; statements by children regarding Detroux were initially dismissed by the police, allowing him to commit further crimes. In Australia, in the `Mr Bubbles' case, young children's allegations against the accused were dismissed as bizarre and implausible at trial, only for a television company to discover `similar fact' allegations concerning the accused from an earlier period in his life in Papua-New Guinea.

14. What factors might lead an investigator to give weight to the allegations of sexual abuse against adults by young children? Clearly, some factors will be those which apply to all forensic investigations. These will include:

  • medical evidence: does the child have a sexually transmitted disease or does examination suggest that sexual penetration has occurred?
  • forensic evidence: does the alleged crime scene contain forensic evidence consistent with the child's allegations?
  • privileged knowledge: does the child show knowledge of the suspect's house, property or personal appearance which would be unlikely to be acquired, other than by first-hand knowledge?
  • corroboration: do statements about tactics and style of abuse from one child coincide with those provided by a second child, where the possibilities of inter-witness contamination of evidence are minimal?

15. In addition, there are a number of more contentious factors which hinge on the very naivete of young children. Children of this age do tell lies, but the lies are not elaborate: usually a simple assertion or denial (`she did it'). Coaching children to tell more elaborate stories is possible, but existing research suggests the deception is normally readily detectable under questioning. Unlike the sorts of experiences children are invited to elaborate upon in experiments, most children aged 4-6 years have a very poor understanding of sexual matters in general and of the mechanics of sexual abuse in particular (Volbert and van der Zanden, 1996). Precocious knowledge, especially when expressed in terms appropriate to the child's current level of understanding (orgasm described as a sneeze or cry of pain; semen as cream or `wee' etc.), is seen as particularly diagnostic. `Sexual details misunderstood' forms one of 18 criteria which the German Courts use as part of their system for assessing the credibility of children's evidence in allegations of sexual abuse (Criterion Based Credibility Analysis or CBCA). The system is based on the idea that there are certain features of young children's accounts which would not readily occur to children who were deliberately setting out to fabricate an account. While the reliability and validity of CBCA is a matter of continuing debate among psychologists, many professionals find the criteria useful in assessing children's accounts (see Davies, in press for a review and Appendix 1 for a brief description of the criteria).

16. In sum, the results of recent research suggest that children of 5-6 years of age are capable of providing accurate accounts of events to which they have been a party. However, particular care needs to be exercised in the questioning of such children. They do appear to be more vulnerable to repeated questioning and suggestion and may come to own beliefs, which though seriously held, have no basis in fact. These problems are likely to be increased when considerable delay occurs between the time the events are alleged to have occurred and questioning, especially when the individual(s) concerned are negatively stereotyped or subject to suggestive comment in the interim. In evaluating a child's statement it is important therefore, to scrutinise very carefully the history of how it came to be made: did the statement emerge spontaneously or only as a result of direct and repeated questioning by parents? Was the name of the suspect suggested to the children as a possible wrong-doer or did the children's statements independently and consistently identify the accused and his actions from the earliest stage?

17. Unlike many child care cases in the United States, in the current case the interviews with the children have been generally well documented with proper record keeping and full videotape records of all the interviews with the six child witnesses whose allegations were upheld by the jury. It is thus possible to examine these interviews with the lessons of previous cases in mind and also to measure them against accepted guidelines for the conduct of such interviews. These guidelines in turn have been developed to counter precisely the kinds of interviewing errors which Ceci and others have identified. At the time these interviews were conducted in 1992, there were no official guidelines in New Zealand for interviewers in cases of alleged child abuse (Sir Thomas Eichelbaum, personal communication, 12/00). However, training in good practice was available and it is possible to assess such interviews against current and contemporary norms. I now turn to the series of interviews with each of these six children.

18. O. This child was interviewed once by Linda Morgan on 12.05.92. when she was approximately 7 years 8 months old. The interview lasts just over an hour; O and Ms Morgan sit side by side facing the camera. As with all the tapes I reviewed, the technical quality and audibility is excellent.

19. O comes across as a socially assured and somewhat precocious child who copes well with the interview situation. There is an absence of explicit leading questions, though many statements about Peter are triggered by reference back to conversations with her parents. While she makes clear her personal dislike of Peter and the reasons for it, only one serious allegation emerges and is not provoked by a reference back: Peter repeatedly touched her in the crotch in the course of play while other carers were not present. She demonstrates this by gesturing to the area between her legs. Her manner is noticeably subdued in discussion of this allegation, which she illustrates with a cupping movement of her hand. She says that this happened "six .... teen times": a literal estimate to which little significance should be attached (see (3) above). She says he desisted when she told him not to. She also volunteers that he did the same thing to Katy and Thomas. Peter peeked at her while she was being assisted on the toilet by Danny. Asked if she would ask Peter to help her in toiletting, O replies "Oh no way, I would never choose him". She claims to have had a small cut in her vagina as a result of Peter's sharp nails, though her parents do not recall this. The term `cut' is often used quite loosely by children of this age and can mean no more than a small scratch which might heal rapidly in such a vascularised region. I am satisfied that the interview itself is well conducted by an experienced interviewer: the truth or otherwise of the allegation rests on the reliability of the testimony of O's parents regarding their prior conversations with O. I could find nothing in the transcript to suggest the hand of an adult in the child's charges.

20. S. S was interviewed three times by Sue Sidey, between 1.5.92 and 3.8.92; she would have been 6 years 9 months old by the final interview.

Interview 1 on 1.5.92. lasts approximately I hour 25 minutes with breaks. S is cooperative and attempts to answer all questions but is fidgety and restive for long periods. There are two short breaks toward the end of the interview, but an interview of nearly an hour and an half is asking a lot of the concentration of a child of 6 years, 6 months. S makes three principal allegations: that Peter painted her bottom; that Peter urinated over her face and that Peter took a bath with her at his house, during which he defecated in the water. The import of the painting incident is unclear: it happened after swimming; there were people about and someone took a photograph and these are "at home". As to whether she was undressed, the interviewer leads: "he painted your bottom without your clothes on, right?" in the absence of a spontaneous statement from S. The urinating incident is much more clearly described with some spontaneous detail and she demonstrates with model figures the position of herself and Peter. It happened to Z and Hazel at the same time, but S is inconsistent on the order in which Peter saw the three children. It happened to her `just once'. As regard the third incident, the interviewer has to prompt S by reference to earlier conversations about the bath and S is vague as to the location: she first says the creche and then Peter's house; earlier in the interview she has twice said nothing untoward happened at his house. She is reluctant to discuss the defecation, but when she does so, she provides some vivid detail ("the poos were floating around.."). S wants to finish the interview, but the interviewer presses her and says "it would be good if we could just do it before we leave today".

21. In my view, this interview was less successful. It went on far too long, with a hint of coercion toward the end. Given that the main outlines of the allegations were known, it could usefully have been broken up into two shorter interviews. There are also far too many focused questions asking for detail on the key incidents; the incidents described took place between 18 months and over two years previously. A 6 year old is unlikely to remember the colour of Peter's pants, for instance, but if she is asked a direct question, with no explicit reminder that `don't know' is an acceptable reply, she will provide an answer (Waterman, 2000). There is also a great deal of leading around the bath incident, which may reflect the child's fatigue at this stage of the interview.

