[Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback]
	The Law Commission
You are here: BAILII >> Databases >> The Law Commission >> Evidence of Bad Character in Criminal Proceedings (Report) [2001] EWLC 273(6) (October 2001) URL: http://www.bailii.org/ew/other/EWLC/2001/273(6).html Cite as: [2001] EWLC 273(6)

[New search] [Help]

 


PART VI THE GENERAL APPROACH: THE OPTIONS

6.1      In the consultation paper we set out six general approaches. They were:

option 1: adduce the defendant's criminal record at the start of every trial;^[1]

option 2: adduce the defendant's record of sex offences in sex cases;^[2]

option 3: allow evidence of the defendant's previous misconduct to be adduced only where it is an ingredient of the offence charged;^[3]

option 4: a single inclusionary rule with an exception for evidence whose likely prejudicial effect outweighs its probative value;^[4]

option 5: an exclusionary rule with a single exception for evidence whose probative value outweighs its likely prejudicial effect;^[5]

option 6: an exclusionary rule with separate exceptions for evidence admissible in chief and for evidence subsequently becoming admissible.^[6]

6.2      Most respondents agreed with our provisional preference for option 6, although not necessarily without reservation. A substantial minority favoured option 1. Some favoured a more limited version of this option, namely the automatic disclosure of offences similar to that charged. This is a kind of halfway house between options 1 and 2, and we now include it as option 1a. There was little support for any of the other options. A few respondents thought that none of the options we identified would be an improvement on the present position.

OPTION 1: ADDUCE THE DEFENDANT'S CRIMINAL RECORD AT THE START OF EVERY TRIAL

6.3      Option 1 had by far the most supporters, after our preferred option, and we therefore reconsider the main arguments for and against this option.

Arguments in favour of option 1

6.4      Arguments put forward for introducing the defendant's record as a matter of course are: that to do so would simplify trials, get rid of the injustices and anomalies created by the present rules, it would be less prejudicial than the revelation of previous offences in cross-examination, that the record is relevant, and that the fact-finders will know anyway that the defendant has one.

Simplicity

6.5      Option 1 appealed to some because of its apparent simplicity. In the consultation paper we said:

A further argument is that, given the tortuous nature of section 1 of the 1898 Act and the uncertainties of the similar fact rules, the course of a trial would be more predictable if a criminal record were put in automatically.^[7]

6.6      We certainly agree with respondents who criticise the current law for its complexity; indeed, we made this criticism ourselves more than once. However, in our view, the apparent simplicity of option 1 would, in practice, turn out to be illusory, because, assuming that only the bare record of offences was initially disclosed, it would still have to be determined whether the fact-finders should hear details of the previous misconduct. It is generally agreed, for example, that such details should sometimes be admitted as part of the prosecution's case. Automatic disclosure of the bare record alone would do nothing to solve the problem of when the details should also be admitted. The only way to ensure simplicity would be by automatically disclosing not just the bare record but the details too; and, given that some or all of the previous convictions may have no relevance whatever to the issues in the case, this would be a waste of time which would puzzle the jury if it did not irretrievably prejudice them against the defendant.

6.7      Even if it were true that option 1 would simplify the law, we would not accept that simplicity should be achieved at the expense of justice; and, for reasons discussed below,^[8] we do not think option 1 would be just.

Removal of injustices and anomalies in the current law

6.8      An attraction of option 1 for some was that it would get rid of some of the perceived injustices in the present law – for example, the sometimes capricious effects of section 1(f) of the 1898 Act. We agree that option 1 would rectify many of these defects in the existing law. It is by no means the only option that would do so. This is, therefore, not a reason per se for preferring it to alternatives which would address many of the same defects.

Avoiding the prejudicial effect of revealing the defendant's record in cross-examination

6.9      More than one respondent referred to the prejudicial effect in a jury trial of the way in which a defendant's record is introduced. If admitted in crossexamination under the 1898 Act, it has more impact than if it were routinely read out at the beginning of the trial, because it is perceived as relevant information which the defendant has tried to keep from the fact-finders. Defence advocates often try to take the sting out of the record by introducing it in the defendant's examination in chief, in a "cards on the table" gesture. Some respondents thought this factor pointed towards option 1, because the dramatic effect would not persist. We agree that the manner of introduction of the record is a factor to consider. We do not agree that this is a good argument for automatic disclosure in cases where, under the present law, the record would not be admissible at all.

The relevance of the record

6.10      A number of respondents cited, as a reason for preferring option 1, the relevance of a criminal record to the question of the defendant's guilt. This argument is, in our view, crucial to the case for option 1. In the consultation paper we tried to assess whether, and if so how, previous offences really are relevant for this purpose.

RELEVANCE TO THE DEFENDANT'S PROPENSITY TO ACT IN THE MANNER ALLEGED

6.11      Clearly, in some cases a defendant's previous convictions may show a propensity to act in the way in which he or she is now alleged to have acted, and may thus help the fact-finders determine whether he or she did act in that way. However, we argued in the consultation paper that

the bare bones of a criminal record do not give adequate information from which fact-finders may reasonably make an assessment of whether the accused is likely to have repeated past behaviour. If the record is admitted automatically, evidence of little or no probative value is admitted.^[9]

This argument was based on our understanding of the psychological research, which we summarised as indicating that

Past behaviour can be probative on the question whether the defendant is likely to have acted in the way alleged, but the probative value of a single previous instance can be easily over-estimated. The research supports the present approach, that past misbehaviour can be admitted where there are close and unusual similarities between the past and the present situations.^[10]

6.12      Nearly all of the respondents who expressly commented on this conclusion agreed with it.^[11] None of those who commented on it – not even those who thought we had underestimated the relevance of previous misconduct – supported option 1. We therefore cannot tell whether the supporters of that option would dispute the factual conclusions suggested by the research, and if so on what basis, or whether they think that evidence of previous offences should be admitted automatically even though we are right in thinking that the relevance of such evidence is usually limited.

