A so-called lie detector featured in a new ad supporting Herman Cain uses voice-stress technology that is no more reliable than “flipping a coin,” according to one university study by a criminal-justice researcher. Two other academics called it “charlatanry.” The ad says that “one of the foremost lie detector experts in America” found that Cain was being truthful when he said he did not do anything improper, and that one of his accusers was not. But the science behind that claim has gotten mixed reviews, at best, from academics.
After an improbable meteoric rise in the Republican presidential polls, Cain’s standing has dropped dramatically — to the point that he is now reportedly considering quitting the race — after several women accused him of sexual harassment and another claimed to have had a 13-year extramarital affair with him.
Enter the 999 Fund (aka Americans for Herman Cain), a super PAC independent of the Cain campaign, which released an ad attempting to clear the candidate’s name. According to a CNN story, the group is spending just under $100,000 to air the ad in Iowa and parts of Minnesota.
The ad begins with a display of newspaper headlines about the sexual harassment claims as a female announcer speaks:
Announcer: It’s time for the truth. The media won’t tell you what one of the foremost lie detector experts in America said about Herman Cain.
T.J.: Ward: From my exam, he is being truthful … But the allegations of saying that she’s been sexually assaulted by him did not occur.
Announcer: Now that you know the truth, let’s focus on what matters.
The clip of Ward comes from an interview he did with a CBS affiliate in Atlanta. Ward used his company’s Layered Voice Analysis software — which sells for about $15,000 — and analyzed the televised comments from Sharon Bialek alleging sexually inappropriate conduct by Cain, and another from Cain denying that he remembered any such incident.
In the news report, Ward states, “In my exam, taken from his speech, he is being truthful about not having any knowledge about these allegations being brought against him. But the allegation of saying that she’s been sexually assaulted by him did not occur.”
Cain cited Ward’s findings in an interview with Neil Cavuto on Fox News.
Cain, Nov. 11: And here’s the other thing. There’s a private investigator by the name of T.J. Ward out of Atlanta, Georgia who has some sophisticated technology that a lot of people may not have heard about. He took my statement from my press conference, ran it through his software, and was willing to go on record — because many law enforcement agencies use this software — and said, ‘Herman Cain is telling the truth.’
He did the same thing for this woman who accused me the other day when she was with Gloria Allred, and went through and said, ‘I’m sorry, but there were a lot of untruths in that statement.’
Ward’s most recent claim to fame is that he has worked with attorney Lin Wood on the Natalee Holloway disappearance case. Wood is also an attorney for Cain. Ward said his analysis was done at the behest of the Atlanta news program, not the Cain campaign. In fact, Ward said he has had no contact with the Cain campaign.
We spoke to Ward, a private investigator, by phone and he said, based on his analysis using the LVA software, that Bialek’s statements contained “false statements” and that “Mr. Cain was being truthful about Ms. Bialek.”
Ward said he doesn’t characterize himself as “one of the foremost lie detector experts in America,” as the ad did, but he said he is the only private investigator in the country licensed to use Layered Voice Analysis (LVA) 6.50. His website says he and business partner Lynn Robbins, president of Voice Analysis Technologies of Madison, Wisc., have sold the software to more than 70 law enforcement agencies around the country.
In the Atlanta TV report and in our interview, Ward claimed the software was 95 percent accurate. Asked for backup, Robbins pointed us to an as-yet unpublished study conducted by J. Michael Adler, a psychological examiner in Tennessee, that concluded there were “no significant differences in the findings determined by the polygraph and the LVA examinations” when used on a number of sex offenders. In his report, Adler notes that past studies of polygraph examinations have found them to be 93.6 percent to 98 percent accurate. Hence Ward’s claim for LVA’s as well.
But there are at least two published studies that found LVA’s did not perform any better than chance.
In 2006, Drs. Harry Hollien and James D. Hansberger at the University of Florida evaluated voice stress analyzers for the Department of Defense’s Counterintelligence Field Activity agency.
CIFA study, March 17, 2006: The findings generated by this study led to the conclusion that neither the CVSA [Computer Voice Stress Analyzer] nor the LVA were sensitive to the presence of deception or stress.
In 2008, researchers with the Oklahoma Department of Mental Health and Substance Abuse Services performed a study, funded by the National Institute of Justice (NIJ), of two voice stress analysis programs, including LVA. They questioned more than 300 arrestees about their recent drug use, and evaluated the programs’ findings of truth when compared against urine drug test results.
One of the study authors, Dr. Kelly R. Damphousse, associate dean of the College of Arts and Sciences and professor of sociology at the University of Oklahoma, wrote this in an article about the study:
Damphousse, March 2008: According to a recent study funded by the National Institute of Justice (NIJ), two of the most popular VSA programs in use by police departments across the country are no better than flipping a coin when it comes to detecting deception regarding recent drug use.
A review of the software by two Swedish linguists, Anders Eriksson and Francisco Lacerda — in a 2007 article in the International Journal of Speech Language and the Law — was even more damning in its assessment of LVA.
Eriksson and Lacerda, 2007: Indeed, our review of scientific studies will show that these machines perform at chance level when tested for reliability. Given such results and the absence of scientific support for the underlying principles it is justified to view the use of these machines as charlatanry and we argue that there are serious ethical and security reasons to demand that responsible authorities and institutions should not get involved in such practices.
We emailed Eriksson and asked him how much weight people should place on Ward’s conclusions about Cain and his accuser. Here’s part of his response:
Eriksson, Dec. 2: As we show in our paper, the principles upon which the LVA is based have no more to do with emotion or deception detection than throwing a pair of dice. So the people investigating Mr. Cain would have obtained equally “reliable” results by doing precisely that.
Eriksson called the technology behind LVA, “total nonsense” and, he said, results can be easily manipulated by “changing the threshold values” so that “any setting produces a vague set of predictions pretty much the same way horoscopes do.”
Robbins, who sells the LVA program, vehemently contests the study findings, calling them “indisputably disconcerting” and an attempt to “discredit any technology in any way possible to benefit polygraph.” And she and Ward point to many law enforcement testimonials about the software’s reliability and utility.
A study by two Duke professors earlier this year looked at LVA technology for business use and noted that, “Several studies document that LVA algorithmic metrics for detecting deception perform no better than chance levels … Other research suggests it would be premature to dismiss LVA as invalid.”
We’re not going to wade too much more deeply into the back-and-forth about the LVA technology — nor are we making any claims about whether Cain’s or Bialek’s statements were truthful or not — but there’s enough academic research questioning the reliability of the LVA technology as a truth detector to treat the 999 ad’s claim with a healthy dose of skepticism.
— Robert Farley