22. The second interview took place nearly a month later (28.5.92.) and this time lasts for around 50 minutes. After the necessary preliminaries, the focus returns to the bath incident and various statements made by S in the interim to her parents. S's manner is more relaxed on this occasion though the same restlessness and wandering attention returns after about 25 minutes. S asks for her mother and the interviewer replies: " Maybe we could see Mum after we have finished talking about it, okay?" An early focus concerns what was washed in the bath and the interviewer produces conventional dolls in an effort to elicit the information; some facilitation is necessary for the child to repeat the allegation The interviewer then moves on to Peter's behaviour with his pet dog which S has mentioned to her mother. Again, some prompting is required for S to repeat the allegation that Peter had put his finger up its bottom, which provides a starting point for exploring whether Peter might have digitally penetrated S. This allegation is eventually confirmed but not before some fairly heavy prompting from the interviewer and an understandable reluctance to confirm from S. Later, the interviewer asks S if it really happened; S replies "that really happened". There are increasing signs of loss of concentration, as the interviewer presses S over whether threats were made about disclosure: S initially says no but later claims Peter did warn her off "else I'll smack your bottom"-hardly an extreme sanction.

23. This is another interview which perhaps went on too long and contained too many focused questions for the age of the memories: the "don't know" and "cant remember" options were never explicitly stated to the child (this would now be standard practice). Little fresh evidence appears to have been elicited, beyond that already disclosed and discussed with her parents in the interim. S and her family were also friends with the Z's, who were aware of the allegations concerning the visit to Peter's house and the creche toilet. As in the first interview, S brought along a book of drawings as an aid to recall; questions inevitably arise about the circumstances under which the originals were produced, though there is now good research evidence that drawing while talking can act as a facilitator to recall by young children (Gross & Hayne, 1999). I understand that this tape was not played at court because of concerns over the use of leading questions about digital penetration. This in retrospect can be seen to have been an important issue, given the subsequent medical examination which supported a history of such penetration.

24. The third interview with S took place over three months later on 3.8.92. and lasts for just over an hour. The coffee table has been cleared away and Sue and S are sitting on the floor with toys. S has come to talk some more about the bath incident. She now says that Peter put his penis into her vagina " a little"; she describes this as " a sexing". Her manner is composed and her tone conversational. There are repeated attempts to reenact the bath incident with miniature figures: they are so small it is difficult to know what is being shown. S now says she drove to Peter's house in a car which was green. This contradicts her statement in interview 1 that she walked there with Peter. This is yet another example of the interviewer asking a series of direct questions with the expectation that S can provide specific answers; when the interviewer asks " Are you sure about the colour or not?" S hedges " I think it was green". There is nothing in this exchange or the later one concerning the `needles in the bottom' allegation which hints at additional superfluous detail or elaboration of the kind looked for in CBCA analyses. The reference to the needles is heavily facilitated by the interviewer, when S does not remember and the prompt is strongly leading ("What else could have touched it and made it bleed?"). There has been further contact and discussion with Z since the last interview ("Z said it happened as well") and Z is named as having bathed with Peter at the house. Once again the interview is too long for the age and attention span of the child, who asks for her mother after half an hour and begins to roam the room after 45 minutes.

25. It is extremely difficult to interview young children in ways which are non-leading, but nevertheless elicit information on which a jury can reach an informed decision. At times the interviewer shows a proper regard for suggestion (the range of body parts offered to diagnose touching in the bath; the checks as to confidence and certainty). In my view, however, this series of interviews, given the age of the child, are too long and contain too many specific questions, which produce answers of questionable reliability. The interviewer is tempted, at several key points, to ask questions which are leading. This strategy could have been justified if it produced new convincing detail or additional information which could be corroborated, but all too often this is not the case. In my opinion, the allegation of greatest evidential weight is that from the first interview concerning the toilet cubicle incident and the least is that concerning the needles. With regard to the bath, the version in tape 1 seems the least prompted: the later versions follow conversations with Z and others and the possibility of negative stereotyping from exposure to Ellis's case on TV (see interview 3), notwithstanding the medical evidence.

26. Q. There are three interviews with Q spread over 9 months, all conducted by Sue Sidey, who also conducted interviews with her brother P. The first interview with Q took place on 09.03.92. and lasted for approximately 40 minutes. Q is cooperative and forthcoming throughout. She has come to talk about Peter, who is "a very mean man who wants children to feel all scared". She complains of undue tickling (a complaint mentioned by other children), which occurred first when she was "three or four" (she is now 6 years old); she repeats these dates later in the interview concerning the other allegations. She states unequivocally, with no prompting, that Peter had put his penis in her mouth in the toilets ("he calls (it) drinking games"): "he just holding us on the shoulders to make sure we are still and that". She was "all scared and wanted to tell someone"; told `Marie' and later mentions telling Jan and Gaye (all creche helpers, whose statements can be checked). Peter's penis "stayed still" in her mouth and she accepts the interviewer's suggestion that it was "floppy and hanging down". Later she says Peter did the same thing to P (her brother), Z and H. She talks of Peter "touching our private parts" and understands and uses the terms anus, clitoris and vagina appropriately. There is an allegation that Peter touched her anus and vagina with his hands over her clothes and directly ("secret touching") while she was using the toilet. This happened "lots of times". She demonstrates rubbing on a doll. The interviewer establishes that this was not during routine toileting. The interviewer elicits an allegation that Peter threatened she would turn into a gherkin if she told anyone; she told her parents, she says, when there was no sign that this had happened by age 5 ("I realised he was not telling the truth"). Peter also took them for walks with his dog, but there were no untoward incidents.

27. This is an interview of appropriate length, that avoids obvious leading questions and over-dependence on specific questions; Q is forthcoming and clear in her allegations, which cannot be traced to any obvious prompting or leading by the interviewer. The allegations are quite explicit. The main question which has to be addressed. in my view. is whether these are real experiences she has undergone, or whether they could have arisen from discussions with her brother, the Z children or her own parents. Careful examination of the evidence and time-scale will be required to answer this question. Certainly, she is frank about having discussed the matter with her parents, as this precipitated the current interview. However, apart from the relatively advanced vocabulary for sexual parts and a reference to `our' private parts, I could see nothing in the interview that was redolent of coaching. Further, the interview contains some explicit detail, which in my view, a child of this age would find difficult to invent.

28. The second interview with Q takes place nearly 7 months later on 6.10.92: it lasts just over 30 minutes. A more formal interview setting- sitting facing the camera- and Q more thoughtful and softly spoken. The motive and purpose of this second interview is not immediately evident. The interviewer establishes that `just tickling' but no `secret touching' took place outside of school (ambiguous in interview 1). Q claims she told Gaye and that Gaye had peeped while Peter put his penis in a child's mouth (it is not clear which child). Q says "... when Peter got out of the toilets, she (Gaye) said don't do that and Peter said `oh yes, I will' and then he went back in the toilets with the children and started doing it again". Q also denies explicitly being driven anywhere by Peter. She repeats her allegation that Peter put his penis in her mouth. She comes up with a fresh allegation that Peter went with them to "a big room with lots of escalators" where secret touching took place. It is evident that she has discussed this with her mother who has suggested that this must be the City Council Building ("My Mummy thinks that it could have been the City Council or something"). There is a further discussion of the nature of the tickling, which does not elicit any inappropriate touching, nor is there any demonstrated on a doll despite repeated invitations. There is a brief return to who was present when Peter put his penis in Q's mouth. She now names three boys and the two girls, Z and T (H was named previously, but no boys).

29. This again appears to be an interview with no very obvious flaws. The interviewer explicitly cautions Q to tell her only what she knows. The earlier allegation is repeated and a new allegation made concerning secret touching in the big room. However, it is not clear how much this has been shaped and developed by conversations with mother: Certainly, Q seems to have wholly taken on board the idea of it being the City Council Building and rather bizarrely in this context, refers to "little wee kids" working there for the City Council. Given the elapsed time since her period at the creche, a variety of explanations could be offered for this odd remark. It could arise from contamination due to discussions with parents or other children. Alternatively, it could be a source monitoring error: pressure to recall has led to the conflation of two separate memories, one of a visit to a large building and the other of Peter's alleged behaviour at creche. Q is less clear in the second interview about who was present at the times Peter allegedly put his penis in her mouth: "children" when Gaye is supposed to have peeped; toward the end of the interview, a different selection of children is named for the occasion covered in interview 1 (this confusion may be the product of too many specific questions to a tiring child).