6.13      Option 1 involves placing all previous convictions before fact-finders regardless of whether the previous offences had any similarity with those currently charged and of how long ago those offences may have been committed. We should be very surprised indeed were it to be supposed that a single offence of indecent assault, committed twenty years previously could have any relevance to the question whether a person had committed an offence of theft.

RELEVANCE TO THE DEFENDANT'S PROPENSITY TO LIE

6.14      One of the defendant's propensities which may be in issue is his or her propensity to lie. The current law assumes that a person's character as a whole is relevant to his or her general tendency to be truthful in a courtroom. In the light of our interpretation of the psychological research, in the consultation paper we doubted this. It does not appear to be generally true that a person who acts dishonestly in one situation is likely to do so in another. A fortiori, misconduct which does not involve dishonesty is unlikely to indicate a tendency to lie on oath. Conversely, the more similar the circumstances, the more likely it is that the defendant will act in the same way. We therefore provisionally concluded that

(1) the instances of previous misconduct most relevant to credibility are convictions for perjury;

(2) convictions for dishonesty may be relevant in some circumstances;

(3) behaviour not involving dishonesty is unlikely to be relevant to credibility;

but we do not think it appropriate to prescribe in a statute which kinds of conviction are and are not probative.^[12]

6.15      Most of the respondents who commented on this conclusion agreed with it, though some thought that all convictions are relevant to credibility, while others thought that they are rarely relevant in this way. In this connection the example posed above is apposite. If the person had pleaded guilty to the single offence of indecent assault of which he was convicted twenty years previously it would be surprising in the extreme were that bare history to have any relevance whatsoever to the question of his credibility when giving evidence denying a charge of theft now.

6.16      In our view there is more to the issue whether a witness should be believed than his or her general credibility – that is, the extent to which the witness is the sort of person who can in general be trusted to tell the truth. Of far more significance is what in the consultation paper we called the witness's specific credibility – that is, the extent to which, in the circumstances of the individual case, the witness appears to have a reason to lie. Such reasons might include a grudge against the defendant, the hope that the witness will profit from a conviction, or the existence of pressure to give particular testimony.^[13] As a magistrate of 17 years' experience put it, "Any witness is likely to divert from the truth if it is in his/her interest to do so." Previous misconduct may be relevant to a witness's general credibility; but, since little significance can in any event be attached to that credibility or the lack of it, the witness's bad character is generally of little value in determining the issues in the case.

6.17      If we are right in concluding that it is not difficult to envisage circumstances in which a person's previous convictions have no conceivable relevance to his or her guilt, either on the basis of propensity or credibility, then option 1 necessarily involves admitting non relevant evidence, presumably on the basis that, if it has no relevance, then no harm is done anyway. As the research to which we refer below indicates, that is by no means a sound conclusion. The prejudice attached to the fact of conviction, particularly for certain types of offence, is such as to give the lie to the assertion that fact-finders, even experienced ones, can be relied on to disregard irrelevant, though prejudicial, evidence. Thereby the risk is run that the fact-finders will convict where they are not persuaded of the defendant's guilt on the relevant evidence to the requisite standard, because they are influenced by irrelevant evidence which they should have disregarded. The defendant may thus be convicted, not on the basis of what he or she has done, but on the basis of one aspect of his or her character.

The fact-finders know anyway

6.18      A criticism often made of the current rules is that they do not acknowledge the reality, namely that the fact-finders will usually realise that the defendant is of bad character in any event. In the consultation paper we acknowledged that

Any lay magistrate or jury with a week's experience knows that if the court is not told that the defendant is of good character then he or she must have a criminal record; so, the argument runs, it might as well be put in.^[14]

6.19      Several respondents supported this argument. Michael McMullan, formerly the resident judge at Wood Green Crown Court, argued not only that a substantial percentage of jurors will assume the defendant to be of bad character if they are not told otherwise, but that that percentage is likely to increase as knowledge of the legal rules becomes more widespread; and that it would be better to tell jurors the truth than to proceed on the increasingly unrealistic assumption that they will not work it out for themselves.

The present system may actually generate speculation about such convictions or, worst of all, suggest that there is an element of charade about the trial generally and that nods and winks are expected to play some part. A regime of complete openness is not possible in anything so intricate as a jury trial (improperly obtained confessions, for instance, must be excluded) but, if we wish to retain trial by jury, the jurors should not be allowed to become demoralized or cynical. Where possible open procedures are best; the jury should be trusted and feel they are trusted. … [T]here is no point in deciding that the convictions should be kept away from the jury unless an effective and fair means of doing so can be devised.^[15]

One member of the public who had served on a jury wrote: "many of us felt that we were witnessing a 'legal game' rather than a search for 'the truth'."

IS THIS TRUE?

6.20      The first question is whether it is in fact true that the fact-finders know anyway. In the two research studies which she undertook for us,^[16] Dr Sally Lloyd-Bostock found that, if given no information about the defendant's character, a majority of lay fact-finders will tend to assume that he or she has at least one previous conviction; but a substantial minority will assume that he or she has none.^[17] Perhaps surprisingly, there was no significant difference in this respect between the study of simulated juries and that of real lay magistrates.

