Kenyon v R. [2010] EWCA Crim 914

[2010] EWCA Crim 914Case No 200806813 C1
IN THE SUPREME COURT OF JUDICATURE
COURT OF APPEAL (CRIMINAL DIVISION)
Venue ON APPEAL FROM LEEDS CROWN COURT
MR JUSTICE GRIGSON
T20027586
Venue Royal Courts of Justice, Strand, London, WC2A 2LLDate 11/05/2010
Before
LORD JUSTICE HUGHESVICE PRESIDENT OF THE COURT OF APPEAL CRIMINAL DIVISIONMR JUSTICE MACKAYMR JUSTICE LLOYD JONES
Between
Julie KenyonAppellantThe QueenRespondent
Mr P Dunkels QC and Miss S Shapiro (instructed by Morris and Warren) for DefendantMr R Smith QC (instructed by Crown Prosecution Service) for CrownHearing Hearing dates : 19 th February 2010
Judgment
[1]This appeal against conviction comes to the court by way of reference by the Criminal Cases Review Commission.[2]The appellant lived with her grandmother. The grandmother died overnight on 4/5 December 1996 at the age of 89. At the time no foul play was suspected. Subsequently, however, over a period of several years, evidence emerged that the appellant told three different people that she had smothered her grandmother. She was tried before Grigson J and a jury in July 2003. The case against her depended upon her confessions. She denied making one of them. The other two she admitted making but contended that they were false. The issue at trial was therefore whether such confessions as she had made could safely be relied upon as proof that she had indeed done what she had said she had done. The jury convicted her.[3]The appellant now seeks leave to adduce before this court the evidence of two experts as to her psychological profile and in particular the likelihood of her making false confessions. One expert, Professor Gudjonsson, is a non-medical psychologist of considerable experience whose particular area of interest is unreliable confessions. The other, Dr Johns, is a forensic psychiatrist.[4]Both a psychiatrist and a psychologist were consulted on behalf of the appellant prior to her trial. The experts consulted prior to trial specifically addressed the question whether the appellant was of such a nature as to be likely to make false confessions. That is hardly surprising, since plainly the principal issue which was going to arise at the trial was whether she might have done so. In the end, however, neither expert was relied upon at the trial. That was a considered decision. The appellant’s case, however, is that the evidence of the two new experts shows that the conviction is unsafe. It is said on her behalf that the evidence now available goes further than that which might have been called at the trial.[5]This court has power to receive evidence which was not adduced at the trial. The rules governing the exercise of the power are set out in section 23 of the Criminal Appeal Act 1968 and are very well known. The critical test is whether the receipt of the evidence is “necessary or expedient in the interests of justice.” As this court emphasised in Erskine [2009] EWCA Crim 1425, that is a highly fact-specific question.[6]We thought it right to hear the evidence of Professor Gudjonsson and Dr Johns de bene esse. It was apparent that the evidence which they could give went somewhat beyond the expert evidence which was available at the time of the trial. Their evidence occupied approximately a day before us. We have heard also from Professor Grubin, who is a forensic psychiatrist, called on behalf of the Crown on the same basis.

