Back in January I mentioned a friend of mine who discovered she’d been the victim of a mistake by the CRB. She does voluntary work with a play group organised through her church and they finally got round to insisting all volunteers had a CRB check.
Rather to her surprise, she discovered she’d been convicted back in the 1980s for Taking a Vehicle Without the Owner’s Consent, Excess Alcohol and No Insurance, which she was sure she’d have remembered if it had happened. Then, which really annoyed her, the following day a snotty and sanctimonious letter arrived from the national organisers of this children’s group suggesting they weren’t really sure she was suitable.
Simple enough mistake, it seemed. Same (rather unusual) first name and, similarly rather unusual, surname as her maiden name (she was, at the time of the offence, using her married name), though different middle name and about 10 years’ age difference.
Anyway, I said I’d post again to say how she got on. I saw her the other day and it’s, at last, resolved — 5 months later. This involved her arranging to be fingerprinted at her local police station, which apparently took several weeks to set up. Not quite sure why this was, though it may have been that they had to find an old ink roller and pad fingerprint kit (and someone who knew how to use it, I suppose) of the sort used a the time of the original offence, rather than use the digital equipment now in use, to make the comparison easier.
That, I think, took a couple of months, and now she’s only just been told, not that they’re sorry they made a mistake but that they can’t find the original file so they can’t actually prove it was her. Which seems somewhat grudging to me — no apology, and it took quite a bit of pushing from her to get Thames Valley (her local police area) to tell the CRB people to issue a new, clean, certificate rather than for her (or the charity, rather) to reapply.
She’s not at all happy about this, and who can blame her? What particularly irritates her, as she said, is that throughout she’s been telling them that the Excess Alcohol conviction would have led to an automatic driving ban, so here’s her licence and why didn’t they ask the DVLA if it had ever been suspended? Since the DVLA would — one hopes — confirm it hadn’t, nor ever had any endorsements on it, that would, surely, prove it wasn’t her who had been convicted. Not interested, though.
This reminded me of another case of mistaken identity, when some time ago another friend’s son, whose DNA is on the Police Database because he was convicted for common assault (punch-up between him and another drunk 18-year-old late one Saturday night) found himself hauled in for questioning in connection with a burglary, which he’d certainly had nothing to do with. They’d found DNA evidence at the scene of the break-in, though, and it seemed to match his.
He was sure he hadn’t done it, and had, in fact, been at work at the time of the burglary, so they ran a more detailed analysis and, sure enough, it wasn’t his DNA at all. Apparently the way it works is that they do an initial, quick and cheap, comparison, in the hope that whoever they get the match for will, on being confronted with the forensic evidence, realise the game’s up and confess. They only bother with the more detailed, time-consuming and expensive tests if, despite the apparent evidence, he’s still adamant he’s innocent.
The full implications of this become horribly clear when I read Philip Johnston in today’s Telegraph. After discussing the reasons we’ve apparently got
the world’s largest DNA database, half as large again as the rest of the European Union put together. It represents more than five per cent of the population, compared with just over one per cent in the rest of the EU and 0.5 per cent in America
— not only do we keep DNA from people like my friend’s son,
convicted criminals, [but this ] was extended to people who were tried but acquitted, then to people who were charged but never tried and then, last year, to people who were arrested but never charged
and now they’re talking about retaining samples from suspects, witnesses, people whose DNA has been taken for elimination purposes and, if John Reid gets his way, taking and retaining DNA for people committing ‘non-recordable offences’ like dropping litter and speeding (I can’t see it happening for speeding, since it would upset too many people, but who knows?).
Well, if you’ve nothing to hide, you’ve nothing to fear, or so we’re told. Other, I suppose, than someone making a mistake or a sample getting contaminated or even, as happened with my friend’s son, being hauled in and questioned at quite some length by detectives who’re convinced they have what appears, prima facie, to be pretty good evidence against you.
