Why do we keep sending innocent people to rot in jail? Why do we let obviously violent criminals wander the streets? Why is it so difficult to punish the right people? These are reasonable questions, but answering them involves facing up to some very unpleasant decisions. This article reviews some of the more alarming cases of recent times - some well known, some less so - then attempts to extract some common features from them. It will be shown that before any real progress can be made, we as a society will have to decide to what extent we agree with the words of Sir William Blackstone: ``It is better to let ten guilty men go free than to incarcerate one innocent man''. Until we can do that - and it isn't easy - things will not improve.
|
The cases of Judith Ward, the Guildford Four, the Birmingham Six and the Macguire Seven are the archetypal miscarriages of justice; between them they reveal falsified and concealed evidence, police brutality, untrustworthy forensic evidence, and forged confessions. The cases themselves are well-known, so only an overview is presented here.
In 1974, the IRA embarked on a vicious campaign of bombings, which
left many people dead and injured. In response to the public outrage
against the bombers, the Government introduced the Bill that was to
become the Prevention of Terrorism Act (1974). This legislation,
arguably the most draconian ever enacted in modern Britain, passed
through the Commons and the Lords in one day. The Act gave new powers
to the police to arrest on suspicion of terrorism, and to detain
suspects for up to five days without charge. It also made membership
of the IRA an offence, probably the only time mere membership of an
organisation has been criminalised in English legal history.
The `Maguire Seven' and the `Guildford Four' were the first to be
arrested and questioned under these new powers. Their arrests were
for alleged involvement in the bombings of pubs in Guildford and
Woolwich. The Four were charged with multiple counts of murder,
the Seven with manufacturing and supplying them
with explosive devices. Judith Ward was arrested for bombings
of a coach on the M62 and at Euston station,
while the `Birmingham Six' were arrested for two other pub bombings.
It is probably fair to say at the outset that some of those arrested
in these cases were not exactly pillars of the community. One of the
Four, Paul Hill, may have been involved with a number of IRA
activities, including the abduction and murder of ex-serviceman Brian
Shaw. However, he strenuously denies this and, indeed, any connection
with the IRA. Among the other members of the Four and the Seven, some
are known to have had IRA sympathies and may have committed other
offences at its behest. Judith Ward confessed to being a member of the
IRA, and variously confessed to, and denied, active involvement. The
Six have always denied any connection with the IRA.
Whatever else they may have done, none had any involvement in the
crimes for which they were convicted, as is now clear.
Of the various cases, that of the Guildford Four is the most
contentious, partly because of the youth of the defendants (the
youngest, at 17, still legally a minor), and the absence of any solid
evidence outside of their confessions.
To cut a long story short, the evidence collected by the police from
witnesses in the bombed pubs and from informants was inconclusive and
contradictory. One of the Four - Richardson - even had an alibi.
Although she was unable to remember her whereabouts on the night of
the Guildford bombing, a friend later came forward and told police
that he was with her that night from 6.30 onwards, along with three
other people. He was was arrested and detained for three days without
charge. His statement said that he and Richardson and her friends went
to a concert at South Bank Polytechnic, where the doorman remembered
seeing them at about 7.30. If Richardson had been in Guildford when
the bombs were planted, these timings would have required her to make
the Journey to the South Bank in about 50 minutes, on a Saturday
evening. Even with 1974 traffic levels this would be a virtual
impossibility, although the police claimed to be able to drive it in
48 minutes.
All this evidence was
withheld from the defence lawyers until just before the trial.
In the end,
the prosecution's case hinged on confessions alleged to have been
made by the defendants.
Even at the trial it was clear that the confessions of the defendants
did not tally, and were factually erroneous.
In Hill's confession he described, for example, a flat in Brixton that
was used for storing explosives. When police visited the address,
however, they found it to be the home of an elderly couple, and no
evidence of any explosives. Some of the defendants confessed to
things that would have required them to be in two places at once. All
in all, even at the trial there were over a hundred discrepancies in
the confessions. The prosecution put this down to deliberate
disinformation, intended to confuse the police.
