By Rich Scinto, New Haven Register
NEW HAVEN, Conn. (AP) — Time is of the essence for law enforcement immediately after a large crime such as the Boston Marathon bombing or the mass shooting at the New Orleans Mother’s Day parade. Hundreds of witnesses have to be screened, and some may be hiding pertinent information.
But now, Southern Connecticut State University associate professor Kevin Colwell has developed a forced-choice test that can allow law enforcement officials to scan a group of 30 witnesses within 10 to 15 minutes. It works on the premise that those telling the truth tend to add more details to a story if questioned properly. “Liars don’t, they keep telling the same story in the same order because they don’t want to screw it up,” Colwell said.
The same principles can be applied to test whether someone is attempting to feign incompetence to stand trial or amnesia.
There is usually plenty of time to question a single witness during a questioning session, Colwell said. The forced-choice test is usually between 20 and 40 questions and can be completed when the last person finishes answering all questions. The only manpower needed is someone to administer the test. His studies so far have a 90 percent accuracy rating, he said.
“If you know a little bit of information and have one or two trustworthy witnesses, you use their information for a test to weed out those that are attempting to deceive,” Colwell said. Although he couldn’t give away all the details, lie detection is based on mathematics. “With a pool of people, it quickly becomes evident who is trying to deceive,” he said. “Math lets us know who is hiding information. It’s fairly simple and it works well.”
As the Boston bombings have shown, the quick use of multiple investigative tools can lead to better success, said John DeCarlo, an associate professor of criminal justice at the University of New Haven in West Haven. “Things have really changed considerably over last 10 years with the proliferation of technology in field,” he said. Cameras mounted throughout city streets and smartphones can act as virtual eyes after an investigation. Psychology and its deception detection branch also are powerful tools. “Kevin is taking it to a whole other level,” DeCarlo said. “The whole area of deception detection holds much promise (in law enforcement).”
John Jay College of Criminal Justice associate professor and lie detection expert Maria Hartwig is familiar with Colwell’s work, and said his new test seems promising. “It’s a very sound and simple test, and in all likelihood a reasonably effective technique. It does lend itself well to suspect elimination,” she said. “All it requires is a person is corporative and they take the test, but that’s true for all lie detection tools.”
Many law enforcement techniques for questioning people aren’t backed by empirical science, Hartwig said. “As far as I know, there is no research on suspect elimination techniques, which is quite interesting, I think, because it seems to be a fairly frequent situation where a pool of people has to be narrowed down,” she said. “I’m glad to hear someone is doing systematic work on it.”
She wasn’t familiar enough with the exact methodology of the test to comment on whether it screens for people who are deliberately deceiving versus those who unknowingly withhold information. Colwell said the mathematics behind the test help tell the difference between the two.
The Federal Law Enforcement Training Center approached Colwell two years ago and asked him to develop the test, Colwell said. The test is built on the prior work of Colwell and Anisman on assessment criteria indicative of deception, also known as ACID. “It’s nothing utterly new because we’ve been using this in forensic assessment and clinical assessment since the ’80s,” Colwell said. But the technique should eventually trickle back to local law enforcement, Colwell said. Federal officers have to go to training at the Federal Law Enforcement Training Center and take both basic and advanced courses.