30. The third interview with Q occurs after a further three months on 9.12.92 and lasts for just over 54 minutes. Q and the interviewer sit side- by-side and carry out various craft and art activities as they talk: this arrangement seems to work well and Q is absorbed for most of along session (Interviewer: "Q I know you are getting tired" after 35 minutes). Q has told her Mother more things and Q begins by describing walks with Peter to the park and Art Centre. When asked whether she liked the walks, Q twice says "they were good". The interviewer turns the conversation toward "mean things Peter has done". There are more references to the room with the desks: Peter, Jeremy, R and P were present but no other "big people": she went there "just once" and had to run around the room. The interviewer raises Peter putting his penis in her mouth which Q says happened in the big room and the interviewer moves to correct her ("it happened at the toilets in the creche, right?") but later Q says it happened at both sites. The interviewer attempts to raise the issue of secret touching involving a group of men, which evidently she has heard about via her parents, but Q thinks she is still talking about the `big room'. After some facilitation, an incident involving sexualized touching by three men is elicited in response to a series of direct questions. A "man with black hair" touched her vulva with his penis; the other men did so too; she had been too scared to tell earlier. Other children were present, including L and later, P. In response to a series of focused questions, Q draws and describes "Peter's friends" but the descriptions are skeletal, normative and lacking in any distinctive detail.

31. In my view, interviews 2 and 3 are of limited evidential value. Q's account in interview 3 of the meeting with the three men seems implausible: lacking in detail or incident. There is nothing in the descriptions of the incidents with the three men which has not already been mentioned in the earlier interview. In the absence of external validation or mutual corroboration, I found these particular accusations unconvincing. In interview 3, the interviewer emphasises to Q the importance of correcting her if she misunderstands anything Q tells her and the warning about telling the truth is given at the beginning and again, significantly perhaps, toward the end of the interview. The interviewer picks Q up when she alleges the men were standing and she sitting when the penis was placed in her mouth and Q then says they were squatting down. Interview 3 is also somewhat longer than the earlier two tapes and Q appears tired and restless as the interview wears on. If the incident with the three men did not have a basis in reality, it is unclear where it came from: Q and her family have been in touch with the Z's in the interim and is presumably aware of the ongoing legal proceedings against Mr Ellis. It is noteworthy that the Prosecution chose to show only interview 1 at trial and not to use interviews 2 or 3.

32. X.' X was interviewed five times over a 5 month period, in the first interview by Linda Morgan and subsequently by Sue Sidey. He was just over 6 years old at the start and 6 years 7 months old at the time of the final interview. The first interview lasts for around 45 minutes. The early part provides a developmental check: X seems a bright and alert little boy. Truth, Lies and promises is followed by some dilatory discussion of his early schooling in which Peter figures. Initially he says "He was alright when I did go there... and now he's not". She raises what he has told his mum and dad and he says "he fiddled with my rude parts". Appropriately, she then goes on to establish what for X is, and is not, a `rude part'. She then goes through a series of direct questions which elicits that Peter `wobbled his dick ` in the toilets. X says it happened when he was three, when he "done poos in his pants" and had to get changed. Peter assisted X in this. Later the interviewer elicits from X that he was lying on a board at the time, later still, the interviewer establishes through models that the board was outside the toilet and X says "I was just a baby then", but no further detail of the incident is forthcoming. He repeats an allegation made by his parents, against John, but seems to have no personal experience of this ("`cos my Mum and Dad told me"). The interviewer establishes that X has also told Robert, his brother and Pauline Selby.

33. It is not clear when this single incident involving Peter took place. X says he was three, but the mention of a board suggests nappy changing, but he talks about `a poo in his pants'. Even if aged three, X is talking about an incident from the dawn of his memory, so that there is unlikely to be significant detail retained and X resists the interviewer's invitations to elaborate on his initial allegation. It is also evident that the story has been told several times to parents and others aware of the main thrust of the allegations against Mr Ellis, so that there is some risk of elaboration on what could have been a routine aspect of toiletting, carried out clumsily. I was generally impressed with the quality of this interview which included necessary checks on X's comprehension and was not judgmental or accusatory (for instance, the interviewer asked X what aspects of Peter he really liked).

34. Interview 2 lasts just over an 'hour and takes place nearly three months later, the first of a group of three carried out in as many days, by Sue Sidey. After the usual developmental checks and truth aced lies, X says he has come to tell some more about what Peter did to him. He swore at him; "he made me go in the bath with him". "It was upstairs in some room" in Peter's house. Peter's friends were present, one beginning with R: "something Like Robert". There was also an old man with a beard ("I think so anyhow"). There were other children there. He got in the bath with Peter; he made him eat poos ("he ah-made me eat his pons-ah, no, no, he made me eat my poos and then he said next time you come here I'll eat mine and he didn't"). This happened "about twice". "He made me touch his penis a long time" and white stuff came out. He shows how he dipped his penis but says his hands stayed still. The white stuff went in the toilet and the floor. It happened at the creche and at the house; no one else was about. His penis was hanging down; when the interviewer queries whether the penis was sticking out he says "yeah it was".. He was "too scared" to tell anyone else. In response to the question "did any other part of you touch Peter's penis or not?", X says "No, he put his penis up my bum". The interviewer fetches a miniature bath and some figures and invites X to demonstrate how this occurred. He says Peter put his penis in his bum, but that Peter did not move. Afterward, they went back to the nursery and a teacher asked if they had a nice walk: Peter said "oh yeah, we had a good walk".

35. X goes on to allege that they drove to Peters house; Peter drove and "one of his friends was in the back"-the old one. When the interviewer suggests this was Robert, X corrects her: the old one was not Robert. There were two or three men waiting at the house. There were other children, but he can only remember Jimmy (later in the interview he mentions Z and Y). Peter put his penis up his bum and "white sticky stuff' came out. X initially says no one else did anything to him but later he says Robert "done hurt me the same way as Peter" and demonstrates on a conventional doll, pointing to its bottom. Robert did this to X, not in "Peter's house, it was one of his friends". Robert had not undressed but just taken down his pants and Robert had done the same. It took place on the bed or on the floor; the curtains were closed and the light was on. His penis did not go in the "poo hole" but just in "the crack" between his buttocks. Initially he responds negatively to the question as to whether anyone else had done that to his bum but then almost immediately mentions "three other guys and one of them was old". X needs to go to the toilet. When he comes back he says Peter watched and took photographs of "all the nasty things". In response to the question: "Where else could a penis go?", X says "He put it in my mouth as well"; it tasted yucky. Later he says it made him sick "and I spewed", but it is unclear whether this is a result of the alleged incident with Peter. The second time he did it, X says he pushed his penis away.

36. Interview 2 described above is long, but X shows no overt signs of distress: he talks as he plays, first at the table and later on the floor. There are few obvious leading questions but there are a lot of direct, focused questions, but on this occasion, a good deal of relevant detail is elicited (the interior of `Peter's friends' house; the appearance of the three men). The allegations are generally plausible, given the known mechanics of abuse, and the circumstances seem reasonable, with the exception of the claim that Peter himself drove to his house. X has clearly discussed his claims with various members of his family, but there is little indication of a rehearsed account: production is consistent, plausible but unstructured (all positive CBCA criteria). The only hint I could detect of possible parental influence was his frequent references to his penis, whereas 'dick' had been his preferred term in interview 1. I could see no indication that the style of this interview was oppressive or accusatory: X was very much in control.

