4 January 2010

Tobin's Law

I once saw a presentation by a cognitive psychologist, interested in jury reasoning and evidence analysis. Studies of juror behaviour and processing of information in the UK are very scarce, scholars being scared off by the provisions of the Contempt of Court Act 1981. Instead, researchers compose their own panels of ordinary and irregular members of the public, run them through a dramatised trial, or alternatively pose questions about particular pieces of evidence in the abstract. The psychologist I encountered had conducted a pilot study attempting to determine how people reason about a prior conviction for a similar act. The fact-pattern described a crime of violence. The researcher invited the people being questioned to vote ‘guilty/not guilty’, based on the evidence provided. The fact-pattern was then embroidered with extra detail, and the same question about conviction or acquittal posed. The goal to discern, more or less probabilistically, how particular pieces of information might affect jurors in the real world.

In one version, the traumas which marked the victim’s body indicated that the attacker may have been right-handed. In an alternative version, the damage suggested a left-handed attacker. If the accused person is left-handed, this pushed up the number of jurors who thought he was more probably guilty and so on. The final section of her study, the psychologist added another detail – that the accused person had been convicted of an analogous crime previously. Guilty, not guilty? Although I cannot recall off the top of my head the exact percentage leap – the jurors veered radically and illogically towards voting ‘guilty’. All of this, simply on the evidence that the attacker was left-handed and the man in the imaginary dock had been convicted of something similar before. The study was small, the numbers not terribly statistically significant, nevertheless it suggests that the old legal saw is probably empirically verifiable – the prejudicial influence of adducing evidence of prior convictions outweighs its probative value. All the more interesting, because the view does not emanate from some crusty, claret-soaked QC – but from science-totting psychologist.

At present, this rule is enshrined in the Criminal Procedure (Scotland) Act 1995, s101. Sections 266 and 270 make exceptions, however, when “evidence is led by the defence, or the defence asks questions of a witness for the prosecution, with a view to establishing the accused’s good character or impugning the character of the prosecutor, of any witness for the prosecution or of the complainer or the nature or conduct of the defence is such as to tend to establish the accused’s good character or to involve imputations on the character of the prosecutor, of any witness for the prosecution or of the complainer.” In these circumstances, the could “may” permit the prosecutor to “lead evidence that the accused has committed, or has been convicted of, or has been charged with, offences other than that for which he is being tried, or is of bad character”. One caution. It is important not to get too caught in the structure of “ordinary rule and exception”. There is, as far as I’m aware, no data available clearly demonstrating how ordinary the ordinary rule is. Or to put it another way, how exceptional the exception is. My hunch is that the accused’s prior convictions being entered in evidence to rebut his or her imputations of good character happens far more regularly than the bare text of the Act seems to admit.

All of this is relevant because Kenny MacAskill has asked the Scottish Law Commission to examine the admissibility of prior convictions. The ever-helpful Criminal Justice Scotland collates the press-cuttings. It is claimed that the impetus for this move is Peter Tobin’s gradual accumulation of life sentences for killings across Britain. In his most recent trial, in Chelmsford, it is my understanding that prosecutors were able to present the jury with Tobin’s past bad acts, under England’s different legal provisions. The nationalist MSP, Stewart Maxwell, time weighing heavily on his hands since he was booted out of ministerial office, seems to be agitating for something similar in Scotland. The usual caveats are made – only serious offences, only affecting a small handful of cases, these cases described as ‘special’. The brutal fact is, laws don’t work like that. The exception may become the rule, depending on how legal actors behave. Other provisions may lapse into comparative desuetude, despite the powers they seem to provide. We simply can’t be confident about the number of cases affected. Moreover, if the change is justified by a desire not to exclude 'relevant' information, why only extend it to 'serious' cases, here presumably defined as cases tried by jury? What is the justification for excluding 'relevant' information in more piddly criminal fare? To put the other point of view, is it not terrible naive to imagine that the jurors were empanelled in subsequent cases against Peter Tobin, totally ignorant of his crimes? Given this, is it not preferable that we are honest about the difficulties presented by the Google age and let the jury have access to detailed information, rather than the digested version, available in the popular press? The response to this doubting little thought is pretty obvious - Tobin's case is exceptional, a poor justification for a reform which might empower prosecutors far beyond a serial-killer fact-pattern, however monstrous and regular its operation.