Outline history

[7]Grandmother lived in a one-bedroom flat. For several months before she died, the defendant had been living there too. The defendant slept in the bedroom and Grandmother on the settee in the living room. Grandmother was 89 and frail. She had osteoporosis, arthritis and vascular disease and was very thin at 32 Kg. She was in some pain, taking a variety of medication, and sometimes complained that she should not be long for this world. But she was fairly doughty. Two days before her death she had had a chip pan fire in the flat; the policeman who visited had found that she had put it out herself with a damp towel and was cleaning up, rather annoyed with herself. Although there had been some smoke, the post mortem examination subsequently ruled out any contribution to her death from this accident.[8]Grandmother had a neighbour in a nearby flat called Mrs Ledingham. She had visited Grandmother that evening when the defendant was out, had given her some cigarettes and had parted upon Grandmother’s remark that she would see her in the morning. At about 0130 or 0200 on the morning of 5 December 1996 the defendant banged on Mrs Ledingham’s door and told her that she could not wake Grandmother up. Grandmother was dead.[9]At the time, the defendant’s account was that she had come home from the public house, had gone to bed at the same time as Grandmother at about 2230, both having taken sleeping pills, and that she had then been awakened at about 0130 by hearing Grandmother call her name. Pausing only to pull on a pair of jeans, she had gone into the living room but the lady was beyond rousing.[10]There was no sign of injury on Grandmother, but if she had been smothered with a pillow or cushion there was no reason why there should be. Nothing sinister was suspected by the authorities and cremation followed.[11]Nearly five years later, in October 2001, the defendant’s sister Carol reported to the police that the defendant had admitted smothering Grandmother. It turned out that by then there had been three (strictly four) occasions when the defendant was said to have made such admissions. In chronological order they were as follows. i) At a time which was not clearly proved, the defendant had told her mother that she had put a pillow over Grandmother’s face. The defendant had been on very bad terms with her mother since childhood, it may well be because Mother had been complicit in sexual or other abuse of the defendant by one of her partners. Mother was not called to give evidence at the trial. The confession to her was proved by the evidence of Carol (see (iv) below) and was admitted by the defendant to have been made. On the defendant’s evidence this occurred in October 1997. ii) A man called Kevin Donegan gave evidence that in August 2001 or thereabouts he had chanced to meet the defendant in a public house called the Railway. She was then the girlfriend of someone he knew, called ‘Lefty’, and he also knew her from childhood. He and she fell into conversation whilst Lefty was playing pool. The conversation turned to Grandmother, and eventually the defendant told him that the old lady had been poorly and had wanted to bring her life to an end. None of her family was prepared to help, but one night she, the defendant, had crept into the room whilst Grandmother slept and had held down a pillow over her face until she died. According to both Donegan and Carol, he had subsequently told Carol what the defendant had said. iii) On Monday 17 September 2001 the defendant’s sister Carol met the defendant in the Union Castle public house. Carol taxed the defendant with killing Grandmother. Initially the defendant denied it, but when pressed she admitted it. Her account was that she had put a pillow over Grandmother’s head, and that Grandmother had asked her to do it. iv) As a result of this, Carol arranged to meet the defendant again on Friday 21 September and this time equipped herself with a hidden tape recorder. A transcript of the conversation was before the jury. The defendant said that she had come home and made a pot of tea for herself and Grandmother as usual. Grandmother had been in jolly, but also reminiscent, mood and they had talked for some time. They both went to bed. Some time later, the defendant heard Grandmother call her name. She went in and Grandmother told her “I want to go” and asked her to help. She, the defendant, refused, saying that she would get ‘done for murder’. But Grandmother persisted and put a pillow over her own head. Said the defendant, she had reluctantly put her hand on the pillow. Grandmother had turned her face into the pillow and she, the defendant, had ‘finished it’. In this taped conversation, the defendant can be heard volunteering, on three different occasions, that she had also told her mother what she had done.[12]In due course, at her trial, the defendant admitted the confessions to Mother and Carol. But she denied that any conversation such as Donegan spoke of had ever taken place. She asserted that she would never have said such a thing to him, in particular because she did not like him because when a paperboy years beforehand he had fiddled her family’s paper money.