37. Interview 3 took place the next day and lasts, again, just over an hour. X and the interviewer are sitting either side of a large floor cushion. After preliminaries , X has some more things to tell about "What Peter did to me". He relates how Peter drove him to a library which had a trap door. Peter's mother was there. When asked whether it was close to shops, X says it was, then says he is not sure. He alleges Peter "tried to make me go up and hurt children". His friends were there: Timmy, Y, Z and Daniel. They went through a trap door, below the building was a maze, but they did not enter it. There were women present and two men called Spike and Boulderhead. Peter kicked and punched him and put a stick up his bum, causing it to bleed: "he stuck a burning piece of paper up my bum five times, so that means it (his bum) bleed ..uh..ten times". Spike, Boulderhead and 19 of their friends put their penis in X's bum. X and his friends were tied up. They were given yucky pills to take by Peter's mother, who tried to get them drunk on beer. By 12.30, X is getting visibly restless. Peter put a needle in his hand; he did the same to the other kids. The interviewer produces a female anatomical doll and X identifies its vagina. He has seen his mother's vagina and also the two women at the library ("They just pulled their pants down to show us for some stupid reason"). The women kicked and punched the other children. By 12.50 X is tiring (Int: "I know your very tired aren't you. X: "Mm. Its boring just sitting here with no one in it"). After a break for consultation, there is more discussion of whether the friends did anything to each other; X says he punched Y in the balls and kicked his bum.

38. The fourth interview is again over an hour long (1h., 10m.), the setting as for the previous interview. After truth and promises. X says he wants to talk about "the things Peter's friends did to me". The location is now specific: "The two storied house in Hereford Street". Marie, Gaye and Andrew were there, and Robert and Peter. The kids were in the middle of a circle and had to kick each other; Peter and his friends and family were playing guitars. "Marie, Gaye and Jan pretended to sex" . X demonstrates with dolls; he briefly juxtaposes the dolls but seems embarrassed and puts them down. They pretended to be cowboys. The interviewer asks whether he was describing a game at creche: "No, it wasn't a game-and the kids in the middle were naked". The children were encouraged to hit each under threat of being stabbed; photographs were taken. Andrew put needles up the children's penises. Minutes later he says Andrew did not put the needle in his penis but in his bum. Later he returns to this theme and talks of needles in "the kids' penises and vaginas". They put the children in an oven and Peter pretended to eat him. This was in the building with the trapdoor. Spike drove them there in a black car. Peter put his penis in his mouth in this building. The girls hurt each other too; they went back to the creche in an old grey car. The interviewer asks whether the kids touched the boy's penises "Cos I heard the kids had. .." he says he saw kids putting penises in each other's mouths. At one point the interviewer asks: "So what stopped you telling me about this yesterday". X: " I just remembered". He talks about Timmy and Y being involved. X is plainly restless and complains of feeling sick and drinks a glass of water. There is a brief discussion about Marie and Gaye "sexing" then shortly after this, the interview closes.

39. The final interview with X takes place over two months later. X and Sue Sidey sit side by side with X colouring and drawing as he speaks in an easy conversational tone. The interviewer establishes that he wants to talk about a trap at the creche. Peter would push kids down it: it led to Room 20: he has seen the room recently. This happened "when I was about four". Peter's mother would hang them up from hooks on the ceiling in cages. Various children had burning paper and sticks placed up their bums. At this point, the interviewer interjects " are these things that really happened or things that you think might have happened?" X says they really happened. There was another trap in the Womble area: a little ladder and tunnel. Room 20 was entered by a secret door. You went up a ladder through a secret door to get to Gaye's office: "most of the creche" knew about this secret way. Peter's mother kicked them: that was how he got this big bruise on his leg; when challenged he turns it into a joke. The interviewer asks "Is there any other stuff that you've been telling that's not true?" X says no. Robert took photos with a video camera. Other named kids had sharp sticks and burning paper on their bums and various creche staff were involved and he produces a list of names when requested. The interviewer asks "Where does the trap go?" and X says "It just goes in the ground and I've been watching the news lately and I've heard the police have found it".

40. These later videos show a gradual spiral into more elaborate allegations embracing a wider and wider circle of helpers and teachers at the creche. A child may disclose more information over a number of interviews as trust is established between interviewer and child and the effect of any threats or prohibitions made by the abusers) weaken. However, such elaboration may have other causes of the kind highlighted in the earlier sections: repeated requests to recall may lead to construction of nonfactual accounts, with the young child unable to distinguish between their retrieved memories, self-generated imagery and content derived from other, later sources. In X's case, it emerged at trial that all the video sessions were proceeded by extensive discussions with his parents, who admitted to asking him very direct, but not leading questions. This was contrary to the advice they were given at the first briefing session and that there was also contact with other parents and notes were compared. It is also accepted that X was taken to various addresses, perhaps in an effort to jog his memory. The addresses certainly figured in his subsequent interviews, but not in ways which increased the credibility of the accounts. He also seems to have followed developments on television news or at least heard or been questioned about them: his last interview seems to have been triggered by the news of police investigation of trap doors at the nursery. Some themes from his earlier videos reappear in the later ones: the grey car; the burning paper and sticks up bums; strangers who hurt children. The only new themes concern the `circle game', the presence of secret passageways and the hanging of children in cages from the ceiling. Clearly, the police investigation explored carefully the creche and other named properties with these allegations in mind and certain traps and air spaces were found at 404 Hereford Street and the creche. This might be construed as broadly consistent with some of X's allegations, but no forensic links could be established between these areas and Mr Ellis or the children involved. Other explanations need also to be considered. Secret passageways are the stuff of children's fiction (interestingly, they also figured in the testimony of children in the McMartin case, referred to earlier) and cages containing children might reasonably be traced to a similar origin. However, the circle game, with its quite detailed description, is an unusual theme, which I would be at a loss to explain in purely fictional terms.

41. R . R was interviewed three times by Sue Sidey over a six month period. R was 5 years 6 months at the start and just 6 years old at the end. Interview 1, which took place on 3.4.92., lasts just under an hour. It is evident that R has not been briefed by his parents as the focus of the interview, but he guesses the creche. They sit on the floor and as they speak, he and the interviewer build a schematic layout of the creche: the use of models to stimulate recall is an acceptable forensic practice with children of this age and can benefit recall, provided it is not linked to suggestive questioning (Davies & Westcott, 1999). It does, however, take some considerable time. R mentions that "mean things" happened at w-: creche, that "Peter said I'd die if you tell" which he now realises was a trick In response to a specific prompt ("what's the other mean things that you told" (your mum and dad)) he says that (Peter) "did wees in people's faces" and this happened "quite a lot". This happened when the child was sitting in the toilet (he demonstrates this with a model). It seems he was present when this occurred, but he can't remember or is reluctant to tell, who was involved ("because I've just forgotten"). He says he feels angry about Peter, but seems unwilling to talk about any personal involvement. Eventually the interviewer asks another direct question ("I heard that you told mum that Peter had done something mean to you as well") and he then says specifically that Peter did wees in his mouth; he elaborates on this, in response to a series of direct questions: a little went in his mouth and it was "yucky"; Peter was standing up with his pants and undies around his ankles. It happened one time. He claims he told Debbie and that she had told a policeman.

42. This interview is conducted in an orthodox manner and departs from the ideal only in the interviewer's use of prompts concerning what R had told his parents. The danger with this approach is that the child recalls what s/he said to the carer, rather than what actually occurred. It could, however, be justified on this occasion, by the information elicited. The interview is a little on the long side for a child of this age, but R shows no obvious signs of distress or distraction. R is reasonably consistent, but occasionally contradicts himself in response to direct questions. For instance, as regard what happened at the park, R first says nothing happened, then he says the "things with the penis" happened at the park and the creche. The interviewer checks this with a leading question to which he accedes, but after careful follow up questioning, he says nothing happened at the park. In general, I believe the interviewer did a good job in elucidating R's main complaints and following up any apparent contradictions.

