The Law Commission

 


PART XV THE QUALITY OF TENDERED EVIDENCE

INTRODUCTION

15.1      In this Part we consider what approach the court should take to the quality of the evidence which a party wishes to adduce. Should the court make some assumption about the veracity or reliability of the witnesses and the overall quality of the evidence, or investigate for itself? We conclude that in considering admissibility, the court should not make its own enquiry, but should take the evidence at face value (unless it is unrealistic to do so).

15.2      The problems posed by this issue arise most acutely where it is alleged that the quality of the evidence is affected by contamination or collusion. We therefore consider additionally what a court should do where bad character evidence has been admitted but turns out to be so affected. We conclude that, in the Crown Court, the judge should be under a duty to discharge the jury where a conviction would be unsafe because of the doubt cast on the quality of the evidence by contamination and/or collusion.

WHAT ARE CONTAMINATION AND COLLUSION?

15.3      In the leading case, Lord Mustill explored the meaning which "collusion" had been given in earlier authorities, and differentiated the true meaning of collusion from contamination:

…, it is important to note the ambiguity of the word "collusion". In its more limited sense this may denote a wicked conspiracy in which the complainants put their heads together to tell lies about the defendant, making up things which never happened. It is however clear that the argument for the appellant, and the authorities on which it is based, give the word a much wider meaning; wide enough to embrace any communications between witnesses, even without malign intent, which may lead to the transfer of recollections between them, and hence to an unconscious elision of the differences between the stories which each would independently have told. …

For convenience, the two situations may be labelled "conspiracy" and "innocent infection".^[1]

15.4      The two situations thus described are the extremes of the spectrum: deliberate fabrication of allegations resulting from an agreement between two or more witnesses, and unconscious alteration of evidence, resulting from having become aware of what the evidence of another will be or has been. There are variations in between. We attempt below to illustrate the range within the spectrum:

(1) A and B have together decided to make false allegations against D.

(2) C, having become aware that E has made allegations against D, goes to the police (or responds to a trawl by investigators) to make false allegations on her own initiative.

(3) F and G make an agreement both to say that D was wearing a dark blue top. They may be pooling their recollection and honestly agreeing what their recollection must have been. Neither of them may be sure what colour he was wearing but they agree to say the same thing so as to give the impression that they are sure. The agreement may be the result of negotiation in which F is persuaded by G, or agrees without really being persuaded, to adopt G's recollection.

(4) H learns that J has said D's top was blue, and decides to change or add to her evidence to the same effect, without colluding with J at all. She may be telling a deliberate lie, or have been unconfident about what colour D's top was and so, on hearing that J thinks it was dark blue, resolves that she will also say it was dark blue, and believes she was mistaken in thinking it might have been green. In this scenario, the alteration is conscious, but there is no collusion in the sense of conspiracy.

(5) K is unconsciously influenced by L and changes her recollection and account so that she too says D was wearing a dark blue top, without being aware that she has changed it. That is Lord Mustill's "innocent infection".

15.5      Situations (1) and (3) involve collusion (in the first, complete concoction, and in the third, collusion as to an element of the evidence) and (1) and (2) involve concoction. Situations (2), (4) and (5) may all be described as "contamination", but only (5) is "innocent infection".

15.6      It is mistaken to correlate collusion as in (3) with false allegations, and contamination (as in (4) or (5)) with an essentially truthful story. F and G may be essentially right that D did what they allege, even though they have agreed to gild the lily to make it sound better. Conversely, K's central allegation, which chimes with L's through innocent infection, may be a complete (though unwitting) fabrication.

15.7      A jury or bench of magistrates could allow that there had been collusion (in the sense of (3) or contamination (as in (4) or (5)) and yet properly convict. They could not, though, accept the falsity of concoction (1) and properly convict. In (2) they could only properly convict of E's allegations, not C's.

THE QUALITY OF EVIDENCE AND ADMISSIBILITY

15.8      Contamination and collusion have the potential to affect decisions at two stages in the trial: at the time the evidence is ruled admissible or inadmissible, and after the evidence has been given. We now consider what view the court should take of the quality of bad character evidence at the time of ruling on admissibility.

