AS COURTROOM dramas go, Angela Cannings鈥檚 appeal hearing in London takes some beating. Last month, she was exonerated of killing two of her babies. Last week, after judges gave their reasons for freeing her, the government ordered reviews of 258 similar cases, and ministers plan to check thousands of cases in which children have been removed from their families because of suspected abuse. Why? They fear the scientific evidence in these cases was worthless.
Cannings is the third mother in two years to walk free after being accused of killing more than one of her babies. The prosecution in each case relied on testimony from retired paediatrician Roy Meadow. He gave his name to 鈥淢eadow鈥檚 law鈥�: one infant death in a family is tragic, two are suspicious and three are murder. He also championed Munchausen鈥檚 syndrome by proxy, a bizarre disorder in which parents harm their children to gain attention.
Yet neither his syndrome 鈥� which has been used to remove children from supposedly abusive parents 鈥� nor his law have stood up to scrutiny. The syndrome is poorly defined and its diagnosis loosely applied. His 鈥渓aw鈥� crumbled when it came to light that some families carry genes that seem to make sudden infant deaths more likely than usual.
Advertisement
Meadow鈥檚 theories spread widely, so these events are likely to be replayed in other countries. But it would be wrong to blame Meadow alone for what has happened. The courts have been gullible, and have not been well served by forensic science. This applies even in the US, where 鈥淒aubert hearings鈥� are supposed to separate the pseudo from the science (see 鈥淔orensic evidence stands accused鈥�).
There are still large areas of forensic science that have not been subjected to systematic study and open review. Big questions remain, for example, over techniques for examining hair, handwriting and even bullets. Worse, we do not even know the rate of errors made by experts trying to match fingerprints. Official reluctance to allow greater scrutiny is understandable, as doubts raised would be like manna from heaven for defence lawyers. But it also means that for many people, justice is based on wilful ignorance.
Fingerprint examiners still talk of zero error rates, a notion that makes no scientific sense. They seem to fear the idea of probabilistic results, despite the success of DNA profiling. Every DNA profile comes with a built-in health warning 鈥� the odds that it belongs to more than one person. Every aspect of the technique has been openly debated by mainstream scientists. As a result, it is now probably the most trusted and persuasive of scientific evidence.
Consider this. In London last week, a judge freed Mark Dallagher on appeal after seven years in prison for murder. His conviction rested on an ear print left on a window, which an expert witness identified as Dallagher鈥檚. He backed his assertion by arguing that everyone鈥檚 ear print is unique. His evidence was eventually undermined by DNA taken from the ear print, which turned out not to be Dallagher鈥檚. Which technique would you bet your freedom on?