All four suspects in the Guildford case later repudiated their
confessions. All four alluded to violent and threatening behaviour
from the police. Hill, in particular, was apparently threatened
at gunpoint. Gerard Conlon claimed that he had been beaten up by
a detective inspector during questioning, who had taken off his
jacket and rolled up his shirtsleeves. As a result, Conlon was
able to identify a tattoo on the inspector's arm at trial. On one
occasion during the trial, Hill arrived late and badly bruised -
the result, he said, of police ill-treatment. Richardson
was interrogated and tried whilst still a minor, without legal
representation or even a parent present.
Paul Hill, who was 21 years old at the time, was sentenced to life
imprisonment with a recommendation that he never be released. His
sentence reflected his leading role in both the Guildford and the
Woolwich bombings. The other members of the Guildford Four were
sentenced to 20 to 35 years' imprisonment. Passing sentence, Donaldson
J (as he then was) expressed regret that the death penalty was no
longer available.
The Macguire Seven were also sentenced to long prison sentences.
Here, however, the prosecution's case was supported not primarily by
confessions, but by forensic evidence. The defendants were found to
have traces of nitroglycerin on their skin which, it was claimed,
could only have come about by contact with bomb-making equipment.
Similar forensic evidence was obtained against Judith Ward and the
Birmingham Six; like the Four, the latter were also, it appears,
ill-treated by the police in the hope of obtaining a confession. The
Six only came to the attention of the police because they were
attending the funeral of a friend who was known to be an IRA activist.
Ward had a history of mental illness - a fact withheld from the jury -
and made various contradictory confessions during interrogation and at
trial. All were convicted and sentenced to long prison terms.
|
As an aside, it has to be questioned whether it is the remit of the Court of Appeal to conduct an evaluation of the evidence in a case - this is, after all, what the trial process is for. The case of the Guildford four is not the only one at which Lawton and Roskill applied this process, and were later criticised by the House of Lords.
Over the next ten years, evidence gradually came to light of police
incompetence and bad faith, and the deliberate concealment of evidence
from the defence. Eventually the Home Secretary, responding to public
pressure, referred the cases of the Guildford Four and Birmingham Six
to the Court of Appeal again. This time it held that their
convictions were `unsafe and unsatisfactory'. They had all
spent about 15 years in prison.
As a result of the outcome of the Guildford Four appeal, another
hearing was held on the Macguire Seven case. Their sentences were
quashed in 1992, although by then all had served their time - apart
from Patrick Conlon who had already died in prison. Judith Ward was
also released in 1992 after serving 18 years.
To date, a total of 38 people allegedly involved in IRA activities have had their convictions quashed. In 2000, the Prime Minister made an official apology to the Guildford Four.
Activities of Irish terrorist organisations have always had a high
profile, as have the miscarriages of justice that resulted from them.
It is essentially the activity of journalists and campaigning lawyers
that have kept these cases in the public eye. Without this, one has
to wonder whether the Guildford Four wouldn't still be in jail, along
with all the others wrongfully convicted. For victims of miscarriages
of justice with no public support, the prospects are not so good.
A particularly shocking case is that of Carol Hanson, who spent an
incredible 27 years in prison before committing suicide. Hanson was
convicted along with her husband for the brutal murder of a
10-year-old schoolgirl. Her husband, it later transpired, had
confessed his guilt - and his wife's complete innocence - to his own
barrister, but insisted on pleading `not guilty'. It remains unclear
why his barrister was prepared to go along with this, but it had the
effect that the Hansons were tried together on a not-guilty plea,
rather than separately. This was unfortunate for Carol Hanson, whose
testimony could clearly not be distinguished from her husband's by the
jury. Under the rules of evidence, a person cannot be compelled to
give evidence for the prosecution of a co-accused so long as both are
pleading not guilty. Thus Carol's counsel was unable to cross-examine
her husband in the presence of the jury.