6.21      A member of this Commission wrote an article for "The Magistrate" inviting readers to let us know whether they assume that a defendant who does not claim to be of good character must have a criminal record.^[18] Nine magistrates responded. They were unanimous that a defendant's bad character was not discussed in the retiring room, but not unanimous as to the assumptions made. Some thought it obvious that such a defendant must have a criminal record; others said the question did not cross their minds; others said they would not make such an assumption where the defendant was unrepresented, or represented by an inexperienced advocate. It is as well that they do not all make this assumption: research cited by Dr Penny Darbyshire shows that a defendant's good character is not always put before the justices, even where the defendant is represented.^[19] We would not accept that if justices are told nothing about the defendant's character they will necessarily assume that he or she must be of bad character, or even that that assumption would necessarily be correct.

6.22      The position is somewhat different in the Crown Court, since the judge will almost certainly have ascertained whether the defendant is of good character and will remind a defence advocate who forgets to adduce evidence of that fact. While Dr Lloyd-Bostock's research suggests that by no means all jurors will infer that the defendant has a record unless they are told otherwise, it seems likely that most juries will include individual jurors who will draw that inference and may voice it; and that, where it is drawn, it will be correct. We therefore agree that, in the Crown Court at least, there is a problem which needs to be addressed. The question is whether this consideration points inexorably to option 1.

IS OPTION 1 THE ONLY SOLUTION?

6.23      We agree that it is unsatisfactory for jurors to be left with neither information on the subject of the defendant's character nor guidance on what to make of the absence of such information. The point at which we part company with the argument is where it is suggested that the jury should therefore be given information on the subject in every case, by disclosing the defendant's record at the start of each trial, rather than guidance on how to deal with its absence. It seems to us that the proposed solution goes much further than is necessary to meet the difficulty. The difficulty, in essence, is that there is information which jurors would be likely to regard as relevant but which is withheld from them, and which they are likely to realise is being withheld from them. The defendant's record is by no means the only information in relation to which this difficulty can arise. Indeed, the Judicial Studies Board recommends that the following direction be given where the judge thinks it may be of assistance in the particular case:

You must decide this case only on the evidence which has been placed before you. There will be no more. You are entitled to draw inferences, that is come to common sense conclusions based on the evidence which you accept, but you may not speculate about what evidence there might have been or allow yourselves to be drawn into speculation.

6.24      It seems to us that, in appropriate cases, this direction might usefully be extended along the following lines:

One matter on which you have not heard evidence is whether the defendant has committed offences before. You might from that be tempted to infer that he or she has. There are good reasons for your not being told about any previous convictions that he or she may have. If fairness both to the prosecution and the defence required you to be told, you would have been. As it is, however, there is no evidence on the matter. It therefore has no bearing on your verdict, and you must not speculate about it.

Where (in the Crown Court) no mention of good character has been made, we assume that this will be because the defendant does have a criminal record. It therefore seems reasonable to direct the jury in terms which, while not expressly confirming that fact (which would be inconsistent with the judge's role), nevertheless confront the likelihood that the jury will be aware of it. Moreover, since the purpose of such a direction would be to protect the defendant, it would not be appropriate to give it without the agreement of the defence. The defence could, if it so wished, insist that no mention of character should be made.

6.25      Arguably this would not go far enough. Michael McMullan points out that

Competent defence advocates often prefer not to leave the jury to make guesses about what the defendant has been up to in the past, if the convictions are innocuous. But if the number of previous convictions is large, includes unpopular crimes, or some are relevant to the offence being tried, the defence will probably prefer not to reveal them. If this pattern became general, the sophisticated juror could infer not only that there are convictions but also that everyone is keeping quiet about them because their nature is, in some sort, damaging to the defendant. "If the convictions didn't matter", they may learn to say to themselves, "we would be told what they are". Innocence cannot be so easily preserved.^[20]

6.26      This danger might still exist even if the jury were expressly warned not to draw inferences from the absence of evidence as to character. In our view, however, the availability of such a direction would probably lead to a reduction in the frequency with which defence advocates chose to adduce evidence of their clients' records. Some advocates might feel that the purpose of doing this, namely to head off uninformed and damaging speculation, could be more safely achieved by invoking the new direction. It is true that, if permitted to do so (and we do not suggest that it should cease to be permissible), advocates might continue to adduce details of the defendant's record where it is clearly irrelevant and the ensuing prejudice is likely to be negligible. However, we think that this would be done far less often than at present, and this would in turn make jurors less ready to draw inferences from an advocate's failure to do it. If a direction along the lines we suggest were available at the defence's request, we think that it would largely meet the concern. Thus option 1 is not the only solution to the problem.

IS OPTION 1 THE BEST SOLUTION?

6.27      There is a significant difference between the fact-finders' guessing that a defendant probably has a record and their knowing exactly how many convictions he or she has and for what crimes. The inevitability of the former therefore does not in itself justify the latter. Option 1 has a number of disadvantages, to which we will shortly turn. Does it have any advantages, in this respect, over the alternatives?

6.28      At present the defence can choose between (a) disclosing the record and (b) not mentioning it, and thus risking adverse speculation by the fact-finders. If our proposed new direction were available, the defence in a jury trial would have a further option, namely (c) invoking that direction. Why should it be fairer to the defendant to deprive the defendant of options (b) and (c), by making it mandatory for the record to be revealed at the outset? As far as we can see, the only circumstances in which this might be fairer are, first, where the defence advocate miscalculates the risk, and, second, where the defendant is of good character but the defence advocate fails to mention that fact – a situation which is only likely to arise in the magistrates' court anyway.^[21] In both cases, some defendants may in fact benefit from having the record revealed rather than having the fact-finders speculate; but others, and probably many more others, will suffer from the prejudicial effect of having their record revealed.