[13]The defendant’s case is and was that she had been “hounded” over the years by other members of the family accusing her of killing Grandmother, and that gossip about this allegation was rife. Insofar as that depends on her unsupported assertion, as in many places it does, it needs to be treated with considerable caution. On any view she is an extremely unreliable historian. A reading of the careful chronology compiled by the CCRC confirms this in relation to many issues, and it is the proposition on which the whole of this appeal on her behalf is founded. But there is some other evidence which appears to demonstrate that there was such suspicion and that from time to time it was voiced. The first complaint, made by another half-sister Shirley in the course of a family row on the day after the death, seems not to have been an accusation of killing, but rather a complaint that the defendant (and indeed Carol) had neglected Grandmother. But there is objective evidence that this family was riven by a good deal of internecine strife, with fluctuating loyalties. Shirley gave evidence at trial that there had been such gossip and that the defendant had been aware of it. Carol’s own evidence was that confession number (iii), the first to her, was made in response to an accusation, and one which was pressed in the face of initial denial. Subsequently Carol has given a journalistic interview confirming that she entertained such suspicions from at any rate a few months after the death and had asked the defendant about them then. At trial, the defendant’s explanation for what she had said to Carol was that she was just telling her what she wanted to hear, in order to ‘get her off my back’. She added that she was “drunk, angry and disgusted” at the time.[14]That was not the defendant’s explanation for the confession to Mother. That one, she said, had been made because she hated her Mother as a result of childhood abuse, and wanted to see pain in her eyes. On her account, it had not achieved that impact and so she had then told her mother that it was not true anyway.[15]The defendant was arrested on 19 October 2001. Her interviews with the police are of some significance to the issues raised before us. She began by roundly refuting any suggestion that she had killed Grandmother. Asked about her mother, she readily said that she hated her, but far from telling the police that she had, for that reason, confessed falsely to killing Grandmother, she denied making any such confession, and with some indignation. The part of the interview containing this denial was not, so far as we can see, before the jury, no doubt because it depended on quotation from a statement made by mother whose evidence was not adduced at trial. The defendant went on to deny equally firmly that she had made the confessions to Donegan and to Carol; such a suggestion was, she said, ‘sick’ and ‘pathetic’. It was only after the taped conversation with Carol was read out in full and played over that she accepted that she had made that confession and also that she had told her mother that she had done it. There was, as we have shown, an assertion on the tape of having told mother this. She continued to deny the Donegan conversation, as she did at trial, but also to deny that she had made the earlier Monday confession ((iii) above) to Carol, which she was later to admit at trial. At one stage in the interview the policeman questioning her raised in very tactful terms the concept of ‘mercy killing’ and spoke in generalities about the possibility that the terminally ill might be ‘helped on their way’ by relatives. The defendant assertively refuted this possibility:
“A: You are asking me if I put a pillow over her face and…. Q: Not out of spite but out of about helping her. A: ….and did it to help her. No I did not. I held her hand.”
The defendant became upset. After a break of about two hours and the introduction of a solicitor on her behalf she went on to give a detailed and circumstantial account of Grandmother holding her hand and asking her to end the pain and misery which she was in, but of her refusing to do so. At that, she said that Grandmother had “turned onto her front, face into the pillow” and she had left her. She added that Grandmother had told her what she was going to do. She said that she, the defendant, was “scared and upset and did not want to be there.” Then, later, she had heard her name called, had gone back into the living room, and had been unable to rouse Grandmother. A: ….and did it to help her. No I did not. I held her hand.”[16]The arrival of the solicitor was the result of the decision of the Inspector who reviewed her detention, who was concerned that the defendant ought to have advice. The custody record had noted from the outset that the defendant “suffers from depression and is taking medication for the same” and the sergeant had accordingly directed that, no doubt like others with a similar history, she should be under constant supervision and her cell door was left open. On arrival, she was offered a solicitor but declined. She was interviewed within the hour of arrival and the offer was repeated but again declined. Later in the afternoon, after she had become upset at the conclusion of the third interview, the inspector positively advised her to accept and she did so. She was examined by a doctor at the end of the afternoon and found to be fit for detention, and he advised that there should not be further questioning after she had received the valium which he prescribed. In the course of the interviews she described the interviewers more than once as having treated her “brilliantly” all day, and the transcript supports that description. In her evidence at trial she was to suggest that whilst they had been kind on tape they had been less so out of formal interview and overall had been “breaking her down” by their questions. Psychiatric and psychological evidence[17]This case was clearly prepared with care on behalf of the defendant. As part of it, the defendant was examined by both a forensic psychiatrist, Dr Bhattacherjee and by a forensic psychologist, Mr Wright. Dr Bhattacherjee was concerned first and foremost whether there was any mental illness or medical defence which ought to be raised, which there was not. But he investigated the defendant’s undoubtedly complex and at times sad social history and he addressed the issue of the confessions. He said this:
“Given her vulnerabilities and difficulties coping, I think that Julie Kenyon is particularly susceptible to coercion in social situations. Given the significance of this aspect of her personality in relation to the current allegations I think that it would be appropriate to obtain formal psychological testing in this case. The object of this testing would be to formally assess her degree of suggestibility and susceptibility to coercion in social situations.”