Later the husband told Carol's lawyers that he had killed the child on
his own, and implicated his wife out of fear that she would be
unfaithful to him if she were acquitted (who, after all, can fathom the
twisted logic of a nutcase?) This was sufficient to get Carol an
appeal but, incredibly, Widgery J ruled that her bad fortune was not
sufficient to raise doubts about the safety of her conviction. So she
remained in prison without complaint for a further 19 years before
becoming eligible for parole. This was refused without any
explanation. She was found drowned in the prison swimming pool; there
were no suspicious circumstances.
What is incredible about the Hanson case is the way that the accused
simply vanished from sight after her conviction. It is almost unheard
of for a woman to serve more than twenty years in prison, so how did
Carol Hanson get overlooked for so long?
One has to wonder how many other people are wrongfully languishing in prison, lacking anyone to fight their corner in public. Consider the case of Sally Clark, the solicitor whose two babies tragically died in their cots, and for whose murders she was convicted last year. In her case, as in many others, a conviction was secured on the basis of misleading and inaccurate expert testimony. Without a determined (and wealthy) husband to promote her case, would her appeals (two were required) have been heard so soon? Or even at all?
Campaigning organisations claim that there are dozens, possibly hundreds, of people serving long prison sentences for crimes they most likely did not commit.
An examination of the well-known cases reveals that some or all of the following factors are present in every one.
Many, but not all, quashed convictions have been associated
with dubious police-work, occasionally straying into outright
criminality. The Police and Criminal Evidence Act (1984) was
supposed to prevent oppressive questioning and reduce the
likelihood of fabricated evidence finding its way to the jury. It
gives judges the power to disqualify prejudicial evidence,
and puts the burden of proof onto the prosecution to show that a
confession was obtained fairly. There are strict rules about
the questioning of suspects and the handling of evidence. For
example, it is now a requirement that interviews conducted in
a police station must be tape recorded, and that a suspect
who makes a confession outside a formal interview be given
an opportunity to repeat it on tape.
|
It is not clear whether PACE would have protected the Guildford
Four and the other IRA suspects wrongfully convicted. Perhaps not: a
tape recorder will record a confession extorted by beatings and
threats just as well as a real one.
What is clear, however, that questionable police-work continues to be a
problem. Of particular concern at present are the activities of a
number of CID officers in South Wales, which have given rise to a
string of wrongful convictions over a twenty year period. Last year,
the `Newsagent Three' were released from prison after an appeal in
which it was revealed that one suspect was questioned while handcuffed
to a hot radiator. The officer responsible has since retired on full
pay, but is currently under investigation in connection with eight
other cases in which convictions were later overturned. Sixteen
officers from this force - some of whom are still working - are
associated with a total of 19 wrongful convictions. Investigation of
these cases has revealed bullying of suspects, destruction and
falsification of records, long off-the-record interviews, forged
confessions, and non-disclosure of evidence.
If recent proposals to install 24-hour video monitoring equipment in police custody areas come to fruition, some of the more flagrant abuses of the criminal process may be avoided.
Although police misconduct has been implicated in many cases where a confession was later shown to be false, this is by no means true of all cases. Judith Ward walked into a police station and gave herself up for a crime we now know she did not commit. It used to thought that only people of pathologically low intelligence would confess to other people's crimes, but pioneering research by Gisli Gudjsonsson and James MacKeith (now a member of the Criminal Cases Review Commission) has shown that even people with `normal' IQ make false confessions as well. It appears that some people are abnormally suggestible, and can be convinced that they committed crimes. Sometimes they can even convince themselves. This was first accepted by the Court of Appeal in 1991 when the three alleged murderers of PC Blakelock were released. The Court held that one of the confessions should not have been admitted at trial, as the defendant was abnormally suggestible. The others were also unreliable as handwriting analysis showed that notes made at interview could have been falsified.
The convictions of Judith Ward, the Birmingham Six, and - more
recently - Sally Clark have all been supported by forensic evidence
that was later revealed to be defective.
You may be surprised to know that the practice of forensic science is,
essentially, unregulated in the UK. Anyone can be employed by a
solicitor to appear as an `expert witness' in a court hearing. If
the expert is clearly a charlatan, then the evidence offered may not
be believed by the jury. However, juries have not shown themselves
well suited to deal with scientific evidence - and why should they be?