6.29      While we accept that the risk of speculation is a good argument for a change in the present practice, we do not accept that it is a compelling argument for option 1.

Arguments against option 1

6.30      A number of considerations militate against option 1. It would involve the admission of irrelevant material, and material which (even if relevant) is highly prejudicial; the likelihood that defendants may find the need to explain away or minimise the impact of their previous misconduct would make for longer trials, and risk distracting the fact-finders from the real issues in the case; and it might result in the criminal justice system as a whole becoming (or at least being perceived as) less fair.^[22]

Irrelevance

6.31      If the defendant's criminal record were automatically disclosed without the need to demonstrate its relevance, it would not be clear what use the fact-finders were expected to make of it. As Sir John Nutting QC pointed out, "if the convictions are simply read out at the beginning of the prosecution case as part of the background, there is a real risk that they will float vaguely in some evidential limbo, to find an uncertain level of significance in the jury's deliberations."

6.32      Moreover, even the respondents who supported option 1 did not suggest that the defendant's record is always relevant. Sometimes it clearly is not; and in such cases we must ask what purpose would be served by introducing it. Having heard it, the fact-finders would then have to ignore it. This might not matter greatly if the irrelevant material were not prejudicial; but very often it will be.

The risk of prejudice

6.33      In the consultation paper^[23] we distinguished two kinds of prejudice – "reasoning prejudice", the tendency to give bad character evidence undue weight in determining whether the defendant is guilty as charged, and "moral prejudice", the tendency to convict through distaste for the defendant without being truly satisfied that he or she is guilty as charged at all.^[24] Moral prejudice is particularly likely where the evidence of bad character discloses crimes for which the accused has not been punished. We thought there was a danger that bad character evidence could give rise to both these kinds of prejudice.^[25] We said that "in the absence of convincing evidence that fact-finders will not be affected by prejudice, our provisional view … is that proposals should err on the side of caution", and regarded this as a serious objection to option 1.^[26]

6.34      The respondents who considered this provisional conclusion were almost equally divided on it. Supporters of option 1 thought the danger of prejudice was overstated. Some did not address the danger at all. Many asserted that the Commission, and society generally, should "trust the jury".^[27]

6.35      Several respondents thought it patronising of us to think that jurors will be susceptible to prejudice. On the other hand, Phillips LJ wrote "Why do we assume that the jury will give it more weight than it deserves? Perhaps because we feel that we should risk doing so ourselves – that human nature carries with it the risk that such evidence will carry greater prejudice than its probative weight merits." We think it is legitimate to expect the training and experience which magistrates and judges have to reduce their susceptibility to prejudice. We recognise, however, that it is impossible to know how successful the training and experience is in this respect.^[28]

6.36      Some argued that we must trust juries because we have them. We are unpersuaded by this argument. It seems to us that the crucial question is whether there is any foundation for such trust. The best way of determining this is through empirical research; yet few of the respondents who argued for greater openness referred to the empirical studies discussed in the consultation paper.

THE RESEARCH

6.37      Dr Lloyd-Bostock's research on the effect of bad character evidence on mock jurors, which we summarised in the consultation paper, found that a conviction was more likely to result if the jury were told that the defendant had either a recent conviction for an offence similar to that charged or one for indecent assault on a child (irrespective of the offence charged). The former tendency may well be explicable on the basis that it is entirely rational to attach probative weight to a recent similar conviction. However, it is hard to see any rational basis for the latter phenomenon, which would appear to be based on prejudice alone.^[29]

6.38      This conclusion seems to be supported by the experience of respondents. Even those who had faith in juries found that that faith deserted them when it came to sexual offences. One QC wrote that "it is insulting to juries to assume they cannot be trusted properly to evaluate the worth of previous convictions"; yet he was unsurprised by the results of the Jury Study, explaining that "Any criminal hack could have told you that he was terrified of letting his client's character in whatever the nature of the offence where it included [a conviction for indecent assault on a child]."

6.39      Several respondents emphasised the limitations of the research. The most frequent observation was that a mock jury is no substitute for the real thing, and that the results should therefore be treated with circumspection. We acknowledged and emphasised the limitations of the research in the consultation paper,^[30] and accept that studies on real jurors might produce different results. However, given that such research as has been done does reveal a risk of prejudice, we do not accept that we would be justified in asserting that there is no risk of prejudice to real defendants in real trials: we could only do that on the basis of research into real juries.

6.40      Moreover, the results of the Jury Study can now be compared with Dr Lloyd- Bostock's more recent research, which, being concerned with real magistrates, is not open to the same objection. It was found that magistrates tend to regard a defendant with a previous conviction for indecent assault on a child as more likely to commit not only an indecent assault on a woman (which is understandable) but also an offence of violence.^[31] A previous conviction for a section 18 assault, on the other hand, was perceived as increasing the likelihood that the defendant would commit not only another offence of violence but also an offence of dishonesty.^[32] These findings are not easy to reconcile with the view that lay justices (and, by inference, real jurors) are not susceptible to prejudice.

6.41      While we recognise the limitations of the research so far carried out, it would in our view be irresponsible to disregard its findings altogether. Unless and until further research demonstrates that the prejudicial effect on lay fact-finders is acceptably small, we believe that the wisest course is to maintain a general rule against the disclosure of the defendant's criminal record.