The current law

15.9      The question how a court should approach such evidence was considered in H.^[2] The defendant appealed against his convictions for sexual offences against his adopted daughter and step daughter. There were similarities in the two complainants' accounts but each denied that they had collaborated or concocted a story. The House of Lords held that, in relation to "similar fact evidence" the assessment of credibility should be left to the jury: for the purpose of deciding whether the evidence is admissible, the trial judge should not make any enquiry into the quality of the evidence. Thus the issues of collusion and contamination are not relevant considerations for the judge on the question of admissibility.

The options in the consultation paper and the response on consultation

15.10      In the consultation paper we looked at three possibilities:^[3] replicating the law in H (option A); requiring the judge to assess the cogency of the evidence on the basis of the documents (option B); and requiring the judge to hold a voir dire to make an assessment of the quality of the evidence before ruling on admissibility (option C). We provisionally rejected option B, but expressed no view as between options A and C. We asked respondents' whether "juries can now be entrusted with the task of evaluating for themselves evidence which exhibits appreciable risks of contamination and collusion".^[4]

15.11      Forty respondents addressed this issue. Twenty agreed in unequivocal terms with option A. Three agreed with it but stated that in exceptional cases option C ought to be available to the judge. (This is in fact the position under H, because their Lordships left open the possibility of a voir dire, although they did not give examples of when it might be appropriate.) Two expressed a preference for option B. Two suggested the use of option B as a precursor to option C. Nine preferred option C, and the remainder did not favour any of the three options.

15.12      The main argument in support of option A was that allegations of collusion and contamination essentially raise issues of credibility and ought to be left to the arbiters of fact. As one respondent wrote, "If we cannot expect juries to deal with cases in which two witnesses may have fabricated their stories together, or in which one may be feeding off or contaminating the other, we should abandon hope and do away with trial by jury."

15.13      The Law Reform Committee, General Council of the Bar thought option B merited further thought and that where there was cogent evidence of collusion and/or contamination, the judge ought to be able to express a view at the outset. We have no doubt a judge would do so where the contamination was glaring from the papers alone. Such a case would be rare, as Lord Mustill said in H, "I find it hard to envisage that where the committal papers are so frank or artless that they disclose on their face a real risk of fabrication the matter will ever be brought to trial." ^[5]

15.14      Those who supported option C appreciated that it had its failings, but nevertheless thought it preferable. As Professor Tapper wrote,

… even granting the valid objections to it, the least bad option is for a voire dire; to adopt any other solution must run a very severe risk of prejudice. The whole rationale of these proposals is that some evidence is too dangerous to admit; it must follow that these very dangers demand that jurors not be exposed to such evidence upon the simple decision of the prosecution, without some attempt to filter out the worst examples. It is not feasible to instruct them to eschew prejudice once the evidence has been led.

Those who criticised option C tended to do so out of hostility to voir dires, and a fear that it would be close to impossible to demonstrate no risk of collusion.

The policy considerations

15.15      The heart of the problem is the fact that the question of admissibility turns on the weight to be given to the evidence. In this context credibility is a crucial factor in deciding the probative force (and hence one limb of admissibility) of the evidence. In many cases where this problem arises, for example cases of alleged sexual misconduct, the evidence in question is likely to be very prejudicial.^[6] In these cases, if the judge does not pre-test the quality of the evidence, highly prejudicial evidence of dubious quality is likely to be admitted.^[7]

15.16      It appeared at the time of the consultation paper that the choice was between following Hoch, (option C, the Australian approach)^[8] and H (option A). The essence of the former is encapsulated by Lord Mustill:

The logic of the exception which permits similar fact evidence depends on the unlikelihood of coincidences between truly independent events. In some cases, the happening of the other events is undisputed, and the prosecution invites the inference that it was the defendant who was responsible for them all. In other instances the happening of all the events, including the one charged, is admitted, but the defendant maintains that they were accidental, to which the prosecution replies that whilst one accident is a possibility, it is unlikely that the same kind of event was brought [about] by the same kind of accident on so many separate occasions. In yet other cases … all the alleged events are in dispute, and the prosecution relies on the improbability that different complainants will have spontaneously invented broadly similar accusations against the same person. In these and other situations the justification given for admitting evidence of the other occasions is that the jury should be presented with the whole picture, so that a choice can be made between an inference of guilt and an acceptance of an implausible coincidence.

