DOE reduces use of polygraph technology
Agency will end general screening of job applicants, employees
- By Aliya Sternstein
- Oct 16, 2006
DOE final rule on polygraphs
The Energy Department plans to end its across-the-board polygraph testing of job applicants and employees, according to a rule that department officials published in the Federal Register. The policy change becomes effective Oct. 30.
Some researchers say there have been no major physiological or technological advances in the past few years to justify the use of polygraphs for employee security screening at federal agencies. Opponents of polygraph testing have argued for years that DOE and other agencies should scrap such evaluations. Now, after ignoring studies that show that polygraphs are not reliably accurate, DOE has decided to decrease its reliance on such testing for screening prospective counterintelligence employees.
DOE screens employees using a computerized polygraph system to prevent insiders from leaking classified information to the country’s enemies. The department has been administering polygraphs to all employees, consultants and contractor employees before granting them access to sensitive information.
According to the rule, published Sept. 29, DOE will eliminate polygraph testing “for general screening of applicants for employment and incumbent employees without specific cause.” The change conforms to standard practices within the intelligence community and a National Academy of Sciences report, according to the rule.
The 2003 NAS report states that the polygraph’s accuracy in detecting security violators is not impressive enough to justify its use in employee screening. “Overall, the evidence is scanty and scientifically weak,” the final report states. “Our conclusions are necessarily based on the far from satisfactory body of evidence on polygraph accuracy, as well as basic knowledge about the physiological responses the polygraph measures.”
Stephen Fienberg, co-author of the NAS report and a professor at Carnegie Mellon University’s Statistics and Machine Learning departments, said nothing has changed since the report’s release. “There hasn’t been a single published study that suggests that anyone can do polygraph testing better than our original assessment” of the existing research, he said.
Flaws in evaluation
The rationale for using computerized polygraphs is that an examinee’s lies will trigger a specific array of physiological responses that signal deception to the computer and examiner. But Fienberg said no evidence indicates the existence of an unambiguous physiological response that means someone is lying.
In many cases, examiners tell the subjects that the test is completely accurate, Fienberg said. That is a bluff.
“They really want to use it as an interrogation tool,” he said. “It’s not like water-boarding, thank God, but it’s deception because we know that it only works with substantial error.”
Polygraph testing also uses a flawed scoring technique, Fienberg said. Although computerized scoring eliminates subjective analyses, computers cannot improve the weak physiological science that underlies the concept of polygraph testing.
“The computer isn’t doing the exam, the computer isn’t fixing the physiology,” Fienberg said. “It’s only as good as the data allows. It’s not clear that this is going to save the polygraph.”
In addition, he said, some subjective analysis may aid what little benefit polygraphs have to offer.
“If an examiner thinks he or she sees something [revealing] in responses to questions, he or she may focus in on them,” Fienberg said. “Maybe that’s good.”
If there is a signature of deception, a mix of the examiner’s intuition and the standardized computer scoring might produce a more reliable means of spotting it, he said.
“But nobody’s done that that I know,” Fienberg added.
Newer polygraph digital technology looks sleek, but the advancements are all cosmetic, he said.
“They’ve replaced the printer with a flat screen,” he said. “That’s nice. You get pretty pictures. You can see them in real time in front of your face. There’s a convenience associated with it. You could have it on your portable [device]. You could probably have it on your iPod.”
Technology can aid the ruse
DOE’s examiners must be experienced counterintelligence or criminal investigators with extensive training in using computerized instrumentation in the psychophysiological detection of deception, according to the Federal Register rule.
They also must have training in psychology, physiology, interviewing and interrogation. Fienberg scoffed at the computer training requirement, adding that that means the examiner must be
able to attach the instrument’s cuff and know when there is a problem with the recording.
“We’re not talking high-level technology,” he said. “Technology can aid the ruse here, but it’s not changing any of the fundamental things that in our report we suggested were quite problematic. I know that I wouldn’t want to be subjected to a polygraph exam, and I have nothing to hide.”
DOE officials said polygraphs are only part of a more comprehensive security review that also includes background checks and — if appropriate — personal interviews, reviews of financial and credit information, net worth analyses, analyses of foreign travel, and foreign contacts and connections.
“The new regulations have re-emphasized the department’s commitment and resolve to utilize polygraph examinations as one element of the investigative evaluation program,” said DOE spokeswoman Megan Barnett.
DOE follows the Defense Department’s standards for polygraph testing developed by the DOD Polygraph Institute, she said.
Officials at the DOD Counterintelligence Field Activity, which oversees the institute, said their polygraph instrument works by measuring and recording several physiological variables, including blood pressure, heart rate, respiration and skin conductivity, during the course of questing the examinee.
“DOD is aggressively pursuing scientific research on a variety of technologies other than polygraph that could enhance our capability for credibility assessment,” said Louise Dreuth, a spokeswoman at the DOD Counterintelligence Field Activity. Vendors are marketing other credibility assessment technologies, but none of those have demonstrated the scientific validation and reliability standards that DOD requires.
“Polygraph remains the sole instrument within DOD for eliciting statement veracity,” she said.
Opponents of polygraph testing say DOE’s new rule marks a compromise that other agencies should consider.
“I think DOE deserves credit for responding to employee and public concerns,” said Steven Aftergood, director of the Project on Government Secrecy at the Federation of American Scientists. “That is more than other agencies have done.” He added that polygraph analyses mostly reflect the examiner’s intuition rather than an objective measurement.
Aftergood said employees who must undergo polygraph testing might have other matters on their minds, separate from deception, when their answers raise red flags.
“They may be anxious not due to deception but because they need a job,” Aftergood said.