Barry George: an error of the first kind

The case of Barry George, released yesterday after a retrial jury cleared him of murdering television presenter Jill Dando, has a lot of resonances.

There are, of course, many wrong convictions. It will never be otherwise while twelve human beings have to do their best to become "without reasonable doubt" on the basis of the messy and inconclusive evidence which a real world can supply. I have every sympathy with juries, including the one which did its best when wrongly convicting George. When called to jury service myself, a few months ago, I prayed for a trivial offence where consequences of mistakes in either direction would be as minimal as possible. But this wrong conviction rings particularly loud bells.

On the face of it, George was convicted on the basis of a single miniscule particle of what might have been gun residue, in the pocket of his coat, a year after the murder took place. As always in such technical issues, the jury were completely dependent not on this evidence itself but on what they were told about it by experts and authorities. The discrediting of that scientific evidence is what led to his retrial. The rest of the case against him was circumstantial, the strongest piece of the jigsaw being that someone saw him on the same street some hours earlier.

In practice, though, George was convicted on the basis of his character. He was convicted of being a sad git. He was convicted of being someone known to approach and hassle women on the street. He was convicted of having already served a jail term for attempted rape of one neighbour, and assault on another. He was convicted of being a fantasist about fame and weapons. He was, in other words, convicted of being an unpleasant person in many ways, and someone we didn't much like or feel comfortable about in many more – but not, so far as anyone outside the courtroom could discern, of being a credible single shot assassin capable of killing and disappearing in full daylight then erasing any trace of his crime.

He was also convicted for a less clear cut reason. Jill Dando was young and widely popular figure, and the public wanted to see a killer convicted. There was something of the Princess Diana effect around her, and a year with no perpetrator on trial bred similar speculation. Open, friendly, popular, photogenic blonde female celebrity on one side; unattractive local weirdo on the other; no contest.

There was another trial news story yesterday: dismissal of a jury failing to reach a verdict on three Muslim defendants from Leeds over complicity in the London Transport bombings of 7th July 2005. Here again, there is the combination of a crime over which public demand for a conviction of some sort is high and evidence which is circumstantial. So far as I or any other lay observer can tell, the case hinges around the fact that these three defendants knew the bombers, and whether or not that makes them accessory conspirators.

Mr George has spent eight years in prison. The news media talk of him now having gotten "justice". It's difficult to see how someone who has spent eight years in wrongful imprisonment can ever be said to have received justice.

I repeat my sympathy with juries in such cases. I extend that to everyone involved in the difficult job of trying to find out the truth in an imperfect world, from the police through the lawyers to the courts, along the chain which brings the case before that jury. I am not asking for someone to be blamed – that would be to repeat exactly the mistake which led to this miscarriage of justice. I am, however, asking that we think carefully about the changes in judicial systems which are creeping through western societies, especially the UK and the US.

In the UK, for example, we have seen fairly recent partial removal of the double jeopardy principle. Longer ago, the acceptance of a majority verdict of the type which allowed 11:1 conviction of Barry George when unanimity wasn't reachable. Also recent is the extension to 28 days of the period for which suspects can be held without charge ... and ahead of us lies the prospect of further extension to 42 days.

I am not a rosy spectacled believer in the inherent goodness of every defendant arraigned before a court. It troubles me deeply that many perpetrators of vicious crimes are released to find new victims, because the fairness of the law cannot find sufficient evidence to convict them. Here, I sympathise with the police. But it also troubles me deeply that some have their lives destroyed when they are innocent, and I do not believe that extension of that should be undertaken lightly.

If forced, reluctantly, to the choice, I would prefer than a guilty person go free (what I would, with my statistician's hat on, call a "type II error") than that an innocent person is convicted (type I error). The best estimates are that in the UK justice system type II errors are far more common than type I errors; I would like to keep it that way. Each of the changes mentioned above, however, reduces the likelihood of increases the likelihood of type II by increasing the probability of type I. (And, of course, a type I error automatically brings with it a type II as well: the wrongful conviction of Barry George meant that the real killer also went unprosecuted. But leave that complication aside for now.)

The problem is this: you don't have to be convicted of a crime to become the victim of what I might call a "hidden type I" error.

Think about your own life for a moment. Imagine that the police arrive at your home, or your place of work, unexpectedly by definition, and arrest you on suspicion of murder, or involvement in terrorism. You are completely innocent, but the police have good reason to mistakenly believe otherwise.

You are held for twenty four hours; that's usually the limit, to be fair, in most cases, but it can be more if the system feels it needs more time. Forty eight hours. Seven days. Perhaps four or, in the future, six weeks. That's one day, one week, one month, six weeks, away from your daily life. Now you are charged, and you spend months (as Barry George did) still innocent in the eyes of the law yet in prison awaiting trial. Then there is the trial itself - anything from days to months. Finally, you are (unlike Barry George) acquitted and released.

Now ... at what point in that story would your life be irrevocably changed?

I've worked inside prisons, in the past and, although I'm not allowed to talk about it, I can say that in my own case a single week of detention in there would change me forever. But for simplicity put that, also, aside ... think, instead, about what is happening outside.

Twenty four hours is enough for your family to feel the eyes of neighbours looking at them in a new and speculative way. It's long enough for your colleagues at work to wonder about you. It's long enough for journalists to make life difficult and stressful at both home and work.

A week is long enough for rumours to grow and take root, for your employer to feel irritated by your absence, for your children (if you have them) to become victims of peer cruelty at school from which they will never emerge.

A month, and perhaps bills are going unpaid. Three months and your employer has had to make arrangements for your possible nonreturn; even if you keep your job, you will forever be known as the one who was suspected of... Six months, and how many of the pieces will still be there to pick up? You are acquitted, you are declared innocent, but nothing can ever give you back the life you used to have. You are a type I error casualty despite acquittal.

Some years ago, a colleague working one day a week inside a prison education wing was accused on a Friday of having had an affair with a prisoner. On the Monday, two days later, both her education employer and the prison authorities declared the accusation to be not only untrue but completely impossible (she spent her prison hours in a glass walled office with busy corridors on all four sides). On the Sunday in between, however, a national "newspaper" had run a salacious four page centre spread feature on the accusation, with implicit assumption of guilt. Her family life was destroyed; she found it impossible to live down the assumptions that there is no smoke without fire, and her career collapsed. She was a type I error casualty despite there being no hint of any official suspicion ... imagine if she had been arrested on a charge of murder or terrorism?

And the cases of type I victims who are "extraordinarily rendered" to other jurisdictions where torture on spec is the norm don't bear thinking about, of course ... so those, too, we will put aside for now.

Barry George, now acquitted, is not unique. Nor are Waheed Ali, Sadeer Saleem or Mohammed Shakil, still in custody pending retrial on charges of assisting the 7 July bombers. But George, in particular, does stand as a fingerpost warning us of the perils inherent in extending tools available to the state. There is no need to paint unjustly villainous pictures of police or legal profession to fear what such extensions can do to civil society. Honest human beings, doing their imperfect best in an imperfect world, will inevitably generate more collateral type I miscarriages of justice, of greater magnitude, as they seek to make best use of tools which seem, on the surface, unexceptionable responses to serious crime.