Founding on this rationalisation the appellant maintains that in the third type of case the stark choice between a coincidental independent fabrication by different complainants and a finding that the complaints are true disappears once there is a possibility that the complaints may not be spontaneous and independent but fabricated in concert; and the disappearance of this choice carries away with it the logic of making an exception to the general rule.

It follows, so the argument concludes, that the prosecution, having the burden of showing that evidence which would generally be excluded is in the particular case admissible, must establish that there is no reasonably possible explanation of the similarity of the events other than guilt or coincidence. This means that the exclusion of conspiracy as a cause is a condition precedent to the admission of the evidence, and since it is the function of the judge to rule on whether the condition is satisfied he must investigate the facts, if necessary by hearing evidence on a voire dire in the absence of the jury, and reach a conclusion upon it, just as he does in other instances where the satisfaction of a factual condition precedent is in dispute.^[9]

15.17      Lord Mustill decided that this conclusion cannot be accepted. The judge would, in that case, have to decide the very question that the jury has to determine:

There are many reasons why witnesses may not be believed, and conspiracy, or the suspicion of it, is only one of them; and it may of course be alleged in cases which have nothing to do with evidence of similar facts. Credibility is always for the jury, and I can see no reason why the special feature that the testimony is adduced to support a charge concerning acts said to have been done to the person with whom the witness is suspected of conspiring, rather than to the witness herself or himself, should affect this fundamental principle in any way.^[10]

15.18      The argument in favour of the judge testing the reliability of the evidence before admitting it may also be put in this way: if probative value must transcend prejudicial effect, the court must not assume the truth of the evidence tendered, because to do so means that the probative value is assumed not assessed. It appeared that the Australian courts were interpreting Pfennig^[11] and Hoch to this effect, requiring the judge to pre-weigh the evidence, by a voir dire if necessary.

15.19      There are two possible answers to this way of putting it. First, assuming the truth of evidence tendered is not the same as assuming its level of probative value. Probative value is a combination of reliability and degree of relevance. The latter has to be assessed, as well as some view being taken as to its reliability.^[12] Secondly, the Australian courts have stopped short of interpreting Hoch as meaning that in every such case the judge has to exclude the possibility of collusion before proceeding to admit such evidence. Whereas it was once thought that the prosecution had to prove the non-collusive quality of the similar fact evidence, it then became clear that it was sufficient for the prosecution to disprove a real danger of concoction. It now appears that showing a real opportunity for communication between witnesses will not, without more, suffice to conjure up a possibility of concoction sufficiently real to amount to a reasonable explanation for the concert of allegations.^[13] It is also noteworthy that the issue for the Australian courts is the risk of "concoction",^[14] and does not extend to the risk of innocent contamination.

15.20      Lord Mustill's objections, however, remain valid even under this less rigorous interpretation of Hoch. Even if the court enquires only so far as to discount concoction as a reasonable explanation, it is still engaged in the very task the jury is there to perform. In any case where the complainants knew each other, or mixed together, the court would often have to assess the truthfulness of the complainants.

15.21      The practical consequences of the different policy options cannot be overlooked. First, if a voir dire were held, advocates and witnesses would learn from the voir dire, and ask questions or give answers differently or more fully, so it is not certain that the evidence on the voir dire would be the same as that given at trial.

15.22      Secondly, it is undesirable to require witnesses to testify more than once where that can be avoided, especially where their evidence may be of distressing personal matters. As Lord Mustill held in H, if the policy is that the prosecution is required to disprove collusion or contamination in any case in which more than one instance of sexual misconduct is relied on but there is the possibility of collusion or contamination the prosecution is required to disprove collusion or contamination, then that will necessitate a voir dire in a large number of cases, with all the delays and the possible collapse of such cases which may follow:

… the judge will be required in almost every case to hold a preliminary trial … solely in order to see whether the prosecution have proved a negative which will often be unprovable. This is more than I am willing to accept. The possibility of innocent infection is one amongst many factors which the jury will have to take into account; but to treat it as a unique "threshold issue" loads the scales unfairly against the prosecution, and hence against the interests of those who cannot protect themselves.^[15]

We agree with this view.

