A new study using handwriting recognition software claims to show that everyone's handwriting is unique and that using such evidence in court is therefore scientifically justifiable.
Handwriting experts have had a tough time in US courts since a 1993 supreme court decision that attempted to clamp down on junk science. According to the ruling, expert testimony can only be admitted if it is based on scientific methods published in peer review journals.
But now, Sargur Srihari and colleagues at the State University of New York at Buffalo have used computer software to show that handwriting samples collected from a representative sample of 1500 people were all unique. The study was funded by the US National Institute of Justice and will appear in the July issue of the Journal of Forensic Sciences,
The researchers aim is to overturn an increasingly common defence strategy in handwriting cases - the argument that no one has proven or quantified how individual handwriting really is.
The study is the most comprehensive peer-reviewed attempt to show that handwriting is unique. But critics say it is irrelevant to the discussion over the admissibility of handwriting evidence.
"The real issue in court is what the likelihood of a random match is," says Michael Risinger, professor of law at Seton Hall Law School. "This research provides zero input on this."
Nonetheless, the work is already making an impact in US courts, such as in the US versus Gricco case, decided two weeks ago in a Pennsylvania Court but currently under appeal. The judge cited Srihari's study as a justification for rejecting defence objections to a handwriting expert due to testify about the writing on forged parking tickets.
"The novelty was showing that handwriting is individual, though not who a sample belongs to." says Venu Govindaraju also at SUNY, but who was not involved in the study.
Srihari's software characterises people's handwriting using 11 features, including letter size, swirls and pen pressure. These are a subset of the features used by handwriting experts. The software is not as good at matching samples as human experts, but it is good at processing the huge numbers of samples needed for valid statistical studies, such as Srihari's.
All comments should respect the New Scientist House Rules. If you think a particular comment breaks these rules then please use the "Report" link in that comment to report it to us.
If you are having a technical problem posting a comment, please contact technical support.