Professional scientists spent years learning how to assess scientific
evidence: why should a juror be able to figure it out in a few
days?
Until 1989, most police forensic work as done by the Forensic
Science Service (FSS) - a part of the Home Office. The FSS charged
police authorities for its services, but the charges were based on
the size of the authority, not the amount of work.
The FSS was not without its faults - even now, reports of
improperly maintained equipment and careless handling of evidence
are still coming to light. However, it had at least the merit that
it was under no pressure to produce a particular `result'. Its staff
continued to get paid regardless of the outcome of a particular
case. When the FSS was `privatised' in 1989, this all changed.
It is now the case that police services are able to contract
their forensic services to whichever laboratory can provide them most
cheaply. A more sinister interpretation is that police services
can now contract with the laboratory that produces findings
most favourable to the police case. Strictly speaking, the
expert witness owes his first duty to the court. However,
a number of respected forensic scientists have intimated
that they are aware of cases in which expert witnesses
dishonestly biased their analysis to produce the desired result.
Since the likelihood of being employed again is likely to depend
on producing the results required by one's client, the temptation
to do this must be considerable.
The deregulation of the forensic science industry was based on
an entirely false premise. It assumed that the principles
that favoured privatisation of the gas and electricity
industry would have the same benefits in forensic science.
That is, extending customer choice would lead to a reduction in
costs without a reduction in quality. This premise was false
because the `customer' in forensic science is not the person
who determines the quality of the service provided. The state
largely pays for the service, and is thus the `customer', but
the impact of a poor quality service is not on the state, but
on the accused.
Even where the forensic laboratory behaves conducts itself impeccably, some forensic tests may not be all that powerful anyway. Even that old stalwart - the fingerprint - has come under question lately, after a series of convictions were overturned when fingerprint evidence was found to be unreliable. The most well-known case is that of Danny McNamee, jailed in 1982 for an IRA bombing. One of the main planks of the prosecution case was a fingerprint found on a battery at the crime seance, claimed by the prosecution to be McNamee's. After eleven years in prison, his case was reopened by the CCRC, and a total of twenty experts looked at the fingerprint evidence; no unanimity emerged. Even if fingerprints really are unique to the individual, they are only a perfect identifier if obtained in laboratory conditions.
During a criminal investigation, a large amount of evidence is
amassed: witness statements, forensic results, photographs, and so
on. Not all of this evidence will be adduced at trial; much of
it will not even be admissible. However, some of it may
undermine the prosecution case, and therefore be of
interest to the defence. Historically, the police were frequently
accused of withholding vital information from the defence.
One of the most alarming cases was that of Stefan Kiszko, who
spent 15 years in prison for the sexual abuse and
brutal murder of an 11-year-old girl. It eventually came to
light that Kiszko was medically incapable of committing the
crime, a fact that was concealed by the police (attempts were
made to get senior police officers and forensic scientists
to face trial for perverting the course of justice, but
no prosecutions were ever brought). As a result of this case,
and that of Judith Ward - whose mental health problems were
concealed from the jury - the Court of Appeal held that the
police had to disclose all evidence to the defence.
Until 1996 there thus existing a fairly
haphazard system of disclosure.
Essentially all the evidence was photocopied and put out on a
big table, and the
defence lawyers came and rummaged through it. The problem with this
approach was that it might be very time-consuming to sift through
it all, while the police ought to know where the defence should
be looking. The police also complained that the photocopying was
an administrative burden.
The Criminal Procedure and Investigations Act (1996) - CPIA -
sought to regulate the disclosure procedure. Essentially, it
puts the burden on the police to disclose all evidence to
the defence that they thought might weaken their case. The process
is, in principle, overseen by the Crown Prosecution Service.
In practice, the system is implemented in police stations by
`disclosure officers'. The job of the disclosure officer is
to sift the reams of evidence and select those pieces that the
defence will have an interest in.
The hazards of this system should be obvious. First, if the
police have invested large amounts of time and public money in
building a good case for the prosecution, they will naturally be
reluctant to see it knocked down by their own unused evidence.