[18]The result was Mr Wright’s examination and testing of the defendant. Mr Wright was well aware of the research which has been done upon suggestibility and compliance, in particular by Professor Gudjonsson. Among the several different tests administered were the Gudjonsson scales for each of these two features. They are different concepts, although both may arise in relation to the issue of whether a false confession has been made. We borrow gratefully Professor Gudjonsson’s definitive analysis. Suggestibility is the tendency to believe what one is told and thus to assert, or accept, it. Compliance is the tendency to accept what is put to one, without believing it but to avoid confrontation or disagreement. Those two tests did not reveal the defendant to be either abnormally suggestible or abnormally compliant. She registered a single score which was out of the normal range, and that on a different test, the Eysenck Personality Questionnaire (EPQ-R). Here she scored abnormally highly for psychoticism. Mr Wright described the significance of this for her personality as follows:
“According to Eysenck a person with an abnormally high score on the Psychoticism scale may be described as being solitary and not caring for people. They are often troublesome and not fitting in anywhere. They may be lacking in feeling and empathy and are regarded by others as insensitive. They are said to be hostile to others, even their own kith and kin, and aggressive even to loved ones. They have a liking for unusual things and a disregard for danger. They like to make fools of people and to upset them. According to Eysenck, socialisation is a concept which is relatively alien to high psychoticism scale scorers. Empathy, feelings of guilt and sensitivity to other people are said to be notions which are strange and unfamiliar to them. Eysenck suggests that the term ‘tough-minded’ is the most appropriate to use in describing high psychoticism scale scorers.”
Overall Mr Wright concluded:
“Personality assessment identifies Julie Kenyon as a very tough-minded individual.”
Mr Wright additionally noted that the defendant presented as angry, irritable and somewhat suspicious. There was, he said, a strong possibility that this might reduce her susceptibility to suggestion and interrogative pressure. He did not, accordingly, wholly rule out the possibility that she might have made a false confession to Carol as a result of internal psychological factors evident at the time but less evident at his examination.[19]Neither expert was called at the trial. This was clearly a deliberate decision made by the defendant on careful legal advice. The CCRC recognises that the decision was a wholly reasonable one. We agree and would go further. Whilst the evidence of Dr Bhattacherjee and Mr Wright, if given, would have left open the possibility of a false confession owing to a susceptibility to suggestibility or compliance, overall their evidence would not have supported it. Neither tendency could be shown to exist. On the contrary, she was predominantly hostile rather than accepting. Neither tendency, if it were assumed nonetheless to exist, could explain the confession to mother. The overall assessment of the defendant’s personality was such as to fit the Crown case that she had indeed killed Grandmother. Moreover, the history, as elicited by both experts and more generally, would have been open for examination and this would have been likely to be distinctly disadvantageous to the defendant because of the number of occasions on which it would show that she had advanced untruthful assertions not by way of confession but by way of exculpation. If the reports had been relied upon, the defendant would have had to submit to further examination by other experts instructed by the Crown, and their conclusions might be anticipated to emphasise the indications which contradicted suggestibility or compliance. The defendant was undoubtedly better served by asserting that the confessions were false for the two different reasons she gave, but without expert analysis.

The new psychological and psychiatric evidence

[20]Professor Gudjonsson and Dr Johns have undertaken the most detailed further examinations of the defendant. Professor Gudjonsson has additionally drawn on interviews with others who know her. Dr Johns has examined her GP records more fully than was previously done. The gravamen of the evidence now advanced on her behalf is twofold: i) the defendant’s personality is disordered; this is a permanent characteristic and by itself tends to suggest the real possibility that she might be prone to making false confessions; this approach to her analysis was not sufficiently explored at the time of the trial; and ii) the assessment of suggestibility and/or compliance is now significantly different from that returned by Mr Wright’s tests at the time and indicates an elevated tendency to make false confessions.[21]We accept that the analysis of the defendant’s personality disorder is significantly fuller than was made at the time of trial. The diagnosis is of emotionally unstable personality disorder with perhaps additional features of schizoid disorder. That has been referred back to Dr Bhattacherjee and Mr Wright, who do not disagree and nor does Professor Grubin. All agree that a personality disorder is essentially a persisting condition and, absent trauma or similar, is unlikely to have changed over a few years in adulthood, so that her personality was basically at the time of these events what it is now. Dr Johns identified for us some of the background social factors: an abusive childhood with a collusive mother, inconsistent parenting, consequent difficulty in establishing relationships, losses and estrangements in adulthood, drink and drugs dependency and low self esteem. We think that we should accept as a general proposition that personality disorders of this kind increase the statistical risk that the subject may make false confessions. The exact mechanisms by which that may occur are perhaps not crucial, but we accept that they might include a desire for self-importance or notoriety, a wish to wound, or an aim to tell people what they want to hear. We also accept that there is some evidence of the defendant, as a child thirty years ago, taking the blame for misdeeds which younger siblings had committed.