43. Interview 2 with R takes place just over three weeks later and is just over a half hour in length. The setting is more formal with R and the interviewer sitting at a table drawing. The first part of the interview recapitulates some of the allegations made in the previous interview. His allegations are consistent with the previous interview: he still refuses or can't remember any other names beyond L. He saw it happen "looking through the toilet doors". R is again reluctant to talk about his alleged personal experience and the interviewer has to again refer back, this time to their previous interview, to get him to recapitulate his allegation about Peter weeing in his mouth and the answers he provides are consistent with those from interview 1, except that he can't or won't now remember which helper he told. Later, he says he told a helper on two occasions and that the policeman was only told the second time. There is a great deal of repetition of questions in this interview. Repeated questions can be a problem if the child takes this as a signal that the first answer was unsatisfactory or inadequate and that a different answer is desired. However, this does not occur on this occasion and R repeats his answers in a generally consistent way.

44. Interview 3 takes place some 6 months later and lasts just under an hour. R appears very composed and mature for his age. Again, they sit in front of a low table facing the camera. He has come back "to talk about the things and Peter". He has "got lots of things to tell" about Peter's behaviour in the Cranmer Centre. A follow up question then elicits a long statement from R, alleging that Peter had used a ladder to creep into another room where "mean things" took place. These included poking sticks up kids bottoms; he saw it happen and Peter did it to R; L was there. There were other men and women present, black as well as white, some with funny haircuts; their names were Spike, Boulderhead, Yuckhead and Stupidhead. He gives some details about the layout of the building: he claims to have gone on to the roof "every day". At one point, the interviewer reminds R to be sure to say he doesn't remember, rather than make things up-this would have been better said at the beginning. In response to specific questions, R adds a number of details about the stick incident: the sticks were sharp and yellow; he had to keep his in his bottom most of the day; it happened six times and Peter used the sticks to wipe poos on them. In response to a question, R says Peter put his finger in his bottom. As the interviewer returns to the details of the building, R shows increasing signs of restlessness and a wish to terminate the interview. More specific questions are asked which R answers. One illustrates the danger of too many specific questions, especially to a tired child recalling an event from the distant past (Int: "Were did they get their clothes from?" R: "Um, supermarket they bought them" Int: "Just tell me things you remember, okay?"). After further attempts to elicit more details and names, the interview is terminated.

45. The allegations that R makes in this late interview have clear parallels with those made by X. Such a parallelism is either independent corroboration, an example of contamination or elements of both: only careful scrutiny of the timing of the interviews and contacts between the families will answer this question. Certainly, in both X and R's account, there are some implausible elements (the claims about the sticks; keeping them in his anus all day). It will also be important to check X's and particularly R's account against the known architecture of the Cranmer Centre. An accompanied visit to the Centre seems to have occurred, but before the critical interview, rather than after it, which again leaves open the possibility of accounts being fueled by new information. It is not clear, why X or R left matters so late before describing the alleged incidents involving groups of adults. According to their earlier accounts, they had both been threatened by Peter about what would happen if they told-and both told the interviewer about this at an early stage. What motive could there be for such delayed disclosure? In these circumstances-and in the absence of an alternative explanation or external validation-there is a very real risk that large elements of the incident involving adults could be unreliable, driven by repeated requests to recall, the negative stereotyping of Mr Ellis, the conflation of separate and unrelated events and the sharing of information between families. Despite quite intensive questioning, R fails to produce many features of accurate accounts as reflected in CBCA criteria: there is a lack of spontaneous or distinctive detail, other than that elicited from focused questions. The events seem to occur in some parallel universe, divorced from the known routine of the creche. As mentioned elsewhere, the most reliable accounts from young children tend to be those which occur early on, before the opportunities for elaboration and contamination of the kind mentioned above have had the opportunity to occur. I think this is well illustrated in the contrast in R's evidence between the content of interview 1 and interview 3.

46. Z. Z was interviewed six times, all by Susan Sidey, between 27.2.92. and 29.10. 92. Z was 5 years 6 months at first interview and 6 years 2 months by the final interview. The recording quality is excellent and both participants are audible, despite Z's soft voice. Interview 1 lasts 45 minutes; the interviewer and Z sit on the floor while Z plays with a conventional doll as she talks: playing while talking is a good method with children of this age who can find a formal interview situation oppressive, provided that there are not so many toys as to be distracting (Davies and Westcott, 1999). Peter is introduced by Z ("I want to say something about Peter") during truth and lies: Z seems to have a rather uncertain grasp of the formal distinction, but says twice that she "always tells the truth". Z says in response to an open question ("What did Peter do that was nasty?"), Z says "He used to show me his penis in the toilet" Later she says "it sticked out from, in the side of his body...Peter used to move it up and down and stuff but we didn't like that". Soon after, she spontaneously demonstrates the movement on a toy baby's bottle. "Some very nasty things" happened to Q and P. When the interviewer asked how she knew that, she replies "cos I was watching. I tried to help the little children but they didn't believe me" . Later she expands on this and says, with some emphasis, that she had told Marie, who had not believed her. Later she adds that Q and P "used to suck his penis". After 15 minutes, she begins to get restless, asks for a play break which the interviewer accommodates. When the interviewer asks whether Peter's penis went into anybody else's mouth, she replies "in my mouth", adding that it "feeled rough" and "baby stuff came out of it". Later questioning reveals that Mummy has told her about "baby stuff' adding that "me knew the colour and it was plain white". She says that it happened "lots of times"; "only on Mondays and Fridays". Later, she points to the mouth of a doll when asked where Peter's penis went. When asked "where did you get that sick feeling?", Z says "when I got that stuff in my mouth". She says that his penis did not touch her elsewhere. She demonstrates this again with miniature dolls; the demonstration is not clear from the tape, but the interviewer comments "so he put the penis in your mouth". Later she adds that this happened in the "kids toilets" at nursery. Later, the conversation moves to Peter's walks with the children; when asked whether anything she didn't like happened, Z says "No". After more play, the interviewer asks whether "anyone ever told you not to tell or not" to which Z replies "No. Yes. Mum, I said that to my Mum", adding that "he would hurt me and that he would burn his, my parents up". The interviewer returns to touching and Z responds by indicating her groin area and saying "he only touched me here"; when asked, she says it was with his penis. Z shows increasing signs of wanting to terminate the interview (earlier she has complained of feeling sick). The interviewer suggests she may have to come back another day and closes the interview.

47. Interview 2 takes place a day later and lasts 40 minutes. It begins with an extended play session lasting around 25 minutes, at the end of which, the interviewer introduces a pair of anatomical dolls. Z remarks that yesterday she didn't like it here because "it was really a bit too long" (hence the extended play in this session?). The doll's vagina triggers a discussion of Peter having seen her vagina in the toilet. Z mentions Peter's penis and the interviewer asks her "What did his penis see, er do?" In response to an options posing question, Z says Peter's penis was "hard" and to a later, forced-choice question, it was "standing up". Initially, there is some ambiguity as to whether he touched Z with his penis above or below her clothing; later she demonstrates with the dolls and says "yeah" when the interviewer comments "so it went underneath your pants" and when she poses the alternative "Or over the top?" she replies "underneath".

48. The third interview took place three weeks later and followed further discussions with her mother and lasts approximately 3 5 minutes. Z by now knows the score and interrupts the preliminaries saying "I know a lot of things to do with Peter". The use of `know' in this context may be significant in that it may not imply first-hand knowledge. She now says Peter took her to his house and they begin to construct a schematic layout of the house with toys (this could be checked for accuracy). When asked "what the worst thing he did to your body?" she says "touching it" and further questions cause her to indicate that he touched her bottom. However, it is not clear whether she is talking about at the creche or Peter's house here. She repeats the allegation again later, adding "everyday" at creche (as indicated earlier, children's estimates of frequency at this age need to be treated with considerable caution). It is clear that she is now definitely talking about the creche. She then repeats her earlier allegation about Peter putting his penis in her vagina, adding under questioning that her own underpants were off and on the floor. Given options by the interviewer, she says she was kneeling and Peter standing up. A little later, she says she was standing too. The interviewer queries this and Z insists that this is correct. The interviewer then fetches first a pair of miniature dolls, then a pair of larger, anatomical dolls. Z clearly juxtaposes the dolls, such that the penis of the male doll is placed in the anus of the female. However, she now says that she was standing up and he was lying down. Z then requests a break, which is given and soon after, asks to go home. There is some more play before the interviewer introduces the topic of Peter once more, but Z evades the question and the interviewer draws the interview to a close.