Second, even assuming that the police do play by the rules,
the police disclosure officer may not have the
legal experience to be a good judge of what evidence would be
of interest to the defence. In practice the disclosure officer
is usually a junior detective, and it has been suggested that
senior officers simply keep evidence locked away.
|
Whatever the constitutional merits of juries - and I'd be the last person to kick the jury system - juries are not always blameless in miscarriages of justice. In many cases the jurors are not themselves to blame. For example, no blame can be attached to jurors for being unable to deal with the complexity of some of the cases which they hear.
There's no doubt that a jury is
not well placed to deal with complex scientific and statistical evidence.
This is a fact that is exploited by barristers, innocently or
deliberately. It's probably innocent, because a barrister will express
things in the way that emphasises his argument; with statistical
evidence in particular, different expressions of the facts can have
very different meanings.
Here is an example. A man is killed in a violent struggle; two types
of blood are found at the scene: one type O+, which is the victims,
and the other A-, which - we assumed - belongs to the killer. A- is
a very rare type; let's say it occurs in one person in a thousand.
When a suspect is finally arrested for the crime, it turns out that
he has type A- blood. How strongly should that fact influence the
jury?
Consider these two statements:
1. There is one chance in a thousand that the blood found at the scene did not come from the defendant.
2. One person in a thousand has the same blood type as the killer.
Only one of these statements is true: can you tell which one it is?
For most people, the two statements express the same fact: they
simply can't tell the difference. In reality, they are no more similar
than the two statements `all cats have fur' and `all things with fur
are cats', but because the statements about blood groups involve
numbers, that seems to throw people off.
Statement 2 is true, statement 1 may be true, but probably isn't.
If A- blood is found in one person in 1000 then, since there are
30 million adults in the UK, this means that 30,000 of them
have A- blood. So, all other things being equal, the probability
that the accused committed the murder is 1 in 30,000. There is a big
gap between one chance in a thousand of being innocent, and one chance
in 30,000 of being guilty. Of course, the probability of the accused
being guilty depends on a host of other things: where he was at the
time of the crime, his possible motive for killing the victim,
the other scientific evidence, etc.
|
My favourite (slightly off-track) example of a statistical error of
this type is this: 65% of heroin addicts say they started with
cannabis. Therefore if you smoke cannabis there's a 65% chance of
becoming a heroin addict. Oh yeah? 100% of heroin addicts started
with milk (think about it). Are we all doomed to become addicts?
Of course, the fact that our hypothetical defendent has the same
blood type as the killer is of some evidential value. It
shouldn't be assumed that we should simply ignore it. But how
much value? This is where things get complicated. To assess the
weighting that the jury should put on this piece of evidence, we
need to apply Bayes' theorem. This will tell us not how likely
it is that the defendant committed the crime, but how much
more likely it is, given this particular piece of evidence.
In the case of R v Adams ([1992] 1 Cr App R 377, for any lawyers that
are interested) Lord Bingham refused to allow a scientific
argument based on Bayes' theorem to be offered to the jury.
He said that although it was a methodologically
sound approach, it would confuse the jury. The jury thus had
to fall back on its common sense (for which, in a case like this,
read `guesswork'). Sadly, his Lordship was probably right.
The barmy thing is that the more extreme the numbers are, the more
likely they are to bamboozle the jury. In the Sally Clark case,
the Pathologist Dr Meadows told the jury that the probability that
Mrs Clark's children died of natural causes was about `one in 70
million'. Even had this been true (which it wasn't - see below),
even had it been
true, would it have been correct to say that Mrs Clark had less
than one chance in 70 million of being innocent? Of course not: all
that could reasonably be concluded was that the medical evidence
increased the likelihood of her guilt by some amount (which could,
in theory, be calculated). But since we have no way of knowing how
likely it would have been that she would have been convicted
without this evidence, the evidence itself is of no value. In
principle, it should never even have been admitted to the court.