6.42      We also believe that the research provides a strong counter-argument to the argument that the fact-finders might as well be told of the defendant's record because they will guess that he or she has one anyway.^[33] It appears that both mock jurors' and magistrates' perceptions of a defendant are adversely influenced by the knowledge that he or she has a conviction for specific kinds of offence (though it seems that the kind of offence giving rise to this effect may vary between jurors and magistrates). This tends to confirm our suspicion that it is easier for fact-finders to ignore a vague perception that "this defendant almost certainly has a record" than to ignore specific knowledge that the defendant has, for example, three convictions for burglary, one for indecent assault and one for resisting arrest.

MINIMISING THE PREJUDICIAL EFFECT OF AUTOMATIC DISCLOSURE

6.43      Some respondents acknowledged that the disclosure of the defendant's record involves a risk of prejudice, and suggested ways of minimising that risk.

6.44      Some thought that option 1, which would involve reading out the record at the start of the trial, would in itself be less prejudicial than disclosure of the record at a later stage. In our view this will depend on the manner in which such later disclosure occurs. We agree that it can be particularly prejudicial if it occurs in cross-examination of the defendant, because of the inevitable suspicion that the defence has tried to conceal relevant information. In Dr Lloyd-Bostock's research, however, this did not occur. The information about the defendant's record was first given to the participants in a voice-over commentary as he entered the witness-box, and repeated in the summing-up (with an appropriate direction on the relevance of that information).^[34] Yet the findings suggest strongly that the disclosure of the information affected the fact-finders' approach in ways which are hard to explain on any rational basis. We doubt that it would have made much difference if the voice-over commentary had mentioned the defendant's record at the very beginning of the trial.

6.45      A few respondents who favoured disclosure of the record thought that, in a jury trial, prejudice could be avoided by an appropriate direction from the judge. In the consultation paper we considered the effectiveness of judicial directions and the available research.^[35] We concluded:

We are uncertain whether juries adequately understand or carry out directions given to them by judges on the use they are to make of evidence of previous misconduct. In the absence of convincing evidence that fact-finders will not be affected by prejudice, our view is that the rules of evidence should err on the side of caution.^[36]

Nothing has been brought to our attention to change this view.

The perceived risk of prejudice

6.46      Even if magistrates and jurors are in fact able to ignore irrelevant information, and to withstand the prejudicial effects of bad character evidence, it is very unlikely that the defendant will be convinced that they have done so and may not feel that justice is being done. The Magistrates' Association made this point, stating,

The defendant is likely to be very resentful if his previous record goes in, feeling that he is being tried and convicted on his record, for which he has already paid the price. Therefore the occasions upon which that record goes in, according to law, should in principle be confined or restricted as much as possible, in order to maintain confidence in the system.

We agree.

Longer trials

6.47      Whatever rules were devised about the use to which the defendant's record could be put by the prosecution, the defence might well wish to call evidence to show that that misconduct was in fact irrelevant, to put it in a more favourable light, or even to prove that the defendant had been wrongly convicted. Once the record had been disclosed, it would clearly be unfair to prevent the defence from doing this. As a result, not only would trials not be simplified as much as option 1's supporters would hope: their length could actually increase. As JUSTICE put it, "Any change to an automatic inclusion of past convictions would … have the consequence of taking up a great deal of court time in defendants explaining past behaviour in evidence, in the context of a trial about something else entirely."

6.48      It may be (as Lord Justice Schiemann suggested) that this danger could be avoided in part by adopting a more flexible version of option 1 under which it would be up to the prosecution to decide whether to disclose the record. This solution would have disadvantages of its own – for example, practice would inevitably vary from one prosecutor to another – but, even if it worked, it would not meet the other objections to option 1. It would still allow the admission of prejudicial evidence with little or no relevance to the issues in the case.

Distraction

6.49      Another practical disadvantage of option 1 is that, even if the defence did not adduce evidence about the offences disclosed, the fact-finders' attention would still be distracted from the real issues in the trial. Lord Justice Schiemann, although favouring option 1, acknowledged that this could be a justification for limiting the admissibility of bad character evidence:

I accept reasons of convenience point towards the exclusion of some facts because their investigation is likely to make it more difficult for everyone to keep their eye on the ball. That may provide a rough and ready reason for excluding convictions in relation to matters which happened "x" years ago.

It may also provide a rough and ready reason for excluding allegations of past misconduct which has not resulted in a conviction. ...

The fairness of the criminal justice system

6.50      In the consultation paper we concluded that "There is a danger that unfairness to those with criminal records would be built into the criminal justice system if previous convictions were freely admitted".^[37] We pointed out that

once someone has a record, he or she is more likely to be questioned than if he or she had no record; if at the trial a previous conviction is admitted and is for a similar offence to the current charge, a conviction is more likely to follow;^[38] the record gets longer and so he or she is even more likely to be questioned when there is next a crime of that type committed in that locality. There is thus a cumulative effect,^[39] which could make it difficult for someone with a criminal record to be acquitted.^[40]

Most of those who commented on this argument agreed.

6.51      We went on to argue that

People with criminal records would be particularly vulnerable to conviction on fabricated evidence where other people know their modus operandi, if it is known that evidence of all similar crimes will automatically be admitted;^[41] finally, as McHugh J put it: "law enforcement officers might be tempted to rely on a suspect's antecedents rather than investigating the facts of the matter".^[42]

Some respondents thought our fear of malpractice by law enforcement officers was exaggerated or did not correspond with their experience,^[43] though Sir John Nutting QC referred to "well documented" instances of such conduct.^[44]

6.52      Even if this fear is not justified (and we do not think it can be discounted) we think our concerns about the cumulative effect to which we referred remain valid. We suspect that public confidence in the system would decrease if bad character evidence was admitted which was more prejudicial than probative, and that that lack of confidence would be to some extent justified.