But it’s worse than that. We’ve apparently recently signed up to something called The Prum Treaty. Mr Johnston explains,
This is an agreement to share criminal data across the EU, including DNA.Since Britain has by the far the largest DNA database in Europe, how this treaty operates is of great importance to those whose profiles are retained. Other EU police forces will have access to the UK database on a “hit, no hit” basis. In other words, the British authorities will inform their counterparts in, say, France whether a profile they have forwarded is retained on the British system.
If it is, what would the police in France think?
Let us imagine they are investigating the rape of a girl at a campsite and they take the DNA of all men staying there. Would they not be suspicious of a person whose DNA shows a positive match on the UK’s criminal database, since their own does not have a million innocent people on it?
So you may have nothing to hide but if, like our correspondent now on the criminal database for being beaten up by someone who ran into his car, there may be something to fear. Who knows what would happen to him were he considered a suspect in a serious crime while on holiday abroad on the sole grounds that his DNA was a “hit” on the UK system?
Not Saussure 
Good entry NS. Another aspect of the data sharing is the huge increase in false matches that will presumably ensue. We’re told that false positives are ‘one in a million’ (although I’m not aware of the basis of this claim if indeed there is one), which means in the UK any given piece of DNA will match 60 people, of which 59 will be a false match. If we throw in all of Europe that goes up to about 750 people.
Comment by thesamovar — May 28, 2007 @ 4:15 pm
But perhaps some one more qualified than me could put in a good word for DNA profiling based on the huge number of cases that are cleared up as a result, and even more important to you good liberal folk, the number of innocents who are released based on this new type of evidence. Some sense of proportion is required, I think.
Comment by David Duff — May 28, 2007 @ 4:38 pm
I don’t think anyone is against DNA evidence being used as part of an investigation (e.g. taking DNA from a suspect, to be used only in this investigation). Likewise, innocents being released can be done without a database.
Comment by thesamovar — May 28, 2007 @ 4:48 pm
I have no objection to it being used either, so long as juries are made properly aware that it isn’t necessarily infallible — there is a danger of cross-contamination, certainly, and, as the DNA database grows, the danger of false matches increases. As I understand it, using current matching techniques, if we put everyone in the country on a DNA database, then we’d get about 50 matches for a given sample*.
But, no, it’s certainly a very valuable forensic tool. My objection is to the way we’re, in effect, having a national DNA database created by stealth, with no proper debate about the implications.
That, combined with the disparity between the size of our database and the size of other ones with which we’re sharing data — and the consequent implicatons, as described in Philip Johnston’s article, is what worries me.
*Oops. Sorry, just noticed Samovar made this point. Our figures are different, I see, which is maybe another cause for concern — I was going by what an expert witness said in court recently, and how he knew I’m not sure at all. I do know that if you’ve got two suspects who are related the odds go down dramatically; from memory, I think if it’s first cousins there’s a something like 1 in 850 chance you’ll get a match.
As I recall, come to think of it, the first murder conviction in the UK based on DNA evidence was overturned on appeal because the forensic science expert for the prosecution had quoted the dramatic odds that there was a 1 in a million chance that the sample wasn’t from the defendant (which sounds like a misleading use of stats to me, but anyway) and then, on appeal, the defence produced two equally eminent experts who gave radically different (from the original expert’s figure and from each other’s) estimates. The Court of Appeal decided that, had the jury realised there was so much disagreement between the experts, they probably wouldn’t have put as much weight on the evidence as they probably did.
Comment by notsaussure — May 28, 2007 @ 5:09 pm
The problem with DNA is that our Police are dishonest cretins who will do anything for a conviction (lovely targetism)
Hence they quote (and use in prosecutions) the stupidity about there being one in a gazillion billion trillion billion chance of this being wrong.
This is complete mathematical bollocks of the kind that makes the Sally Clark scenario look like a PhD project. However, it impresses stupid juries.
Two main reasons : (i) it ignores deliberate/accidental contamination and error (ii) it uses the same probability to both make the match and confirm it.
It is, for example, ridiculously easy to put someone else’s DNA at the scene of a crime. ScotPlod refused to have their DNA taken at one point because of this.
Comment by Paul — May 28, 2007 @ 5:52 pm
I was just going by the ‘one in a million’ claim and noting the UK population is about 60m to come up with 60. The one in 50 claim is presumably based on a more accurate model.