[22]That said, the defendant’s personality disorder does not necessarily mean that she will make false confessions. Her personality has to be considered in its entirety. Both Dr Johns and Professor Gudjonsson agreed that the description of it which was set out by Mr Wright, and which we have quoted at paragraph 18 above is accurate, save that Dr Johns says that ‘impaired empathy’ is a better description than ‘tough-minded’. If giving evidence at any trial of the defendant each would tell the jury that those are essentially features of her personality. The defendant’s assessments in prison are wholly consistent with these personality traits. She has a history of bullying and is described in the psychological assessments as ‘over-assertive’. At her annual review in August 2005 she was particular in detailed refutation of a minor detail in what had been recorded as her history and she described herself as having ‘this pent up hate’. As Professor Gudjonsson told us, the same personality disorder which gives rise to an increased risk of false confessions also produces an increased risk of aggression, of callousness and of violence, all of which make the defendant vulnerable to a sudden loss of temper in the course of which she may have smothered her Grandmother. It is also characterised by a much increased risk of lying in her own interest. The experts made it clear that they would have to agree that whilst there was a greater risk of false confession than there would be in an undamaged personality, an alternative explanation of events was that the confessions had been true and the defendant was now lying in an effort to avoid the consequences. Sadly, the tendency to lie to protect herself was only too clearly revealed by the evidence in the case, not least by her determined assertions to the police that she had not made the confessions to Carol and Mother until the contrary was conclusively proved by the production of the tape. A similar tendency to lie may well have been inferred by the jury from the defendant’s evidence about the removal of a ring from Grandmother’s finger, which she asserted had been open and above board either because a policeman had done it or because the neighbour Mrs Ledingham had. The two assertions were mutually contradictory and the latter, on which the defendant relied at the trial, was refuted by Mrs Ledingham, whom the defendant in consequence accused of lying, although it is difficult to see what possible motive she could have for denying it if it had occurred.[23]We accept that the measurements made by Professor Gudjonsson of suggestibility and compliance are higher than those made by Mr Wright. Whether the passage of time has anything to do with that we cannot know, but we think we ought to assume that this may not be so and that it is possible that Mr Wright’s tests were affected by the anger she was then manifesting. We have some misgivings about the technique of interviewing other people as closely connected with the facts of the case as Carol, Shirley and a third lady, Janice Jaggers, with a view to assessing either tendency. We readily understand that Professor Gudjonsson’s object was to avoid being wholly dependent on the defendant’s self-reporting, but the difficulty is that none of the informants was really a disinterested observer. Shirley, who is a strong supporter of the defendant, gave an assessment of suggestibility and compliance which is wholly contradictory of that of Carol, whose stance is likely to be very different. Moreover, Professor Gudjonsson unsurprisingly attached a good deal of significance to Shirley’s report to him of what she now says the defendant told her of her conversation with Carol immediately after the confession had been made, namely that she had ‘told them what they wanted to hear; I’ve told them I killed my Nana’. There is some difficulty about that assertion of Shirley’s, however, because in evidence at trial she said very clearly that her only knowledge of any accusation made by Carol (or anyone else) to the defendant was that a landlord had told her (Shirley) that Carol had been ejected from the public house for making such. If she had been able to say that the defendant had come away from her conversation with Carol saying what she now says, it is difficult to see how she could have given that evidence. Similarly, Professor Gudjonsson’s assessment of suggestibility is founded in substantial part (though not exclusively) on Shirley’s recent assertion to him that the defendant had telephoned her as many as 5-6 times per week in distress at harassing accusations from the family and had said to her “Well, perhaps I have killed her, but how did I do it ?” Of course if this were reliable evidence, it would point towards suggestibility, towards the defendant beginning to believe what she was accused of. But this evidence is flatly contradicted by what Shirley said at the trial. There she was asked directly, in chief on behalf of the defendant, whether she had ever asked the defendant if there was any truth in the rumours and she answered that she had, and that the defendant had looked her straight in the face and said no. The truth is that, apart from this highly unreliable recent assertion from Shirley, there is not the slightest sign anywhere in the enormous volume of material assembled either now or at trial of the defendant coming anywhere near falsely believing that she had killed Grandmother. All the indications are against the operation of any element of suggestibility, whatever may be her theoretical susceptibility to that process. Mr Dunkels QC understandably put her case substantially upon compliance, rather than upon suggestibility.