15.23      We are, therefore, persuaded to favour option A. Issues of contamination and collusion ought to be decided by the jury. Of course, it might be the case that the evidence in the depositions would not, even if taken at their highest, prove the facts it is adduced to prove, in which case the court would not admit the evidence. Thus, if a judge can determine on the papers that there has been concoction, then the evidence should not be admitted. We would hope that this would be a rare occurrence. Where the court cannot make a determination on the papers it has an inherent power to hold a voir dire if it finds it necessary.^[16]

15.24      We are conscious that protection against evidence which is revealed as obviously tainted must be afforded to the defendant. The purpose of the exclusionary rule is to protect the defendant. Thus, if the defendant is not going to receive the benefit of a pre-testing of the veracity of the witnesses by the court in this situation, we need to be confident that the other procedures in place to ensure a fair trial are adequate. We believe that the judge should have a duty to keep the issue of the quality of the evidence under review. If, in the course of the trial, the evidence turns out to have low probative value, because of collusion or contamination, then the judge will have the power to withdraw the case from the jury.^[17]

What approach is a court to adopt to bad character evidence generally?

15.25      The quality of evidence tendered which goes to credibility or to correct a false impression^[18] is significant, not as a pre-admission test for veracity, but on the question how far evidence on a side issue is going to distract from the central issues. The quality of the evidence might fall to be assessed by the court having regard to that factor, but not as a pre-condition of admission.

The recommendation

15.26      We recommend that, where the court is required to assess the probative value of evidence of a person's bad character, it should be required to do so on the assumption that the evidence is true, except where it appears, on the basis of any material before the court, that no court or jury could reasonably find it to be true.

15.27      This recommendation is given effect by clause 14 in the Bill.

A DUTY TO WITHDRAW THE CASE FROM THE JURY

15.28      The House of Lords in H stated that if the question of collusion has been raised during the proceedings, the judge is obliged to draw the jury's attention to its importance, with a warning that, if they were not satisfied that the evidence could be relied on as free of collusion, they could not properly rely upon it for any purpose adverse to the defendant.^[19] Their Lordships went on to consider the case where similar fact evidence is admitted but is later shown to be such that no reasonable jury could accept the evidence as free from collusion. In that case, they believed that the jury should be given a direction that the evidence could no longer be relied on for any purpose adverse to the defence.

15.29      This raises a stark issue: whether jurors are likely to be able to put out of their minds discredited evidence which has already been admitted and which may be extremely prejudicial. We are concerned that when such evidence has been correctly admitted, but it later transpires that the evidence has been discredited by contamination or collusion, the warnings referred to in H may not be effectively followed. The jury will have been seriously prejudiced.

15.30      In the consultation paper^[20] we suggested that a better course might be to require the judge, if so requested, to withdraw the matter from the jury if he or she was satisfied, after hearing all the evidence, that due to the contaminated nature of the evidence given, the prejudicial effect of the bad character evidence outweighs its probative value.

15.31      In making this suggestion, we noted the existence of a similar safeguard against unreliable identification evidence.^[21] A judge is obliged to withhold a case from the jury if the prosecution is based largely or entirely on identification evidence and the judge concludes that there is insufficient evidence on which a jury could properly convict.^[22] The same principle could be said to apply to certain types of similar fact evidence.

Responses on consultation

15.32      Thirty-two respondents addressed this issue. Of these, eleven respondents unequivocally agreed with the proposal, and a further nine agreed with reservations. Eight disagreed, two because they thought the current law already covers the point. Two respondents were confident that a judicial direction would be sufficient to prevent injustice.

15.33      Trial judges have the discretion to direct an acquittal where discredited confession evidence has been admitted, so as to ensure that the defendant has a fair trial.^[23] Recent cases have gone further, expressly stating that the judge may direct an acquittal at any time from the close of the prosecution case to the end of the trial, on the basis that there is no case to answer. This discretion should be used rarely, but it can be used for any reason, so presumably it could be exercised if the judge was concerned that some form of prejudicial evidence should not, after all, have been admitted.^[24] However, this recent development runs counter to the traditional view as to the limits of a judge's power.^[25] Our proposal would oblige the judge to direct an acquittal where discredited prejudicial evidence has been admitted, thereby making explicit what may previously have been unclear and discretionary.