Comment by thesamovar — May 28, 2007 @ 6:01 pm
Sorry, but I have to disagree there, Paul. For one thing, it’s the Forensic Science people, not the police, who give expert evidence on matters such as DNA, and, for another, I don’t accept your characterisation of police evidence in general (and I do see rather a lot of it in my line of work).
In fact, the last case in which I was involved where DNA evidence played a part ended up with the guy’s acquittal, even though he was almost certainly guilty, precisely because Forensic Scientist was honest enough to admit her lack of certainty that the DNA was from the defendant rather than from one of his associates (I won’t go into the details, because the charges were somewhat complicated, but it essentially was a robbery that went wrong, though fortunately no one was too badly hurt).
I think what must have happened in the jury room was that the admitted uncertainty over the DNA evidence — which we know from programmes like CSI is supposed to be infallible — over-rode all the other evidence that pointed to it being the defendant. But who knows?
The fact that prosecutor was spectacularly bad and the defence barrister is one of the best guys on the circuit and was on top form that day must have had something to do with it, too. But even he was gobsmacked when the jury returned a not guilty verdict.
Comment by notsaussure — May 28, 2007 @ 6:13 pm
May I go off-topic for a moment because I need some legal advice.
In a court case last week, two men and a woman who imprisoned a man for four months and gave him regular beatings, had their murder charge dropped by the CPS. Apparently the man was an epileptic but an expert in epilepsy stated that it was very unlikely that epilepsy was the cause of death, *however*, he could not rule it out entirely; and that was the reason the CPS dropped the murder charge.
In my view, even if the epilepsy did kill him, the fact that he was held prisoner and was unable to reach his medicine or medical assistance, makes his captors liable to a charge of murder. What am I missing here?
Comment by David Duff — May 28, 2007 @ 7:11 pm
Well, what the cause of death? Since it’s highly unlikely the epilepsy caused it, that’s neither here nor there.
Briefly, murder is when D unlawfully kills V with an intention to kill V or to do V really serious harm.
From what you’ve said it sounds like manslaughter, which can be committed in one of four ways:
(1) killing by conduct that D knew involved a risk of killing or causing serious
harm (‘reckless manslaughter’);
(2) killing by conduct that was grossly negligent given the risk of killing
(‘gross negligence manslaughter’);
(3) killing by conduct taking the form of an unlawful act involving a danger of
some harm to the person (‘unlawful act manslaughter’); or
(4) killing with the intent for murder but where a partial defence applies,
namely provocation, diminished responsibility or killing pursuant to a
suicide pact.
Comment by notsaussure — May 28, 2007 @ 7:45 pm
“The problem with DNA is that our Police are dishonest cretins who will do anything for a conviction”
Gee, thanks Paul. Please remember to call the RSPCA next time you need to report a burglary… As has been pointed out, it’s the FSS and the other boffins who do the complicated stuff, we just look at the pretty pictures….
The “problem” with DNA is that, contrary to assertions so far, the reliability of a match is improving all the time. I recall a case from the dim and distant past, where a guy was suspected of burglary on the basis of a fifty thousand to one hit, but a further sample eliminated him under the newer technique available. That newer technique is now very old news. I vaguely recall challenging a representative from the FSS on the issue about ten years ago, using the obvious logic about cousins or half siblings coming up as false matches. Her response was that the quoted statistical probability for a false match was considerably lower than the reality. Anyone identified as a suspect from an old sample gets to have their sample rechecked under the newest system available – a process which results in plenty of people being cleared from involvement, as in the original post. A near match is no match.
The old SGM profiling system (under which the son in the post was probably sampled originally) used, to the best of my recollection, the identification of 8 pairs of sites of variance. The current SGM+ system uses 20/22 pairs of sites, making a misidentification much less probable. Some sites have only a few known variations, but others have 20 or so, which gives rise to the unfeasibly high figures quoted currently. The maths is sometimes quite staggering, but unless you’ve got an identical twin brother who follows you around, the chances of someone else with the same profile being in the vicinity of both you and the victim around the same time is somewhat slim to say the least.