[24]As to compliance, the defendant faces equal difficulty. On her case she had been hounded for five years before she succumbed to making any confession save that made to mother, which was for revenge not the result of any accusation. If that is so, she had been resistant to constant accusations for years. We recognise the very real possibility that her account of the accusations is exaggerated, and we are sure that we should accept the proposition that to say that a person exhibits compliance does not mean that she will cave in at the first opportunity. But this defendant’s compliance was put to an acid test, and on this very topic, by her police interviews. Unsupported by either solicitor (until the last session) or appropriate adult, she was interviewed for several hours. She was confronted with the statements of her accusers, including mother, Donegan and Carol. The suggestion that she had killed Grandmother was distinctly put. She stoutly denied it and refuted what others were saying with some spirit. No-one reading the interviews could mistake her assertiveness on this topic. The delicate suggestion of mercy killing would, for a compliant person, have offered a simple and excusable way out. Even that did not occasion any hint of admission. We agree that it occasioned a shift, but the only shift was to an exculpatory account that Grandmother had been wanting to put an end to her life and, by implication, that she must have done it unaided when the defendant refused to help. That exculpatory account, in effect that it must have been self-suffocation, was directly and compellingly refuted by the pathological evidence served in advance of the trial. It was abandoned at the trial by the defendant who admitted that the story of Grandmother holding her hand and asking to be helped out of her misery was a deliberate lie, and who reverted to the account that she had gone to bed unsuspecting of anything amiss and had been awakened in the night by the calling of her name. In her evidence the defendant was asked why she had told these lies, but she had no explanation at all to offer. This lie is likely to have been a significant factor in her conviction; despite the interviewer’s suggestion of mercy killing it was not a lie by way of false admission, nor any kind of compliance, but a lie by way of untruthful self-exculpation.[25]Mr Dunkels contended that the new expert evidence provides a basis on which the confessions to both Carol and Mother could be excluded pursuant to either section 76 or section 78 of the Police and Criminal Evidence Act 1984. We agree that a case could be mounted for the exclusion of the confession to Carol on the basis that it may have been obtained in consequence of something said which might make it unreliable, viz if the defendant were correct, a long period of harassing accusations. We do not agree that the confession to Mother would be likely to be excluded because the defendant’s admission that it was made appeared in the taped conversation with Carol. The topic was raised entirely by the defendant. There was no hint of any suggestion to her that she had confessed to Mother; she volunteered it. In those circumstances the admission to the confession was not arguably the result of anything said which was likely to render it unreliable. Indeed, in the face of clear evidence of confessions to Mother and Donegan we do not think that there would have been a proper basis for excluding the evidence of the confession to Carol. As to section 78, the suggestion is that there might have been a basis for asserting a breach of Code C (in particular note 1G and paragraph C 3.15) in not introducing an appropriate adult. That would involve saying that the defendant’s mental vulnerability was such that she might not understand the significance of what was said. Knowledge that she was prone to depression and was taking medication by no means necessarily carries that implication. She had vigorously declined two separate offers of a solicitor. However, even if there were a case for alleging an inadvertent infringement of the Code, the question which would arise would be the fairness to the proceedings of admission of the interviews in evidence. In the circumstances of this case we have no doubt that if compliance were the issue, the fairness of the proceedings generally required the jury to be in a position to judge whether the defendant had shown compliance or resistance in the course of quite long interviews, the first three without attendance and the last with a solicitor.[26]No doubt it is true that there remains some evidence which could support a case mounted for the defendant on the basis of a personality which might increase the risk of a false confession. But it is plain that the exploration of this suggestion would entail opening up for the jury’s consideration several matters which were only likely to damage her case. We have little doubt that that price was recognised by those advising her at the time of her trial and that that was part of the reason for her decision not to proceed on the basis of expert analysis of personality. Had there been called the evidence which is now suggested, and equally if there were now another trial at which it were called, the following consequences, at least, would ensue. i) Professor Gudjonsson and Dr Johns would give evidence that her overall personality traits, whilst indicating the risk of false confession, are also consistent with both guilt and lying to avoid it. The description of her personality would be much more likely to reinforce the conclusion that she is guilty than to lead to the opposite result. ii) The jury would have to know, which the trial jury did not, that not only had she denied the confession to Carol and subsequently admitted it, but that the same was true of the confession to Mother. iii) The defendant’s whole history would be open for examination in order to assess the likelihood of false confession owing to personality, or the alternative pattern of behaviour involving offending (often aggressive and violent) followed by lies to exculpate herself. It is clear that whatever may be the understandable social deprivations which made her what she