Forensic document examiners have been asked to analyze questionable signatures for a long time. In legal situations all over the world, they are often called to courtrooms to determine whether a signature is genuine, disguised or forged. But until now there has been little research that not only demonstrates the validity and reliability of the methodology these experts use but also examines whether or not the conclusions made by the experts are accurate.
Faculty and students in Kentucky State University’s Division of Behavioral and Social Sciences have recently concluded research that will likely be referenced by many forensic document examiners and attorneys when expert testimony about the identity of a signature’s author is at issue.
Dr. Mara Merlino, an associate professor in the division, and Dr. Tierra Freeman, acting chair of the division, led the study made possible by a grant from the National Institute of Justice. They are now busy preparing to present the research and educate experts in the field.
“It’s cutting edge,” Freeman said. Also, she added, “Our students did this.”
Freeman said the CIA expressed interest in one of the research students at a job fair, and another student has his eye on future employment at the FBI.
Students operated a state-of-the-art eye-tracking system in the division’s lab. Participants in the study, including forensic document examiners and a group of individuals who do not work in the field, were asked to look at approximately 120 questioned signatures during four different experimental protocols, without any tools such as microscopes, while an infrared light tracked retina movement.
In one protocol, participants were shown a screen with five signatures. Four of the five signatures were known signatures of an individual and one served as the questioned signature. Participants’ eye movements were tracked while they examined the signatures and made a decision about whether or not the questioned signature was written by the same person who wrote the four known signatures.
The eye-tracking system created a video that showed where the participants’ gaze rested on the signatures, the length of time they spent looking at that spot and the sequence in which they looked at the features. The participants were also asked to look at a single signature to determine if it was genuine, disguised or forged. And they were asked to make the same call in a protocol that only showed them a signature for a single second. The accuracy rate for the forensic document examiners in all these protocols was about 70 percent.
Merlino said the next study would allow the forensic document examiners the usual tools they use to examine signatures instead of just the naked eye.
“I would predict that their accuracy would bump considerably when they are viewing original signatures with the techniques and equipment they ordinarily use in their labs,” she said.
There are many features to look for in a signature, including the orientation to the baseline and the slant of the letters. It’s also possible to tell where the pen moved the fastest and where the pen may have stopped.
“If it is a fluidly written signature, the speed is going to come in the upstroke,” said Merlino as she described one signature.
People who try to disguise their own signature might change the first letter but revert to their own style and slant, Merlino said. This knowledge would be useful if people disguised their own signature on a Social Security check, then claimed to have not received it, for example.
In recent years, DNA evidence has exonerated people convicted of crimes because in some cases forensic evidence was not properly analyzed, Merlino said. The signature study helps ensure the same doesn’t happen with signatures by providing data on the reliability and validity of forensic document examination methods and accuracy of signature analysis.