A contaminated profile is usually flagged up by FSS, who tend to place their own scientific credibility ahead of any need to assist in an enquiry. Like the CPS, they are notorious fence-sitters, but provide a valuable check and balance to the enthusiasm and faith in technology displayed by many officers.
A colleague of mine is working on an enquiry which has very little to go on, other than the DNA deposit left on the victim. One of the variations is so rare that it is only seen in about a dozen or so of the millions of people on the national database, and none of them is anywhere like a good match to the rest of the sample. The likelihood is that it is a distant relative of some or all of them, given the rarity of the relevant section.
If, as the result of research, they manage to identify an individual and get a full match, they will still need to identify the suspect’s movements etc. for the relevant time, before being able to even consider a charge. The task is not undertaken lightly.
Presumably, your argument against the validity and admissibility of DNA evidence means that I should suggest to him that they forget the whole thing, since you have no faith in it. To be honest, if it helps to catch the offender, very few people will give a rat’s arse what you think.
Oh, and we’ve been checking DNA profiles for serious crime against overseas databases for years – there have been some notable successes, but usually the response is negative. Send us a swab of your own – I’ll arrange for it to be left at the scene of something appropriately seedy ;-)
Comment by maneatingcheesesandwich — May 29, 2007 @ 1:02 am
I am grateful, ‘Saussure’, for your masterful summary which confirms my worst suspicions. In my view 1) and 3) are murder and the law is, yes, you’ve heard it before, an ass!
I am also grateful for the contribution from ‘Sandwich man’ above who has injected some much-needed fresh air into this fetid, alarmist non-topic which seems to have all you liberals fainting with shock.
Comment by David Duff — May 29, 2007 @ 8:13 am
Shorter Duff: threatening to frame people who disagree with you is exactly what a police officer should be doing!
Comment by Alex — May 29, 2007 @ 9:39 am
Given the level of Home Office red tape, demands for statistical analysis and general governmental interference which hampers the day to day duties of the police, your fears of a police state are unfounded. If anything, we’re becoming a state state, which explains the state we’re in.
Oh – and as for framing people, we haven’t got the time…. we’re too busy fiddling the statistics…..
Comment by maneatingcheesesandwich — May 29, 2007 @ 9:53 am
Mr Duff, you may be surprised to learn that the Law Commission — from whom I took the summary of the law, so congratulations are owed to them, not me — pretty much agree with you. Their recentpoposals on reforming the law on murder and manslaughter (press summary here, which proposed introducing first and second degree murder, run on similar lines.
ManEaingCheeseSandwich — I hope you don’t think I’m questioning the usefulness of DNA evidence. My primary worry is about getting sleepwalked into a national DNA database — about which I do have grave reservations — by stealth.
I don’t know what system my friend’s son’s DNA was matched under (this was 2 or 3 years ago), and I can quite see the logic of doing a quick and dirty initial match to see if any of the people to whom the DNA potentially belongs are inclined to confess before you undertake a more time-consuming, rigorous and expensive analysis. But it can’t have been a particularly discriminating test if it identified at least two people in the same area (apparently the actual burglar was also on the database). If we have a national database, how frequently am I likely to find myself pulled in for questioning because some traces of DNA that might be mine have been found at at crime scene, and how many other innocent people are likely to fall under suspicion from the same sample?
Yes, I know that a more rigorous examination of the DNA will exonerate me, but I would like to know how frequently I’m likely to fall under wrongful suspicion if we have a national database. Can anyone tell me?
My other worry is about what happens in court when there’s DNA evidence and the matter’s still contested. In my — admittedly limited — experience of such cases, I worry it confuses the jury as much as it helps them.
OK, the case I mentioned is much on my mind, but I’m sure that if we hadn’t got bogged down in discussions about how sure the expert was it was the defendant’s DNA on the baseball cap, and could she be sure that this wasn’t the baseball cap he said he’d left in a pub the other night that someone must have picked up, since there appeared to be traces of someone else’s DNA inside it, too…. and just concentrated on more comprehensible evidence, he’d have been convicted (and rightly so, to my mind). Presumably it works the other way round, too, and sometimes juries get confused into convicting because they place too much weight on the DNA evidence.