is, the preponderance of evidential material demonstrates her to be both aggressive and untruthful in her own interests rather than prone to make false confessions. iv) That history would be very likely to include the introduction of the fact that her first child, a girl, had died at the age of 5 months in 1977 and her account of it. That was that she had left the child face up on the settee and then had returned to the room to find that she had turned over and was lying on her plastic-backed bib. According to the defendant she had turned the child over, whereupon she had breathed twice but then expired. This was excluded from the evidence at trial at the defence request, although it was potentially relevant at two points. First, Donegan had said that in the course of his conversation with her the defendant had told him that she had gone to prison over the death of the child, whereas she had not. Second, the defendant had concluded her taped conversation with Carol by saying that although she had killed Grandmother, she had not killed the baby. The first would be relevant to the defence contention that she made an incorrect statement potentially damaging to herself, although not an admission of guilt. The second would be relevant to the Crown’s contention that she was able to decide between an admission she was prepared to make and one she was not. If close analysis of her personality were in issue, this aspect of the case could not be shut out. The defendant would be likely to give an account of the event, and thus the evidence about it would be open for examination. Quite apart from the similarity of the defendant coming upon someone who had suffocated and the fact that her account was, at the least, open to question, there was also evidence of inconsistent assertions by her as to whether she was alone or accompanied at the time. There might have been a risk of it also emerging that this had occurred at a time when the defendant and her then partner (not the child’s father) were, on his account, at odds because he suspected her of infidelity, and when she had recently told him that if he did not return to her she would harm herself. v) The defendant’s history would also be likely to open up for examination an incident in 1985 when she was convicted of arson and did go to prison for 9 months. Although at the time she admitted to the probation officer that she had started a fire in order to claim benefit money for a replacement carpet or furniture, she has also said that her then five year old son David started the fire and, to Dr Bhattacherjee and Mr Wright, that she admitted it because the police told her that if she did not, he would be taken into care. That might superficially look like support for a tendency to make false confessions, but it would expose her to the evidence that David had already been taken into care on twelve occasions and that the social workers were very worried that he appeared to have had a number of suspected non-accidental injuries and to be treated badly by his mother, described as ‘volatile’ and ‘explosive’. Meanwhile the defendant’s account to Dr Johns, whilst it attributed the fire to the boy, involved a different explanation for her admission, namely that she had hated being in the police station with men. vi) There would be a number of other examples of inconsistent and apparently deliberately untruthful statements made over the years by the defendant when she was accused of wrongdoing. They would have been likely to include an incident of theft of a purse where the defendant has given the wholly inconsistent accounts that it was mistaken identity and that she only took back what was hers and an assault on a sister (Lynne) where she has said to different people that it was someone else and that it was herself in self defence. Those are only some of a number of illustrations of her tendency not to make false confessions but to make false denials of guilt.[27]We are quite satisfied that the new expert evidence, whilst it adds detail and some fresh assessment to the expert evidence available at trial, would in the end not assist the defendant in contesting this charge and that it would not have been in her best interests to adduce it, nor would it be now. In those circumstances it is not either necessary or expedient in the interests of justice to admit the evidence. We reach that conclusion without it being necessary to bring into the equation the important factor, emphasised in Erskine (see paragraphs 39 and 82) and in many other cases, that the interests of justice lie in there being a single trial at which the defendant and the Crown each presents the whole of its case. It is apt to subvert the process of justice if it is open to a defendant to rely on appeal on something which could have been relied upon at trial, but which he chose not to adduce. In Erskine the court was concerned with the defendant who seeks to advance a new and inconsistent defence on appeal (diminished responsibility when previously he had denied carrying out the killing). A fortiori the same applies, and potentially with added force, to the case of the defendant who seeks to advance the same defence as was advanced at trial but relying on the expansion of a strand of evidence which was deliberately not adduced there. We pause only to record that we do not accept Mr Dunkels’ submission that it is enough that the balance of the decision whether to rely on it or not has shifted in the light of differently expressed or improved evidence in support of the relevant strand. There has to be fresh evidence which is strong enough to justify the conclusion that it is necessary or expedient in the interests of justice to override the interests which justice has in a single trial. In this case it is abundantly clear that the strand of evidence was not adduced because it would be likely to do the defendant more harm than good, and still would. The evidence now proffered would not be likely to afford grounds for quashing the conviction. Accordingly we decline to admit it and the appeal must be dismissed.