49. These three interviews contain clear allegations against Mr Ellis concerning events in the creche. It is difficult to see how these could have arisen as a result of defective interviewing practice. The interviews are, in my opinion, generally well conducted, though there is some leading by the interviewer at critical points. The use of anatomical dolls is appropriate, given the age of the child and they are not accompanied by suggestive questioning. As with a number of the other children, the crucial question concerns the origins of the allegation: is Z talking about a personal experience or relating a coached story? Z has certainly talked to her mother about the allegations in the recent past and this was confirmed at court. The fact that she broaches the topic herself so early on in two of the three interviews is an indication that she has been made aware of the agenda. Her use of adult terms like `penis' and `vagina', may also reflect conversations with her carers. In interview 2, she mentions talking to her mother twice, once in relation to the "baby stuff' remark, which clearly reflects adult input. However, the allegations, though vague in some details (the exact position of herself and the accused: who was sitting and who was standing) are plausible, reasonably detailed given the presumed age of the memory, consistent and plausibly enacted with dolls. Much of it does not arise from the kind of specific questions which characterise some other interviews. Given existing research on coaching, it is hard to see how such an account could have arisen from this source alone.

50. The fourth interview takes place just over a week later and lasts around 40 minutes. Z takes the interviewer through the preliminaries, after which the interviewer leads "I know you have been to some places with Peter. Where are all the places you've been with him?" There are references to Peter's car: "Peter hasn't got a flash car ....he has a white one like this" (it emerged at trial that Z had seen a white car on TV featured in a news report). Later attempts to elicit more information on `Peter's car' are unsuccessful or diverted by Z. When pressed by the interviewer as to where Peter went with her, Z says that he took her "Nowhere else just to the beach and Willowbank", but the interviewer insists "Because I think you've told that you have been to his house before" (reference to an exchange in interview 3). Z says "OK, what I'd like to do is make Peter's house". She goes on to say that Peter lives alone and has "lots of friends, bad friends" but when she asks how she knows (a shrewd question) replies "because mummy telled me" . The interviewer asks what his friends were doing and Z says "showing the penis and `gina". She is invited to name Peter's friends and gives some names and in response to a series of direct questions says that "Joseph... teased me with his penis". The interviewer is interested in Joseph but Z keeps returning to Peter. A further request for names produces more names. When asked a direct question ("how many other big men's penises have you seen?") she replies "lots" but then adds "but not as many". There follows a confused passage where the interviewer tries to elicit more detail and it is unclear whether Z is talking about Peter or the "other men". There is a short break as Z is showing signs of restlessness, before a renewed attempt to obtain more information on `Joseph', but Z gives nothing of significance: for instance when asked how often Joseph's penis touched her, she replies "Only once, every day he did different things" , but when asked what `other things', she replies "Nothing else". Shortly afterward, the interview is brought to a close.

51. There is nothing in this interview which convinced me that Z visited Peter's house or was assaulted by a man called Joseph. It begins in confusion: the first exchanges make it unclear whether Z is talking about what she has personally experienced or what her mother has told her (signified by the use of `know' as opposed to `remember'). At one point, she says Peter took her to the park and Willowbank only, but the interviewer presses her on her visit to his house. As I noted above, it was not clearly established in interview 3 that Z had actually visited Peter's house and most of her comments referred to activities at the creche. The interviewer then asks her to name the people involved and she twice provides a list of names, but this could simply be a response to the demands of the question, rather than having any basis in reality. All subsequent information is derived from forced-choice or direct questions and there is no supporting detail or spontaneous comments of any substance. In one revealing passage, the interviewer wants to talk about Joseph, but Z continues to bring the conversation back to Peter. In conclusion, I felt that the solid evidential content of this interview was minimal and there was some confusion by both parties. A visit to Peter's house may have occurred, but this was not satisfactorily established in this interview.

52. Interviews 5 and 6 took place 7 months later on consecutive days: 28 and 29.10.92 respectively. Interview 5 lasted about 40 minutes and interview 6 barely 20 minutes. At the start of Interview S, Z announces " mum said I don't have to do too much tacking because I haven't got much things to tell you". Later she says she wants to talk about "Peter's friend and people who hurt you... and about Peter's mother". "Peter's friends touched me in my private places and I didn't really like it and hit me". The interviewer cautions Z to tell her if she can't remember and that it is better to say this than make things up. New names are mentioned by Z (not the ones mentioned in interview 4) including `Andrew'. The interviewer tries unsuccessfully to get Z to describe Andrew. Z says Andrew hit Z and touched her vagina with a knife. This happened when Peter and his other friends were present, but later she says it took place in the toilets at the creche: she doesn't know why Andrew was there. She cannot describe the knife beyond saying it was " very, very sharp". Later, she says the friends "had gone away" and only Peter was there and he laughed. The interviewer presses Z to describe the friends and when she does not, she cautions her about only telling what she remembers. Later the issue of the knife is revisited. Z repeats the allegation, but now in relation to Peter ("I only know that Peter has"): it was at the same time as Andrew. Z can say little about Peter's mum ("she had grey hair"): "she just kicked me and hit me like this". According to Z, this happened 17 times and Marie, Gaye and Debbie, helpers from the creche, were present. Later the interviewer asks "What's it feel like having to remember these things?" and Z replies "Not very good, but I can't remember anything much now". When the interviewer pursues the location of this alleged assault, Z says "I can't remember at the creche, I said that a million times... That's all I can remember Sue". Z wants to terminate the interview and after a further question, the interview is closed.

53. Interview 6 starts again the following afternoon. Z says defensively "she told the truth yesterday". The interviewer invests sometime in trying to elicit from Z a clear distinction between truth and lies. Z says "I haven't got much to tell you ", but makes a fresh allegation against Gaye, claiming that she touched her on the bottom (later her vagina) with a knife at creche. Gaye had asked her to take her clothes off Later Z is asked whether "anyone else has put a knife on her vagina before, she replies "No". This seems to contradict what she has alleged in relation to Andrew and Peter. The interviewer again asks about the knife, but Z can only describe it as "a normal sharp knife". Later, the interviewer again asks about other people who touched her vagina with a knife. Z is getting increasingly restless. At first she says "I cant remember" but then repeats her allegation against Andrew and Peter. The interviewer asks whether the allegation against Gaye really happened and Z does not answer; the interview concludes shortly afterward.

54. In my view, these two late interviews subtract from, rather than add to Z's credibility as a witness. The allegations are not accompanied by any clear details, the locations of the incidents are unclear and she appears to be inconsistent regarding who and what was involved. The interviewer, does a good job in elucidating the inconsistencies in these allegations, through open-ended questions; her gentle remonstrance to Z to stick to what really happened indicates that perhaps her own sense of the credible is being strained. I am not surprised that these two tapes were not used at trial. They do not, however, in my opinion detract from the earlier allegations, made some months before, in interviews 1, 2 and 3 against Mr Ellis concerning incidents) in the toilet at creche, for reasons I have given above.