15.34      Professor Tapper agreed with the proposal but felt that there may be "a danger that the judge would not sufficiently address his mind to the issue, and might be reluctant to reverse his previous considered view". We believe that the contrary is the case: a clear statutory duty will make it more, rather than less, likely that a judge would be asked to address the issue. Moreover, the judge would not be asked "to reverse his previous considered view": the evidence would have been admitted without enquiry into its quality. Once the evidence has been heard and tested, the judge need not deal in assumptions about it, but can evaluate it afresh.

15.35      A different respondent thought the issue is one for the jury, not the trial judge. We disagree. The fact that the judge must in our scheme take the reliability of the evidence at face value necessitates the introduction of this safeguard for the defendant. There already exist examples of the judge having to assess the quality of evidence placed before the jury, for example, the second limb of Galbraith,^[26] in the case of identification evidence and, in the case of hearsay evidence, recommended by this Commission^[27] (and accepted by the Government). As Lord Mustill has said, the risk that the jury may act upon evidence which is not worthy to be relied upon "may well be seen as serious enough to outweigh the general principle that the functions of the judge and jury must be kept apart".^[28]

15.36      A further respondent thought it should be left to the Court of Appeal to quash an unsafe conviction, and not for the trial judge to prevent the conviction in the first place. We do not agree. A judge has a duty to achieve a fair trial for the defendant. It is now recognised that there is a high level of coincidence, (if not total identity) between a trial which is not fair and a verdict of guilty which is "unsafe".^[29] If evidence has been admitted which is highly prejudicial, and a direction from the judge may not adequately guard against prejudice, then protection for the accused from an unsafe conviction requires intervention to withdraw the case from that jury.

The recommendation

15.37      We recommend that, in a trial on indictment, where evidence of the defendant's bad character has been admitted with leave, and the judge is satisfied that the evidence is contaminated such that, considering the importance of the evidence to the case against the defendant, a conviction would be unsafe, the judge should be required to discharge the jury or direct the jury to acquit.^[30]

15.38      We make no equivalent recommendation for the magistrates' court: if the magistrates think either that a conviction would be unsafe, or merely that the incriminating evidence has been so weakened by evidence of collusion or contamination that they cannot be sure the accused committed the offence, they will acquit.

Ý
Ü   Þ

Note 1   H [1995] 2 AC 596, 616.    [Back]

Note 2   [1995] 2 AC 596.    [Back]

Note 3   Paras 10.94 – 10.97.    [Back]

Note 4   Para 10.97.    [Back]

Note 5   [1995] 2 AC 596, 616.    [Back]

Note 6   The Jury Study concluded that hearing that a defendant had a previous conviction for indecent assault on a child was likely to colour the mock jurors’ whole perception of the defendant. See paras D42 – D44 of Appendix D to the consultation paper.    [Back]

Note 7   See C Tapper, “The Erosion ofBoardman v DPP” (1995) 145 NLJ 1223.    [Back]

Note 8   (1988) 81 ALR 225.    [Back]

Note 9   [1995] 2 AC 596, 619.    [Back]

Note 10   [1995] 2 AC 596, 620.     [Back]

Note 11   (1995) 127 ALR 99.    [Back]

Note 12   Musquera [1999] Crim LR 857 illustrates this point. The appellate court was of the view that the evidence in question was inadmissible on the other count(s), not because of the risk of collusion but because of the lack of relevance. Sexual offences were alleged against D by three complainants. Two of the complainants’ allegations were of comparatively minor misconduct, while the third complainant alleged she had been raped (later amended to attempted rape). Beyond the fact that they were all allegations of sexual misconduct and that the families knew each other because they lived in the same building, the Court of Appeal could find no features which made one set of allegations relevant to another, and they held that those features were not enough to make the evidence cross-admissible.    [Back]