Conclusion

6.53      In our view option 1 would involve the admission of prejudicial and irrelevant evidence for no very clear purpose, with potentially damaging consequences to the administration of justice and public confidence in it. We reject this option.

OPTION 1A: ADDUCE THE DEFENDANT'S RECORD OF SIMILAR OFFENCES AT THE START OF EVERY TRIAL

6.54      The research suggests, unsurprisingly, that past offences are more likely to be regarded as relevant where they are similar to that charged. A few respondents thought that convictions for such offences, but only those, should be automatically revealed. One favoured admitting all previous convictions for similar offences, subject to a discretion to exclude them in exceptional circumstances if it would be unfair to admit them.

6.55      A practical difficulty with this suggestion is that it is often difficult to classify offences according to their supposedly similar features. For example, it may not be as straightforward as at first appears to determine which offences are "offences of dishonesty". Some drugs offences will necessarily involve an element of dishonesty, such as the illegal importation of a controlled drug;^[45] but many offences commonly thought of as dishonesty offences do not in law require proof of dishonesty.^[46] Producing workable rules for determining when the defendant's record should be automatically disclosed, and how much of it, would be very difficult, and the results would probably be so arbitrary as to bring the system into disrepute.

6.56      In any event, this suggestion is open to much the same objections as option 1. It cannot be assumed that previous convictions for similar offences will always have much probative value. According to the psychological research, the extent to which previous offences of a particular kind suggest guilt of a particular offence of the same kind will depend on the individual features of each situation and the character of the person concerned.^[47] The criminal record in itself does not reveal enough about the previous misconduct for anyone to know how far, or in what way, each conviction is relevant to the current charge. The reason that a defendant's convictions even of similar offences are usually excluded at present is that they would often be prejudicial, and that that fact usually outweighs their probative value. We see no reason to suppose that such convictions are less prejudicial than they have hitherto been thought to be, or that they have greater probative value. Indeed, the limited research so far carried out suggests otherwise.

OPTION 2: ADDUCE THE DEFENDANT'S RECORD OF SEX OFFENCES IN SEX CASES^[48]

6.57      Only the Association of Chief Police Officers was attracted by the prospect of following the lead of the United States Federal Rules of Evidence by putting in the defendant's record for sexual assault where he or she is charged with such an offence.^[49] Several respondents strongly opposed this option. More than one respondent made the point that if, as a general rule, previous misconduct is excluded because it is irrelevant, prejudicial or both, this is no less so in respect of sexual offences. Indeed it is likely to be more unfair to introduce sexual misconduct, because of the greater risk of prejudice.^[50] It would not only be unnecessary to have a special rule for sexual misconduct: it would be wrong.

6.58      It was argued, however, that we had not fully appreciated the potential probative value of sexual misconduct. Professor McEwan wrote:

The Commission concedes that past behaviour can be probative, but does not acknowledge that some propensities can be more significant than others. … Where a personality is abnormal, it appears to be a better predictor of behaviour than situation. The problem is how to identify the abnormal personality.^[51]

She argued that research has shown that serial and sadistic offenders are abnormal, and expressed concern about three recent cases where information about the defendant ought, she thought, to have been available to the jury.

6.59      One of the main reasons mentioned in the consultation paper^[52] for treating sex offences differently from other offences is that the perpetrators are psychologically different from the rest of the population in a way that other offenders are not.^[53] It is argued that sex offenders are motivated by compulsions not shared by "normal people". While burglars, for example, attempt to pursue the normal social goal of acquiring property, but by illegal means, the desires of sex offenders are themselves deviant. This argument entails a belief that sexual offences result from the character traits of the perpetrator to a greater extent than other offences.^[54] If this is the case, the fact that the defendant has such character traits, demonstrated by evidence of previous convictions, would be highly probative.

6.60      In the consultation paper we said:

There are undoubtedly some individuals who form a "small but important sub-group of offenders … for whom clinical or 'special' psychogenic^[55] explanations remain highly relevant".^[56] A possible solution would thus be to allow evidence of a defendant's past conduct to be adduced only if a psychiatrist were able to give evidence that the defendant has a personality defect that "causes" him or her to commit sexual offences.^[57]

We added that the increased danger of prejudice in cases involving sexual offences must not be forgotten.

6.61      If, in a particular case, evidence of some personality defect or disposition is especially relevant because it is abnormal, then (unless it is also especially prejudicial) it should be admissible. For the purposes of evaluating option 2, however, the question is whether evidence of previous sexual offences should be admissible without the need to show such relevance, merely because both the offence charged is also sexual. We are still of the view that the case for such a rule has not been made out. Indeed, since sexual misconduct tends to be more prejudicial than other misconduct, the arguments for a general exclusionary rule seem if anything to be stronger in this case. We reject this option.

OPTION 3: ALLOW EVIDENCE OF THE DEFENDANT'S PREVIOUS MISCONDUCT TO BE ADDUCED ONLY WHERE IT IS AN INGREDIENT OF THE OFFENCE CHARGED

6.62      At the opposite extreme to option 1 is option 3, under which previous misconduct would hardly ever be admissible. The only exception would be where the misconduct is actually an element of the offence alleged. The defendant's previous disqualification, for example, is an element of the offence of driving while disqualified. Subject to that necessary exception, previous offences could not be adduced even if their probative value were overwhelming.