While the Herald and Scotsman host rather bland responses, the Times is joyously, rip-roaringly hostile. In general, I sympathise far more strongly with the critics here than with the proponents of change. Given MacAskill's reference to the Law Commission, this issue will be floating about for a while, with discussion documents, consultations and the like to follow. Consider this one, pre-emptive, negative response.


  1. I recall reading about some of the studies done into the way witnesses of crime recall the incident, as well as the behaviour of juries. I think Richard Dawkins may have made comment on it. It's interesting stuff that deserves further serious study.

    This proposal by the Justice Secretary makes me decidedly uneasy.

  2. Lallands I was hoping you'd write about this once I read about it in the press. Many thanks.

  3. I can see both sides, and I'm certainly aware of the shortcomings, however I do think previous convictions are relevant and should be admissible.

    Mind you I had a thought of a quote from a Scottish judge (possibly Irvine Smith)

    "Not guilty? If you were not guilty you wouldn't be here!"

  4. Jim,

    I'm (loosely) familiar with the literature you mention. Much of it is to do with that far more broad consideration - how our memory functions. I gather that findings from psychology on this front have suggested changes in how identification evidence is gathered. Assessing the veracity of the traditional line up seems to pose a bit of a quandry. Memory seems to be reconstructive. If given an array of six or more choices, witnesses tend to select the individual of the group who looks most like their recollection.

    This 'pressure to choose' can be avoided by showing the witness the six individuals in succession. Without lining them up side-by-side, in short.

    As you say, an interesting and unconsidered area. I'll root about a bit and see if I can find the citation for the original study I mentioned. If I've any luck, I'll post it up here.

  5. Bigrab,

    There are certainly alternative arguments. In the Herald, Colette Douglas-Home proposed a few. You can read it here:


    For my money, her account of criminal trials is hopelessly idealised. In particular, if you think these convictions should be admitted - why? Are they always relevant? If not, how is relevance determined? Does it change with the passage of time? Does it depend on the type of crime prosecuted?

    Maxwell himself claims to accept that "the fact that you have done 96 robberies before does not mean you have done this one" - but argues that there is still "room for evidence to be led". Why lead it then? Presumably Maxwell has the idea that guilt should be based on other, presumably substantive facts. If that is so, why is a change necessary?

  6. Sorry I hadn't replied LPW. They may or may not be relevant. However I believe that in certain cases previous convictions are central to the case - for instance in a driving while disqualified case how would the case be prosecuted without reference to the previous case?

    I have friends in the legal profession who have described cases where guilty men have walked free because jurors did not have the full facts which would in all probability have seen them convicted.

  7. Bigrab,

    Spot on - but in very specific circumstances. The example you mention is absolutely a case of this - that prior conviction and punishment - and an attempt to circumvent that punishment - constitutes the definition of the new crime alleged to have been committed.

    That, I'd certainly argue, is quite different from allegations of criminality which seek to use prior convictions as a means to demonstrate the truth or falsity of the prosecutor's allegations. To put it in the context you mention, this second form of 'proof' would be to use evidence that the accused person had been convicted of driving while disqualified before, rather than simply referencing that a ban was in legal effect. That is quite a different barrel of herrings to me.

  8. What if......

    A criminal with a particular and unusual modus operandi is placed in an id parade and positively id'd by the victims (two of them separately).

    The accused had made a long career out of his speciality and had served several jail sentences.

    The jury failed to convict him because he simply denied being at the crime scene despite the positive id by two witnesses.

    Not proven.

    However had the jury known of his previous crimes would they have reached this decision?

    This was an actual case, about which I can't be more specific but you catch the drift.

  9. Ah ha!

    There is a tougher challenge, presumably the modal sort of case in Mr Maxwell's mind. Our answer will depend on which side of one particular debate you err on - should the balance be for putting things before the jury, letting them whittle right from wrong - or should the evidence be excluded wholesale. In general, I'd say the populist faith of the first argument is its strength - its difficult for a popularly elected tribune to cry moron, and point his or her finger at her electorate. As I tried to argue above, one way of approaching the question is to ask - does the prejudicial effect of that information outweigh its probative value? One of the strengths of this position, I'd suggest, is that it admits your central premise - that an unusual modus operandi may seem to have probative value - but equally, that disclosing that information, whatever its value, would, generally speaking, be unfairly disposed to lead to conviction. Against this proposition, we can fling another. Indeed, the sort of argument Maxwell himself has in mind.