Comment by notsaussure — May 29, 2007 @ 3:05 pm
[…] Saussure has an article on what happened to his friend when a mistake was made with her CRB check, the outcome of which […]
Pingback by Nothing to hide nothing to fear? « UK Liberty — May 29, 2007 @ 4:22 pm
‘Saussure, whilst I am grateful to you for your source, the whole notion of ‘murder 1’ and ‘Murder 2’ fills me with gloom and stands, in my mind, at least, as yet another example of lawyerly make-work. Simpleton that I am, to me, murder is murder and the culprit should hang.
Comment by David Duff — May 29, 2007 @ 5:12 pm
Re use of DNA and databases, I wonder if part of the problem relates to a culture of faith in experts. When someone who has been trained in DNA profiling asserts in court that the odds of being wrong are “1 in a million” (say) are jurors able to assess the veracity of this claim, compared to less scientific matters (e.g. whether someone could have got to the scene of the crime from where they last were)? Are there any lessons to be learnt in this connection from Roy Meadows’ assertion about the statistical likelihood of cot deaths rather than murder?
Comment by heraklitus — May 29, 2007 @ 5:39 pm
NS, last comment was me – was logged in with a WordPress ID which I’m not using any more – apologies.
Comment by Heraklites — May 29, 2007 @ 5:44 pm
Your friend should sue Thames Valley Police. I know she was doing voluntary work but surely she could still get 5 months of lost wages (perhaps paid to the charity for which she should have been working?) plus damages to her character and reputation.
Comment by The Cynical Libertarian — May 29, 2007 @ 6:24 pm
Mr Duff, the problem is deciding what’s murder, what’s manslaughter and what’s a horrible accident. I won’t try and rehearse the Law Commission’s analysis and argument here — you can read it yourself and see what you make of it — but essentially they’re trying to bring some sense and order to 50-odd years’ worth of Court of Appeal and Lords’ decisions, most of which seemed reasonable enough in the particular case but which, taken together, have left us with a bit of a complicated and illogical state of affairs.
Heraklites, you make a very good point, I think. You’ve really put your finger on one of the main reasons I feel uneasy with expert DNA evidence — as you say, the jury just can’t assess, much of the time, the evidence in the way it can other forms of evidence. I don’t know that there’s a simple solution — after all, if it’s evidence, it’s evidence, but I share your unease about, in effect, a jury being told, ‘I’m the expert and, take it from me, he did it’ (or, as in the case I was talking about, being left so confused that they decided there was reasonable doubt, which I don’t think they would have done had they paid more attention to the rest of the evidence).
Cynical Libertarian — I don’t think it was Thames Valley’s fault. CRB’s maybe, or possibly that of the police up in Lancashire, where the offence of which she was wrongly accused took place. Since it was voluntary work, though, I don’t think lost wages come into it!
Comment by notsaussure — May 29, 2007 @ 6:47 pm
“I would like to know how frequently I’m likely to fall under wrongful suspicion if we have a national database. Can anyone tell me?”
The old SGM tests looked at 8 or so sites of variance. I don’t know how many there are to choose from, but I guess that those were the only sites they could clearly identify with the right degree of certainty. People arrested during the period of SGM tests had their swabs analysed and classified to that standard.
The newer SGM+ tests use 20/22 sites (as compared to a standard paternity test, which uses about 15 pairs and tends to satisfy both the civil courts and daytime TV presenters..) and the process goes something like this:
Crime occurs, swab taken from scene is analysed and reveals a full profile (I’ll use 20 sites). The profile is compared to the database and hopefully throws up a full match (20/20). I’d want to know where he was at the time etc., but he’d be Prime Suspect.
If no full match is obtained, the database will throw up partial matches. Lets say it throws up one at 18/20, two at 16/20 and a hundred at 13/20. Subject at 18 is likely to be closely related to the “offender”, probably a sibling. Subject at 16 might also be a sibling, the others may or may not be related. The important thing is that, since they are not 20/20 matches, none of them are the offender and will not be treated as suspects. End of. The siblings of 18 and 16 might get a bit of research though.