Even if the `one in 70 million' figure had some probative value,
it would only be useful if it were correct. However, it was derived
from a basic statistical error: the belief that the likelihood
of two cot deaths was the square of the likelihood of one. This
would only have been true if the likelihood of one death did
not affect the likelihood of another. However, tragically,
the medical evidence suggests that if one baby has
died the second is more, not less, like to die as well.
It is probable, I think, that even a first-year undergraduate in
maths would have spotted this error.
One of the incredible things about the Guildford Four trial is that the jury did not reject the prosecution case outright. Even at the time it was clear that the case was extremely weak, and required the jury to credit things that were virtually impossible and to believe things that were contradictory. It required the jury to accept as genuine confessions made by people who appeared in court covered in bruises. Nevertheless the jury did believe these things. Very possibly they believed them because of the way the case was summed up by the judge. Although a summing up is supposed to be impartial, it is clear that a judge often has a good idea by the end of a trial what the outcome is supposed to be. Frequently this is only to be expected - when either the defence or the prosecution is so weak as to have no realistic prospect of success, it is hardly the job of the judge to level the scales. However, it is widely believed (although I know of no empirical studies) that judges are more apt to believe police evidence than they should do, and that they pass on this bias when summing up.
Jurors may be out of their depth or gullible, but they can hardly be condemned for these very human failings. More worrisome are the allegations of racism and bias that have leaked out. Under English law it is illegal to attempt to impeach a jury decision, and even judges are forbidden to enquire into the way a jury reaches its decision. This is for a very good reason, as supporters of, say, Clive Pontin, will probably admit. Pontin was acquitted by a jury of leaking information about the Government's involvement in the sinking of the Argentinian ship the General Belgrano. His defence, that it was in the public interest that this information be made available, was not a defence at all in law. Nevertheless, he was acquitted, to the consternation of the Government. In short, the invulnerability of the jury is seen by many as a safeguard against the implementation of oppressive legislation.
It does have a downside, of course, in that really bad juries are rarely exposed. Occasionally, however, information does leak out. For example, after the conviction of Sajid Quershi for arson in 2000, one of the jurors wrote to the judge that the other jury members had been making racist remarks about the defendant. Nothing could be done, of course; it was too late.
Successive governments have implemented a number of important measures to prevent further miscarriages of justice. Most notable, perhaps, is PACE, and the codes of practice that accompany it. These were revised only last month to take account of the Human Rights Act. Also important, and to be commended, was the establishment of the Runciman Royal Commission in 1993 and the Criminal Cases Review Commission that it recommended.
On the other hand, many of the reforms to the criminal justice system have increased, rather decreased, the likelihood of erroneous convictions. Only a few examples are presented here.
The current Government has introduced a new Bill based in part on the recommendations of the Auld Report on the criminal justice system. The Government appears to have rejected recommendations that would result in more innocent people being acquitted, in favour of those that increase the likelihood that the guilty will be convicted. This, of course, is it's stated intention; the Bill is designed to `re-balance the criminal justice system in favour of the victim'. Here are a few highlights.
The Bill does not include Auld's proposal to allow scrutiny of jury deliberations, but to its credit the Government did drop proposals to limit the right to jury trial in general.
Many causes of miscarriages of justice could be eliminated, at least
at a cost. It would be possible to regulate forensic science
laboratories more effectively; to investigate police misconduct more
diligently and punish it more severely; to increase the funding of the
CCRC; to recruit juries with better science and maths skills, and so on.
Arguably all these things should be done. Will this eliminate
miscarriages of justice? I would suggest not, although it would
reduce their number, which is no bad thing.
Why can't we eliminate
miscarriages of justice completely? I think the answer is that
we don't have a magic way of knowing what's going on inside the
defendant's head. Ultimately we have to rely on evidence, and its
interpretation by the judge and jury. There is rarely perfect,
definitive evidence in a criminal trial. There will be gaps,
assumptions to be made, and judgement to be exercised. Since
this is the case, we come back to
Blackstone: ``It is better to let
ten guilty men go free than to incarcerate one innocent man''.
|
|
Law glossary index
©1994-2006 Kevin Boone, all rights reserved