DNA pixie dust fails to solve all UK crime, shock horror

The Home Office has boasted of a quadrupling of detection of crime via DNA technology over the last five years, during which period the UK's National DNA Database has trebled in size, and now exceeds 3 million records. An enthusiastic report from the Home Office's Forensic Science & Pathology Unit (DNA Expansion Programme 2000-2005: Reporting achievement) lists impressive improvements in detection rates, thanks to DNA, which is "a powerful aid to crime investigation."

Under, erm, certain circumstances. The report happily shares the glad tidings of how much better clear up rates are with the addition of DNA (e.g. domestic burglary up from 16 per cent to 41 per cent, and theft from vehicle clear-ups boosted from 8 per cent to 63 per cent), but is rather less forthcoming on the extent to which these welcome improvements can be achieved on real burglaries and nickings of car stereos and MoD laptops. As the report bashfully tells us, "the rate of recovery of DNA from crime scenes remains lower than expected and is a bottleneck at the beginning of the process."

Which means? Luckless inner city inhabitants among the Reg readership will possibly have the impression that if they're burgled or their car window gets smashed, they're a lot likelier to get offers of counselling than an actual crime investigation; historically, it frequently hasn't been cost effective for the police to try to find solvable crimes in among the huge pile of unsolvable low ticket ones, hence the low clear-up rates for them. But, as the Home Office figures show, DNA matching produces dramatic boosts in clear-up rates, so it now makes sense for police to get round to the scene of the crime and collect the samples, right?

Not entirely. To some extent the low rate of DNA recovery the report talks of is caused by the failure of some police forces to use DNA sampling as extensively as they might (the report pulls up ten out of 43 forces for this), but largely it's because, contrary to popular belief, forensic investigations don't always produce a useable DNA sample. Au contraire... In a presentation delivered at the end of 2004, Robin Williams of Durham University's School of Applied Social Sciences reported that 17 per cent of recorded crimes were examined by crime scene investigators, that DNA was recovered from 5 per cent of scenes (accounting for 0.8 per cent of all recorded crime), and that 45 per cent of crime scene profiles matched subject profiles when loaded.

The last number is the sort the Home Office is ecstatic about, but the others are more significant. Obviously these numbers will have changed somewhat over the last 12 months, but you can see the bleeding obvious starting to creep in, can't you? A relatively small percentage of crime scenes get the forensic treatment, smaller numbers still produce a useable result, but where a result is produced, there's a reasonably good chance of the police finding someone to nick. Now, you could consider the possibility that if the police got their fingers out and increased the number of crime scenes they examined, they would increase the number of nickable people they discovered. Which they would, of course - but the lowish hit rate produced from the ones they do already examine suggests that the percentage of positive results would fall as the percentage of crime scenes covered rose. So the question of bangs per buck starts to creep in - more resources could be ploughed into DNA work (and will be, it being a major Home Office enthusiasm), but would tend to produce diminishing returns, and lead one to suspect that the resources might have been better deployed on some more fruitful aspects of crime-fighting. Which perhaps explains why the police's failure to show up for your minor burglary, although irritating, is the sensible thing to do (N.B. Fingerprints are also supposed to be magic, right? So how come, pre-DNA, the didn't dust your house as a matter of course? For pretty much the same reasons, right?).

Home Office propaganda might lead the unwary to believe that DNA is the magic pixie dust that will solve all crime, but if you think about it, it quite obviously ain't so. People leak DNA wherever they go, which is an advantage for the investigator in the sense that it's extremely difficult for a criminal to be absolutely certain they didn't leave a trail, but a disadvantage because of the amount of leaked DNA 'noise' the world is full of. Burglars and car thieves will tend to leave less DNA at the scene than the usual inhabitants, and as their awareness of DNA matching climbs, they'll leave less still, because they'll be more careful. They will also be more likely to leave false trails (e.g., as police have been beginning to note, random cigarette butts in stolen cars), or to plant evidence (note that it's a lot easier to plant DNA than fingerprints).

The Home Office announcement offers us numerous satisfyingly large percentage increases to get the message of the DNA Database's 'success' over, but the bottom line is that the impact of DNA on overall crime is considerably less dramatic, and it's unlikely to get much more dramatic with further expansion of the Database. That expansion, however, is the Database's most unqualified 'success'. When New Labour took office in 1997 the national database (which tended not to be referred to as such, because we weren't building a National DNA Database by stealth) stood at 700,000 records, but by dint of a series of legislative changes it has now passed 3 million, and is intended to reach 4.2 million by 2008. We're certainly building it now, but by increment, not by stealth.

Which is possibly more dangerous. The database currently has records of 37 per cent of black men in the UK, 13 per cent of Asian men and 9 per cent of whites. The Home Office justifies this on the basis that most of the records on the Database are for those who have been charged and convicted of crimes, but as not all of those on the Database have been charged and convicted, that is not a wholly adequate response. If there are large numbers of 'suspects' within particular communities, then the tendency will be for more 'suspects' (on the DNA database, we're all suspects) from those communities to be added to the Database, and distortions will tend to be magnified.

There is, one hopes, a ceiling to the size of the DNA Database that can achieved via current means. As of 1st January, the law changed to make all offences arrestable, and DNA and fingerprints can be taken and retained, forever, on arrest. It is therefore conceivable that littering, or even wrongful arrest for littering, could add you to the Database. But despite this expansion and the large number of new laws and offences the Government has added to the statute book over the years, surely some percentage of the population will remain splendidly unarrested and unsampled, meaning that there will on the one hand be a pile of sampled 'suspects', and on the other a pile of unsampled 'innocents'. We wouldn't care to bet on the likely sizes of these piles, but clearly it's not fair, and the logic (as argued by Alec Jeffreys) is either you delete records or you sample everybody.

If you sample everybody then your general clear-up rate goes up a little more, but not a lot more, and you're maybe going to be able to crack a couple more high profile unsolved crimes from days gone by. There's an obvious price to be paid for this in terms of civil liberties as the technology currently stands, and a greater price further down the road as the technology develops. The public may or may not be willing to pay this price, but even if it is, the question of whether, from the point of view of fighting crime, it's actually worth the money, remains. Quite possibly, it isn't. ®