6.63      However sceptical one might be of the value of bad character evidence, there are clearly some cases in which it is so probative that it should be admitted. Understandably, this option received no support at all, and we reject it.

OPTION 4: A SINGLE INCLUSIONARY RULE WITH AN EXCEPTION FOR EVIDENCE WHOSE LIKELY PREJUDICIAL EFFECT OUTWEIGHS ITS PROBATIVE VALUE^[58]

6.64      A few respondents favoured introducing the record at the beginning of a trial as a matter of course, as in option 1, but would also give the court a discretion to exclude previous convictions which were too prejudicial. This amounts to our option 4. While it meets some of our objections to option 1, it also sacrifices one of option 1's main attractions – namely its (supposed) simplicity. We think it would make for complexity and delay if a judicial discretion had to be exercised in every case where a defendant has a criminal record.

6.65      More fundamentally, it is axiomatic that only relevant evidence should be admitted. Not all evidence of bad character is relevant to the issue of guilt. The admission of irrelevant bad character evidence might not matter if it were not prejudicial; but often it is. It can lead to a person being convicted on inadequate evidence, or where the fact-finders are not in fact sure that the charge has been made out. Therefore, bad character evidence which is not relevant should in our view be excluded as a matter of course, not merely as a matter of discretion. We therefore favour a general rule excluding bad character evidence (subject to exceptions) rather than a general inclusionary rule subject to a discretion to exclude.

OPTION 5: AN EXCLUSIONARY RULE WITH A SINGLE EXCEPTION FOR EVIDENCE WHOSE PROBATIVE VALUE OUTWEIGHS ITS LIKELY PREJUDICIAL EFFECT^[59]

6.66      In our view, this option is something of an over-simplification. First, it offers no scope for excluding evidence whose probative value does outweigh its likely prejudicial effect. This might be desirable where, for example, the probative value outweighs the prejudicial effect, but both are of negligible significance, and introducing the evidence would simply distract the fact-finders from the real issues. Second, option 5 amounts to the complete abandonment of any attempt to minimise reliance on judicial discretion, the unpredictability of which we have identified as a defect of the present law. Our preference is for an exclusionary rule subject to exceptions which are not wholly dependent on judicial discretion, but are, as far as possible, objectively defined. We therefore reject option 5. CONCLUSION: THE GENERAL APPROACH WE RECOMMEND

6.67      Our preference, therefore, is still for a structure along lines similar to option 6 – namely, an exclusionary rule with a specified exception or exceptions. Moreover, we now believe that the problems discussed in the consultation paper are best resolved by means of a general approach which extends to evidence of the bad character of witnesses and other non-defendants, rather than defendants alone. In the next Part we give an overview of this scheme.

Ý
Ü   Þ

Note 1   Paras 9.4 – 9.23.    [Back]

Note 2   Paras 9.24 – 9.38.    [Back]

Note 3   Paras 9.39 – 9.57.    [Back]

Note 4   Paras 9.58 – 9.69.    [Back]

Note 5   Paras 9.70 – 9.71.    [Back]

Note 6   Paras 9.72 – 9.73.    [Back]

Note 7   Paras 9.10 – 9.11.    [Back]

Note 8   Paras 6.30 – 6.52 below.    [Back]

Note 9   Para 9.16.     [Back]

Note 10   Para 6.93.    [Back]

Note 11   Professors McEwan and Murphy thought that the consultation paper failed to recognise the particular probative value of evidence of certain kinds of disposition. See para 6.58 below.    [Back]

Note 12   Para 6.63.    [Back]

Note 13   On the distinction between general and specific credibility see A Zuckerman, The Principles of Criminal Evidence (1989) p 248.    [Back]

Note 14   Para 9.7.    [Back]

Note 15   Michael McMullan, “The Wool Pulled Over the Jury’s Eyes is Wearing Thin” (2000) 164 JPN 599, 600–601.    [Back]

Note 16   The research relating to mock juries was summarised as Appendix D to the consultation paper and is referred to as “the Jury Study”. The research relating to magistrates is published at [2000] Crim LR 734, and summarised at Appendix A below.    [Back]

Note 17   See para A.20 below.    [Back]

Note 18   Stephen Silber QC, “Previous convictions?” Oct 1996 The Magistrate 182.    [Back]

Note 19   Penny Darbyshire, “Previous Misconduct and Magistrates’ Courts – Some Tales from the Real World” [1997] Crim LR 105, 109–110. Dr Darbyshire refers to M McConville, J Hodgson, L Bridges and A Pavlovic, Standing Accused (1994) p 216, and J Vennard, Contested Trials in Magistrates’ Courts Home Office Research Study No 71 (1982).    [Back]

Note 20   Michael McMullan, “The Wool Pulled Over the Jury’s Eyes is Wearing Thin” (2000) 164 JPN 599, 600.    [Back]

Note 21   See n 19 above.    [Back]

Note 22   Both the ALRC and the New Zealand Law Commission have reviewed this area of the law, and both concluded that an accused should be protected by special rules about evidence of propensity/tendency and truthfulness/credibility. See the ALRC Evidence (1987) Report No 38, and NZLC, Evidence: Reform of the Law (1999) Report 55.    [Back]

Note 23   Paras 7.7 – 7.15.    [Back]