55. What conclusions can be drawn from this review of the content and conduct of the children's interviews and, more widely, of the methods used in the investigation? I will confine my concluding comments to four main issues which I believe are central to consideration of the children's testimony:

  • The age of the children and the age of the memories
  • The style and quality of interviewing assessed against current standards
  • The bizarre elements in some of the accounts
  • The risks of contamination

56. The age of the children and the age of the memories The first point to make is that these are very young children recalling events which allegedly occurred to them a year or 18 months previously. Even though all of the 6 children who are the subject of this review come across as bright and alert in their interviews, they cannot be expected to provide the kind of detailed and spontaneous accounts which are so useful from the point of view of making judgements on reliability. Young children forget events more rapidly than older children or adults. Accounts will inevitably be schematic and some prompting will probably be necessary to elicit even the most basic details: we cannot and should not expect a vivid and detailed account in these circumstances and nor in general do we get one from any of the children. On the other hand, we should expect (following CBCA) to be able to fit the jumbled pieces of the child's narrative together to form a coherent account which is consistent with the known events in the child's life, those of significant players in the story and the setting and movements should correspond to documented reality. Judged by these criteria, I suggest that this is much easier to fulfill this for the alleged events at the creche, than for those incidents which allegedly occurred at locations away from the creche. All the children reviewed chose to tell about events at the creche in their first interviews and some, when asked, explicitly denied that Mr Ellis had done anything untoward on any walks or visits outside of the creche. Virtually all the allegations concerning events outside the creche emerged in later interviews, often after considerable delay. It is evident that this made for an even longer delay between the original events and recounting them and gave even greater opportunities for cross- tack between families and their children. Of course, as I have noted earlier, such a delay could have been due to a reluctance by the children to discuss the matter earlier, born of personal shame and the threats which the children said they received from the alleged abuser. This possible explanation would have added credibility if the investigation turned up any positive proof that the children had indeed visited Mr Ellis's house or indeed any of the other locations with him (there is some supporting evidence, e.g. Tronson's deposition regarding 404 Hereford Street). Likewise, if they displayed privileged knowledge, which they would have been unlikely to have acquired other than through such a visit. Conversely, evidence that such visits were physically unlikely to have taken place-for instance because of the distances involved or the lack of access by Mr Ellis to a vehicle-would be likely to reduce the credibility of the allegations.

57. The style and quality of interviewing assessed against current standards As I have noted, the interviews were conducted before the introduction of national standards for interviewing. However, judged by today's standards, the quality of interviewing stands up surprisingly well. There is an attempt to set ground rules during the rapport stage (in hindsight, the necessity of children to tell only what they remembered and to make explicit the `don't know' option could have been more heavily emphasised from the start), to minimise suggestible questioning and use open-ended questions whenever possible. There are few of the gross violations of interviewing practice of the kind recorded by Bruck and Ceci(1995) in the Michaels' case. Children are not fed tidbits of information from other children (though the use of the same interviewer for all but two of the interviews left open this possibility). Children are not refused breaks until they tell what the interviewer wants to hear, though some of the interviews do go on too long in my opinion and requests for the children to see their carer or to terminate the interview are on occasion temporarily set aside. It is noteworthy that some of the more bizarre and inconsistent claims tend to occur toward the end of long interviews. Children are sometimes asked leading questions or referred back to earlier allegations made by them; most Codes of Practice strongly discourage the former, but allow the latter. Of more concern, in my view, is the frequent use of direct, closed or multiple choice questions in some interviews, where the age of the child and age of memory would be unlikely to provide an accurate response. Such questions can produce accurate information where the witness's memory for the event is vivid and detailed. However, as Waterman's work (2000) has shown, where there is no memory as such, closed questions can produce answers, particularly from young children, which have no correspondence to reality. A young child's readiness to answer such questions could prompt the interviewer to ask further questions of this type. I hasten to add that this is not a fault unique to these interviews: many studies have reported a preponderance of direct as opposed to open-ended questions in investigative interviews (e.g. Davies, Westcott and Horan, 1999). If there was one major weakness in the investigation, it was the sheer number of interviews that some children underwent over an extended time period. Most Guidelines recommend one, or at most two interviews per child, though the new Guidance currently being developed in the United Kingdom does acknowledge that very young children of the kind involved in the current investigation may require a number of shorter interviews conducted over a limited time period. The risks of suggestible responding with spaced multiple interviews have already been documented, in particular the adverse effect of repeated questioning concerning one key incident. One can sympathise with the interviewers. They were dealing with very young children: this fact alone might have pointed to the need for a number of shorter interviews rather than a single long interview. However, in my view it did not justify interviewing the child again after a delay of anything up to 6 or 7 months. I can also appreciate they were under some pressure from parents and carers, but based on my reading of the case material, little of value came out of these late interviews: they tended to devalue the earlier interviews given by the children. In sum, while mistakes were made (hindsight is a wonderful thing), the general standard of interviewing was good and exceptional for the time the interviews were made, particularly given the age of the children interviewed. Mistakes were made, but these were insufficient in my view, to explain the content of the allegations made by the children, particularly regarding events at the creche. As regard alleged events outside the creche-on walks, at Mr Ellis's house and elsewhere- the repeated use of specific questions, sometimes after initial denials that anything untoward had occurred, may have contributed to confabulations by some witnesses, though as I have noted, both X and S provided plausible accounts of outside events, which could not readily be attributed to undue prompting or inappropriate questioning..

58. The bizarre elements in some of the accounts All but one of the accounts contains material which could be regarded as bizarre or fanciful. Such evidence may be merely implausible or incredible. Among the implausible elements are the accusations of group sex made against the creche workers and children being placed in an oven or hung from cages in the ceiling. As the West and Detroux cases demonstrate, nightmarish happenings can occasionally occur, but their frequency is fortunately very low indeed and accordingly, strong and convincing corroboration would be required, beyond the simple accusation of an eyewitness, whether child or adult. By incredible accusations, are meant those which defy the laws of nature or at variance with the known facts. In this category I would place such accusations as needles being placed in children's penises or vaginas and of a child being forced to stay all day at the creche with a sharpened stick in his anus. Even supposing this were possible, abuse of this kind would be likely to be readily detected by carers. While I understand medical evidence consistent with sexual abuse was found for three of the children, there were no indications to support the view that any of the children whose interviews I examined had signs of injury consistent with physical abuse of this kind.

59. If it is accepted that at least some elements of a child's accusations are unbelievable, where does this leave the more credible elements? In terms of law, such bizarre material is bound to decrease the credibility of a witness in the eyes of a jury. However, some pilot research on young children in abuse cases suggests that such bizarre material can crop up in their evidence alongside telling and accurate testimony: the presence of bizarre elements is not in itself proof that the whole of the child's testimony is tainted (Everson, 1997).

Dalenberg (1996) examined the testimony of a sample of 644 children involved in sexual abuse investigations of whom 2% included fantastical elements in their testimony. Rates of inclusion of such material were no higher among a group whose testimony was corroborated by confessions or medical evidence compared to the evidence of those where corroboration was lacking. The highest rates of inclusion of bizarre material was among the youngest children and those suffering the most severe and prolonged abuse.