The partial list will also throw up all those people sampled at SGM standard whose 8 site result looks like it might match the 20 site scene sample. There could be dozens of them. The investigation team and forensic scientists have to decide whether to (a) re-test whatever remains of the original 8-site swab, to produce an upgraded profile, or (b) visit those people and obtain a new sample. Either way, none of them are suspects unless a 20/20 hit comes back for them.
The sticky bit is where the scene sample doesn’t throw up a full profile, but instead throws up a partial profile where perhaps only 14/20 sites are clearly identified. In that scenario, there is obviously greater scope for Uncle Tom Cobbley and all to be a good match to the scene profile. Neither the police nor the FSS are keen on these….
The scene profile could also be a mixed profile (more than one profile in the sample), whereupon the first step would be to establish whether the victim or other known individuals’ DNA can be eliminated from the mix, before trying to ascertain who’s left.
And yes, you can also get the stray profile – one that shows no match on the system and is on something that’s out of place. Usually a cigarette butt, usually a short distance from the scene. If you’re lucky, it’ll have been dropped by a copper, who can be ordered to give an elimination sample. If you’re unlucky, it’ll have been dropped by a fireman, who will tell you to whistle for it, having been advised by his union that his sample will (a) be searched against everything else we have outstanding – which it won’t unless he consents for that to happen, and (b) be treated by his bosses as an admission of smoking at the scene, contrary to Regs. Either way, the case will become skewed by this stray profile, the DNA equivalent of a man on a grassy knoll.
So, in response to your question – if the scene yields a full profile, you’re extremely unlikely nowadays to fall under “wrongful” suspicion if neither you nor your identical twin have any connection to the scene. If the scene yields a partial profile, it’s statistically more likely, but the likelihood markedly decreases if your profile is of the 20-site rather than 8-site style.
In any event, you are statistically more likely to be hit by a bus, robbed by a “hoodie”, kicked to death by a donkey, or flattened by a falling piano than you are to be seriously inconvenienced by the DNA database.
84.732% of all statistics are made up just to frighten us. Don’t worry, be happy.
Comment by maneatingcheesesandwich — May 30, 2007 @ 12:44 am
The other problem with a big DNA archive is that when you start doing familial matching on it, you run into ethical problems. As a hypothetical example, try this:
A crime is committed which leaves DNA evidence. The police test this, and find that they have someone on file who is a 50% match for it; i.e. he’s the father of the person who left the DNA.
So, they go haul in his children and test them, but come up blank, so none of the kids are arrested for the offence.
Later on, the family read in the local paper that someone’s been convicted for the offence, based on the DNA.
They don’t realise, until the local pub smartarse hears the tale, and starts giggling into his beer.
If (as he explains) the initial hit was on the father of the criminal, and none of the father’s official kids match, yet a conviction ensues because of the DNA hit, then this means daddy must have _another_ kid by some other woman.
Cue arguments and possible divorce, and suing of police for being insensitive in the manner by which they broke the news to the man and his family…
Comment by Dr Dan H. — May 30, 2007 @ 9:19 am
Dan H
“they have someone on file who is a 50% match for it; i.e. he’s the father of the person who left the DNA.”
Thankfully, it tends to be difficult to identify exactly which 50% comes from the father or mother, until you have mum, dad and baby all together. In it’s simplest terms, if mum is 1234 and dad is 5678, child could be 1278, 3456 or any other combination which takes two from one and two from another. Another couple, who are 1698 and 2357 could also throw out 1278. If your suspect is 1278, it won’t be obvious who the parents are, since very few of the sites are sex-specific.
Thankfully it’s more complicated, but the list of potential parents normally runs into the hundreds, since you don’t know who either mum or dad are. In civil paternity cases, you tend to be able to show who mum is, so profiling dad is relatively simple, as you already have two out of three available.