Note 24   Both terms are taken from A Palmer, “The Scope of the Similar Fact Rule” (1994) 16 Adel LR 161, 169. The Jury Study found that if the defendant had a previous conviction for an indecent assault on a child, the mock jurors were not only more likely to convict him, but were also more likely to believe that he would commit other offences in the future, was more deserving of punishment and was more likely to lie on oath. See Appendix D to the consultation paper.    [Back]

Note 25   See para 7.36.    [Back]

Note 26   Para 9.19.    [Back]

Note 27   Several respondents explicitly endorsed the views of David Pannick QC expressed in his article, “Juries can cope with a defendant’s form” The Times 8 October 1996. More recently, though, he has written, “[The Government’s] proposal that juries should be told about defendants’ previous convictions is wrong in principle and flawed in practice. … It does not make sense to give a dog a bad name.” The Times 17 July 2001.    [Back]

Note 28   At paras 9.13 – 9.14 of the consultation paper we referred to studies which cast doubt on the ability of judges to be immune to prejudice. One of these is a study of 35 Dutch cases where Roderick Munday concludes that, in 13 of the cases, the judges’ decisions had been “critically affected” by the defendant’s criminal records: “Comparative Law and English Law’s Character Evidence Rules” (1993) 13 OJLS 589, 597. In the same study reference is made to an experiment which showed that the order in which information is presented to trained judges affected the likelihood of conviction, which was higher where the record was disclosed earlier. We note that, in France, where the defendant’s record does form part of material on which the verdict is founded, a trained judge sits with the lay fact-finders. Both fact-finders and judge deliberate on guilt and sentence (see pp 593–594). See also A Zuckerman, The Principles of Criminal Evidence (1989) p 245.    [Back]

Note 29   There was a further tendency to regard a defendant as less likely to have committed the offence charged if he had a recent conviction for a dissimilar offence (unless it was for indecent assault on a child). This may be explicable, but strongly suggests an element of irrationality in the reasoning process.    [Back]

Note 30   Paras D.54 – D.59.    [Back]

Note 31   By contrast with the mock jurors, however, the magistrates did not regard such a defendant as more likely to commit an offence of dishonesty.    [Back]

Note 32   But not a sexual offence.    [Back]

Note 33   See paras 6.18 – 6.22 above.    [Back]

Note 34   The videos of a jury trial were used in the research on magistrates as well as that on mock jurors. This did not appear to cause the participants any difficulty.    [Back]

Note 35   See paras 7.16 – 7.20 of the consultation paper.    [Back]

Note 36   Para 7.37.    [Back]

Note 37   Para 7.38.    [Back]

Note 38   See Appendix D to the consultation paper.    [Back]

Note 39   As described by, eg, C Tapper, “Proof and Prejudice” in E Campbell and L Waller (eds) Well and Truly Tried (1982) p 207.    [Back]

Note 40   Para 7.23.    [Back]

Note 41   D McBarnett, Conviction (1983) p 113.    [Back]

Note 42   Para 7.24. The quotation from McHugh J was taken from Pfennig (No 2) (1995) 127 ALR 99, 136.    [Back]

Note 43   H M Customs & Excise pointed out that since the consultation paper was published, s 23(1)(a) of the Criminal Procedure and Investigations Act 1996 has created a statutory duty on police officers to pursue all reasonable lines of enquiry.    [Back]

Note 44   Similarly, Judge Rivlin QC warned: I fear that if evidence of previous misconduct were admitted as a general rule it would quickly bring our system of criminal justice into disrepute. Take any of the high-profile appeals of recent years, and add into the equation that at the outset of the trial the jury were told of the defendant’s convictions.    [Back]

Note 45   Misuse of Drugs Act 1971, s 3.    [Back]

Note 46   Eg taking a vehicle without consent (Theft Act 1968, s 12).    [Back]

Note 47   See paras 6.11 – 6.21 of the consultation paper.    [Back]

Note 48   Paras 9.24 – 9.38.    [Back]

Note 49   In 1994 Rules 413–415 were added to the Federal Rules of Evidence. Rule 413 refers to offences of sexual assault, 414 to sexual abuse of children, and 415 to civil cases arising out of such offences. In cases of sexual assault or of molestation of children a party is now entitled to introduce evidence that a defendant has previously committed such an offence if it is relevant to any matter, including, presumably, the defendant’s disposition or tendency to commit the act charged. We take note of a critique of the new Federal Rules, published since the consultation paper, which contains a detailed and persuasive set of arguments to the effect that the new Rules are not only wrong in principle but are likely to backfire in practice: Katharine K Baker, “Once a rapist? Motivational evidence and relevancy in rape law” (1997) 110 Harv LR 563.    [Back]

Note 50   The Jury Study confirmed this view.    [Back]

Note 51   Jenny McEwan, “Law Commission Dodges the Nettles in Consultation Paper No. 141” [1997] Crim LR 93, 95–96.    [Back]

Note 52   See paras 9.25 – 9.27.    [Back]

Note 53   It was said in Congress that rapes are committed by a “small class of depraved criminals”: 137 Congressional Record S3241, 13 March 1991.    [Back]

Note 54   Ie, while non-sexual crimes may be the product of certain situations, sex offences stem from an inherent character defect in the offender.    [Back]

Note 55   Something is “psychogenic” if it has an emotional or psychological origin.    [Back]

Note 56   R G Broadhurst and R A Maller, “The Recidivism of Sex Offenders in the Western Australian Prison Population” (1992) 32 British Journal of Criminology 54, 72 (footnote added).    [Back]

Note 57   Para 9.36.    [Back]

Note 58   Paras 9.58 – 9.69.    [Back]

Note 59   Paras 9.70 – 9.71.    [Back]