Note 13   See Colby [1999] NSWCCA 261, para 111 which draws on Robertson (1997) 91 A Crim R 388, and see OGD (No 2) [2000] NSWCCA 404.    [Back]

Note 14   Meaning conspiracy to make false allegations, and extending to concoction by one complainant of his or her account “after becoming aware of some like event or some like allegation concerning the accused”: BRS v The Queen (1997) 191 CLR 275, 230, per Gaudron J.    [Back]

Note 15   H [1995] 2 AC 596, 617.    [Back]

Note 16   As confirmed in H [1995] 2 AC 596, 612C, per Lord Mackay.    [Back]

Note 17   See para 15.37 below.    [Back]

Note 18   See clauses 9 and 10 and Parts XII and XIII respectively.    [Back]

Note 19   [1995] 2 AC 596, 612.    [Back]

Note 20   At para 10.105.    [Back]

Note 21   Turnbull [1977] QB 224.    [Back]

Note 22   Daley v R [1994] 1 AC 117.    [Back]

Note 23   In the case of Watson [1980] 1 WLR 991, the Court of Appeal accepted that the judge retains control over the evidence that should be submitted to the jury throughout the trial. If a confession is admitted, but the judge later decides that it should not have been, the judge is empowered to direct the jury to disregard it, direct an acquittal or direct a new trial (at p 995). This case was followed in Sat-Bhambra (1989) 88 Cr App R 55.    [Back]

Note 24   Boakye (Court of Appeal, 12 March 1992, unreported); Brown (Jamie) [1998] Crim LR 196;Anderson, The Independent, 13 July 1998; Brown (Davina) [2001] Crim LR 675.    [Back]

Note 25   See Archbold para 4–292.    [Back]

Note 26   [1981] 1 WLR 1039: where the judge is required to consider the prosecution case, taken at its highest, the judge must make some kind of assessment of the quality of the evidence.    [Back]

Note 27   Evidence in Criminal Proceedings: Hearsay and Related Topics (1997) Law Com No 245, para 11.31, Recommendation 47.    [Back]

Note 28   Daley v R [1994] 1 AC 117, 129D.    [Back]

Note 29   Davis, Johnson and Rowe [2000] Crim LR 1012 was the first case to consider whether the defendants’ convictions were unsafe following the Strasbourg Court’s finding that their trial had been unfair. The Court of Appeal agreed that the trial had been unfair, and went on to conclude that the convictions were unsafe. However, the court took the view that this was not an inevitable conclusion. The nature and the degree of the breach of Article 6 would determine whether the conviction was unsafe. Davis, Johnson and Rowe was considered in the case of Togher [2001] Crim LR 124, but slightly glossed: “We would suggest that, even if there was previously a difference of approach, that since the [HRA 1998] came into force, the circumstances in which there will be room for a different result before this Court and before the ECHR because of unfairness based on the respective tests we employ will be rare indeed ... we consider that if a defendant has been denied a fair trial it will almost be inevitable that the conviction will be regarded as unsafe.” This approach was firmly endorsed by the House of Lords in Forbes [2001] 2 WLR 1, 13. Unfortunately the subsequent case of Williams, The Times 30 March 2001 was decided on the basis of Davis, Johnson and Rowe without reference to the more recent cases. Nonetheless, the outcome in Williams is consistent with the Togher-Forbes test, so it is now settled law that “safety” will almost inevitably follow “unfairness”.    [Back]

Note 30   Clause 13 lays a duty on the court to direct an acquittal in a trial on indictment, in specified circumstances. Under ss 6(3) and 7(1) of the Criminal Law Act 1967 and surviving common law provisions (see Saunders [1988] AC 148), a defendant may be convicted of a lesser offence when acquitted of the greater offence. Where the defendant has been found unfit to plead by reason of mental disability, the jury must determine whether the defendant did the act or made the omission under section 4A Criminal Procedure (Insanity) Act 1964. If evidence of the defendant’s bad character is so affected by collusion or contamination that a verdict of a lesser offence or finding that the defendant did the act would be unsafe, then a duty should be laid upon the judge to direct the jury in this circumstance to acquit the defendant. This is provided for by cl 14(3).    [Back]