To set your mind at rest a little more, one case I’ve dealt with in the recent past had an unknown offender but a full profile. The idea of familial searching was thrown into the ring and seized with gusto, based on recent high profile cases. We ended up with no reasonable sibling matches (nearest being 12 or 13 out of 20) and about five hundred possible parents or children of the offender(ie 10/20 or thereabouts, plus a few other telltales). Needless to say, it didn’t look like the most efficient line of enquiry. Thankfully there were other avenues to pursue.
As with old fashioned fingerprinting, there is always room for improvement. I would place as much reliance on full profile DNA evidence as I would on 16-site fingerprint evidence. Both tend to be more consistent than eye witness evidence.
If three witnesses see Man A stab and kill Man B, then run away, the chances are they won’t agree on his clothing, what the weapon looked like, nor which way he went. Finding a knife in a nearby bin, with the victim’s DNA in blood on the blade and the offender’s fingerprints or DNA on the handle, would be good enough for me, whatever he looked like (or wherever his mum says he was at the time). To satisfy the courts, we’d still need to show a raft of other evidence, as I’m not aware of any case where the suspect was convicted on either DNA or fingerprints alone.
Scientific evidence is just that, evidence. It is not proof in itself, but requires supporting fact. If you have no faith in scientific evidence, then it’s a good job the courts still do, otherwise there’d be an awful lot more rapists and murderers walking the streets, as quite often you’ll only have one person’s word against another – and dead men don’t tell tales.
There will always be mistakes or over-emphasis of statistical accuracy, but where do you draw the line? The discrediting of the evidential weight of Roy Meadows’ “expert opinion” has resulted in some spectacular acquittals on appeal, but an awful lot of convictions in which his evidence was a contributory factor remain sound. You don’t hear much about them though. Whose word would you trust as proof of fact in itself? The Pope ? The Prime Minister ? Your next-door neighbour ? And could they trust yours ?
Comment by maneatingcheesesandwich — May 30, 2007 @ 11:12 am
I might be speaking for many of the readers of this thread in saying how interesting and informative it is. I am particularly grateful to the ‘Sandwichman’ for his clearly expressed expertise, and our host who started it all.
Comment by David Duff — May 30, 2007 @ 4:15 pm
For once I agree with Mr Duff, at least about how interesting informative this has become and particularly grateful I am ManEatingCheeseSandwich. My fears are, to a certain extent, assuaged, though I’m still worried both about the principle of what looks like a national DNA database being introduced by stealth — if we had a proper debate that left everyone giving their informed consent, that might be different — and about the way juries are asked to make judgements on evidence they often don’t properly understand.
In point of fact, most of the time DNA evidence is mentioned in court it’s in the context of how the defendant initially denied the offence at interview but, on being confronted with the DNA evidence, made full admissions. It’s when matters are disputed I start to worry a bit.
The two problems I still have, though, are that, based on the experience of my friend’s son, there were enough people on the database, for whatever reason, in a not particularly large town to generate at least one partial match significant enough for the police to interview someone against whom there was no other evidence. Fortunately for him, he hadn’t done it, had a good alibi and the culprit soon pleaded guilty, so he suffered no more than a comparatively minor inconvenience, though one wonders, rather, how it would have gone for him had there not been another suspect, had he not had an alibi, and had it been a rather more serious crime than a small domestic burglary. Doubtless further forensic investigations would have exonerated him eventually, but I wonder after how long.
I’m not completely clear whether — which was my impression — they did a 8 point test to see if anyone would confess and spare them the trouble and expense of conducting a more detailed one, of whether they only had a partial sample with which to work, or what.
My other, more general concern, is how do other countries manage without such a database. Our DNA database is apparently half as large again as the rest of the EU combined. Are the clear-up rates for rapes and murders in places like France, Germany, Scandinavia and the like dramatically worse than ours? I ask because I do not know.
Comment by notsaussure — May 30, 2007 @ 4:38 pm
One of the main issues with the DNA database is that people who have never been convicted of a crime are on it. The only way that is justifiable is if everyone in the country was on the database – and that’s surely not OK.
Comment by thesamovar — May 30, 2007 @ 8:18 pm