    Why not put it to a judge, adjudicate in particular cases on the relevance or otherwise - rather than taking an 'overly broad' decision on either side of things, totally excluding evidence of prior bad acts or sanctioning their total inclusion. I'd suggest this sort of 'consensus-middle' position is misleading - that judicial adjudication on admissibility will be disposed to be arbitrary, or worse, will basically tend towards the wholesale admission of evidence which this consensus position claimed it was attempting to avoid.

  10. But, but, but....................

    If you agree with a jury system, then why not let the jury decide? whilst having the benefit of all the information?

    Unfair or not and notwithstanding your stout defence, the guy did it!

  11. Largely for the reasons I set out in the first post - there seems to be a reasonable basis to conclude that introducing evidence of past convictions in this way will be unduly prejudicial, significantly outweighing the plausible uses it might have in some cases. Moreover, I'm not convinced that the safeguards and checks envisaged (loosely) by the tribunes who have discussed it will in reality come to be used in that manner. This is significant, since it is not unusual for the apparent exception to become the legal rule. Moreover, by no means do juries, in general, have the benefit of 'all the information'. Despite best intentions, that is not what trials are primarily concerned with. The law weeds out context, much of the human significance. Its in this context of profoundly limited information placed before most juries, I'd suggest, we have to consider the prospect of detailing to that jury grim information of past offences. Needless to say, I've no interest in the guilty escaping their lawful reward. That too is a miscarriage of justice. That said, we ought to be able to demonstrate guilt in a specific offence with the material facts of that offence. Not with the easy innuendo of introducing evidence of past bad acts, confident that the psychological findings will be borne out, and prejudicial judgement will overcome the weak evidence of the prosecution.

    As I've noted - I can see that there is an argument, that my position isn't self-evidently true. That said, I remain hostile to the proposal. Even conceding many of your arguments, I'm not persuaded that my doubts, and the dangers faced by accused persons, are sufficiently allayed.

  12. I come along here 6 months after the above remarks - but even at this late stage can I mention another side of the previous convictions coin, viz

    My late mother-in-law told me that she'd served on a jury many years ago in which the accused was a presentable-looking young chap charged with something quite serious (I can't remember what). The evidence against him was quite strong and all the jurors thought that he had 'done it'.

    But, one of the more confident jurors advised the rest that the accused 'obviously had never done this sort of thing before, OTHERWISE THE PROSECUTOR WOULD'VE BRUNG IT OOT'.

    The jurors thereupon decided to 'give the wee boy a chance, him wi' his mammy sitting greetin' in the court' and found the charge not proven.

    It may well be obvious and well-known to lawyers and the better-educated that the accused's record cannot be 'brung oot', but I do not think it's obvious or well-known to all.

    It might be quite awkward to explain to jurors that the absence of a record being 'brung oot' doesn't mean that there isn't one, but otherwise I am sure many juries proceed on the wholly unwarranted assumption that the accused is a first-timer.

  13. You may well be right on that front, Almax. It would take a rather deft formulation to make the point that the jury should neither assume the accused is a criminal tabula rasa nor that they have a well-chalked board of criminality. No doubt an attempt is made to convey that understanding, by stressing that their decision should only be based on the evidence lead before them, rather than extraneous interpretations and additions.

    You gently allude to another point which is problematic too, Almax - that assumptions about the accused's criminal history are likely to be informed by such things as appearance, presentation - or to use a slightly pretentious phrase from sociology - "bodily hexis". In this respect, the shiny looking, respectable accused is far more likely to attract such favourable (and potentially fictitious) inferences of previous good conduct in the absence of evidence than other, less well turned-out or otherwise "innocent" looking characters.

    Defenders of the principle would undoubtedly give this answer - that the balance of doubt favours and ought to favour the accused. While the unjustified inferences you described can certainly occur, such blips seem preferable to the risks associated with the alternative.