60. The risks of contamination A major concern in this and in all investigations involving abuse allegations within groups of all ages is that witness's evidence will be contaminated by discussion between them or their carers. According to the agreed timetable, Mr Ellis had been suspended in the previous November and Susan Sidey had already conducted a number of interviews with children in December which had failed to turn up any statements consistent with abuse. It is not possible to comment on the content or conduct of these interviews in that, with one exception, no tapes of them were submitted as evidence. In any event, I am satisfied that the investigators were aware of the risk of contamination and at the meeting on March 31, warned parents of this danger and of using direct and leading questions with their children (see Ms Crawford's evidence). Unfortunately, this warning probably came too late for some parents: despite the good advice provided at the meeting on March 31, some parents readily admitted in their anxiety to find out what, if anything, had happened to their children, they pressed them for information subsequent to and perhaps prior to the March meeting, using precisely the procedures which they had been warned against. As I have noted, there is certainly indications in the transcripts of the video-recorded interviews that children had discussed matters with their parents in varying degrees: only one of the 6 children seems to have turned up at their interview apparently entirely unbriefed as to its purpose. Such consultation between parent and carer does not in itself invalidate evidence: if every witness, whether adult or child, who had discussed the content of their evidence with a loved one prior to trial were excluded from testifying, there would be precious few trials. Nor are one or two leading questions alone likely to change radically the perspective of the testimony of even a small child. Of more concern, is cross -talk between families, against a background of persistent accusation against a suspect. In my analysis of the tapes, I have already indicated that some of the more improbable incidents reported by two or more children may have their origins in such cross-talk. I am also satisfied that the unsupervised visits of children to various locations of alleged events may also have coloured the accusations made in the later interviews. Careful scrutiny will be required of the timetable to establish which parents and children were in contact at what times and how this relates to the information provided at interview. In conclusion, I should add that I do not think that cross-talk alone is sufficient to explain the similar accusations made against Mr Ellis, particularly in relation to incidents in the creche toilets. This is more redolent of natural variations on a theme, rather than the simple glib repetition of basically the same accusation relayed between a number of children. There is a degree of detail in some of these accusations which is not in itself proof of the charges against Mr Ellis, but does mean, in my view, that such accusations deserve to be taken seriously. They need to be studied in the wider context of the investigation. For instance, do the toilet facilities at the creche correspond in their layout and construction to those described by the children? Did the physical layout of the creche afford sufficient privacy to ensure that abuse of the kind described by the children could have taken place? Where children specify that other children were present, does the evidence of these children corroborate their allegation? These are issues which are beyond my remit, but which the wider inquiry will wish to consider.

Graham Davies
4 January, 2001


Bruck, M., & Ceci, S. J. (1995). Amicus brief for the case of state of New Jersey v. Michaels, presented by Committee of Concerned social Scientists. Psychology, Public Policy and Law, 1, 272322.

Ceci, S. J., Huffman, M. L. C., Smith, E. & Loftus, E. (1994). Repeatedly thinking about a non-event: Source misattributions among preschoolers. Consciousness and Cognition, 3, 388-407.

Clyde, The Right Hon. Lord, (1992). Report into the Inquiry into the removal of children from Orkney in February, 1991. Edinburgh: HMSO

Dahlenberg, C. J. (1996). Fantastic elements in child disclosures of abuse. APSAC Advisor, 9(2), 110.

Davies, G. M. (1991). Research on children's testimony: Implications for interview practice. In C. R. Hollin & K. Howells (Ed.), Clinical approaches to sex offenders and their victims. Chichester: Wiley.

Davies, G. M. (in press). Is it possible to discriminate true from false memories? In G. M. Davies and T. Dalgleish (eds.), Recovered memories: Seeking the middle ground. Chichester: Wiley.

Davies, G. M. & Westcott, H. L. Interviewing child witnesses under the Memorandum of Good Practice': A research review. London: Home Office (Police Research Series, No. 115).

Davies, G: M., Westcott, H. & Horan, N. (2000). The impact of questioning style on the content of investigative interviews with suspected child sexual abuse victims. Psychology, Crime and Law, 6, 8197.

Everson, M. D. (1997). Understanding Bizarre, improbable and fantastic elements in children's accounts of abuse. Child Maltreatment, 2, 134-149.

Geddie, L., Fradin, S. & Beer, J. (2000). Child characteristics which impact accuracy of recall and suggestibility in preschoolers: Is age the best predictor? Child Abuse and Neglect, 24, 223-235.

Gross, J. & Hayne, H. (1999). Drawing facilitates children's reports after long delays. Journal of Experimental Psychology: Applied, 5, 265-283.

Lampinen, J. M. & Smith, V. L. (1995). The incredible (and sometimes incredulous) child witness: Child eyewitnesses' sensitivity to source credibility cues. Journal of Applied Psychology, 80, 621-627.

Leichtman, M. D. & Ceci, S. J. (1995). The effects of stereotypes and suggestions on preschoolers' reports. Developmental Psychology, 31, 568-578.

Poole, D. A. & Lindsay, D. S. (1995). Interviewing preschoolers: Effects of nonsuggestive techniques, parental coaching, and leading questions on reports of nonexperienced events. Journal of Experimental Child Psychology, 60, 129-154.

Volbert, R, & van der Zanden, R (1996). Sexual knowledge and behaviour of children up to 12 years. What is age appropriate? In G. Davies, S. Lloyd-Bostock, M. McMurran & C. Wilson (Eds.), Psychology, law and criminal justice: International developments in research and practice. Berlin: Walter de Gruyter.

Waterman, A. (2000). Why do children respond positively to non sensible questions? Unpublished Doctoral dissertation, University of Sheffield

Criteria-based content analysis (CBCA)
(Reprinted from Raskin & Esplin, Copyright 1991, with permission from Elsevier Science)

General characteristics:

(1) Logical Structure Is the statement coherent? Is the content logical? Do the different segments fit together? (Note: Peculiar or unique details or unexpected complications do not diminish logical structure)

(2) Unstructured Production Are descriptions unconstrained? Is the report somewhat unorganized? Are there digressions or spontaneous shifts of focus? Are some elements distributed throughout? (Note: This criterion requires that the account is logically consistent.)

(3) Quantity of Details Are these specific descriptions of place or time? Are persons, objects, and events specifically described? (Note: Repetitions do not count.)

Specific Contents:

(4) Contextual Embedding Are events placed in spatial and temporal context? Is the action connected to other incidental events, such as routine daily occurrences?

(5) Interactions Are there reports of actions and reactions or conversations composed of a minimum of three elements involving at least the accused and the witness?

(6) Reproduction of Speech Is speech or conversation during the incident reported in its original form? (Note: Unfamiliar terms or quotes are especially strong indicators, even when attributed to only one participant.)

(7) Unexpected Complications Was there an unplanned interruption or an unexpected complication or difficulty during the sexual incident?

(8) Unusual Details Are there details of persons, objects, or events that are unusual, yet meaningful in this context? (Note: Unusual details must be realistic.)

(9) Superfluous Details Are peripheral details described in connection with the alleged sexual events that are not essential and do not contribute directly to the specific allegation? (Note: If a passage satisfies any of the specific criteria 4-18, it probably is not superfluous.)

(10) Accurately Reported Details Misunderstood Did the child correctly describe an object or event but interpret it incorrectly?

(11) Related External Associations Is there reference to a sexually-toned event or conversation of a sexual nature that is related in some way to the incident but is not part of the alleged sexual offenses?

(12) Subjective Experience Did the child describe feelings or thoughts experienced at the time of the incident? (Note: This criterion is not satisfied when the witness responds to a direct question, unless the answer goes beyond the question.)

(13) Attribution of Accused's Mental State Is there reference to the alleged perpetrator's feelings or thoughts during the incident? (Note: Descriptions of overt behaviour do not qualify.)

Motivation-Related contents:

(14) Spontaneous Corrections or Additions Were corrections offered or information added to material previously provided in the statement? (Note: Responses to direct questions do not qualify.)

(15) Admitting Lack of Memory or Knowledge Did the child indicate lack of memory or knowledge of an aspect of the incident? (Note: In response to a direct question, the answer must go beyond "I don't know" or "I can't remember".)

(16) Raising Doubts About One's Own Testimony Did the child express concern that some part of the statement seems incorrect or unbelievable? (Note: Merely asserting that one is telling the truth does not qualify.)

(17) Self-Deprecation Did the child describe some aspect of his/her behaviour related to the sexual incident as wrong or inappropriate?

(18) Pardoning the Accused Did the child make excuses for or fail to blame the alleged perpetrator, minimize the seriousness of the acts, or fail to add to the allegation when the opportunity occurred?